The Belt and Road InitiativersquosEffect on Supply-Chain Trade

Evidence from Structural Gravity Equations1

Tristan Kohl

University of GroningenThe Netherlands

Lund November 13 2018

Annual Conference for the Swedish Network for European Studies inEconomics and Business European Integration

1Forthcoming Cambridge Journal of Regions Economy and Society

Chinarsquos Free Trade Agreements

Source World Trade Organization accessed November 6 2018

Options RCEP TPP

Source Asia Foundation accessed November 12 2018 Note USA withdrew from TPP

The Belt and Road Initiative

Source IRU accessed November 6 2018

This Paper

1 Explores economic outcomes of BRI

2 for southbound westbound and combined land routes

3 for RCEP TPP and a ldquoBRIFTArdquo as trade policy alternatives

4 Accounts for differences in gross and value-added trade

5 Quantifies the effects on economic outcomesmdashie trade andwelfaremdashin general equilibrium

Preview of Findings

1 Distinction between gross and value-added exports mattersmost differences in novel measures of supply-chain tradeless pronounced in the case of China

2 Gains from infrastructural developments far exceed those oftrade-policy alternatives

LITERATURE

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

THEORY

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Full Endowment General Equilibrium

po = (Yo

Y)

11minusσ

1βoΠo

Eo = φoYo = φopoQo

ESTIMATION

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Chinarsquos Free Trade Agreements

Source World Trade Organization accessed November 6 2018

Options RCEP TPP

Source Asia Foundation accessed November 12 2018 Note USA withdrew from TPP

The Belt and Road Initiative

Source IRU accessed November 6 2018

This Paper

1 Explores economic outcomes of BRI

2 for southbound westbound and combined land routes

3 for RCEP TPP and a ldquoBRIFTArdquo as trade policy alternatives

4 Accounts for differences in gross and value-added trade

5 Quantifies the effects on economic outcomesmdashie trade andwelfaremdashin general equilibrium

Preview of Findings

1 Distinction between gross and value-added exports mattersmost differences in novel measures of supply-chain tradeless pronounced in the case of China

2 Gains from infrastructural developments far exceed those oftrade-policy alternatives

LITERATURE

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

THEORY

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Full Endowment General Equilibrium

po = (Yo

Y)

11minusσ

1βoΠo

Eo = φoYo = φopoQo

ESTIMATION

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Options RCEP TPP

Source Asia Foundation accessed November 12 2018 Note USA withdrew from TPP

The Belt and Road Initiative

Source IRU accessed November 6 2018

This Paper

1 Explores economic outcomes of BRI

2 for southbound westbound and combined land routes

3 for RCEP TPP and a ldquoBRIFTArdquo as trade policy alternatives

4 Accounts for differences in gross and value-added trade

5 Quantifies the effects on economic outcomesmdashie trade andwelfaremdashin general equilibrium

Preview of Findings

1 Distinction between gross and value-added exports mattersmost differences in novel measures of supply-chain tradeless pronounced in the case of China

2 Gains from infrastructural developments far exceed those oftrade-policy alternatives

LITERATURE

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

THEORY

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Full Endowment General Equilibrium

po = (Yo

Y)

11minusσ

1βoΠo

Eo = φoYo = φopoQo

ESTIMATION

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

The Belt and Road Initiative

Source IRU accessed November 6 2018

This Paper

1 Explores economic outcomes of BRI

2 for southbound westbound and combined land routes

3 for RCEP TPP and a ldquoBRIFTArdquo as trade policy alternatives

4 Accounts for differences in gross and value-added trade

5 Quantifies the effects on economic outcomesmdashie trade andwelfaremdashin general equilibrium

Preview of Findings

1 Distinction between gross and value-added exports mattersmost differences in novel measures of supply-chain tradeless pronounced in the case of China

2 Gains from infrastructural developments far exceed those oftrade-policy alternatives

LITERATURE

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

THEORY

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Full Endowment General Equilibrium

po = (Yo

Y)

11minusσ

1βoΠo

Eo = φoYo = φopoQo

ESTIMATION

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

This Paper

1 Explores economic outcomes of BRI

2 for southbound westbound and combined land routes

3 for RCEP TPP and a ldquoBRIFTArdquo as trade policy alternatives

4 Accounts for differences in gross and value-added trade

5 Quantifies the effects on economic outcomesmdashie trade andwelfaremdashin general equilibrium

Preview of Findings

1 Distinction between gross and value-added exports mattersmost differences in novel measures of supply-chain tradeless pronounced in the case of China

2 Gains from infrastructural developments far exceed those oftrade-policy alternatives

LITERATURE

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

THEORY

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Full Endowment General Equilibrium

po = (Yo

Y)

11minusσ

1βoΠo

Eo = φoYo = φopoQo

ESTIMATION

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Preview of Findings

1 Distinction between gross and value-added exports mattersmost differences in novel measures of supply-chain tradeless pronounced in the case of China

2 Gains from infrastructural developments far exceed those oftrade-policy alternatives

LITERATURE

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

THEORY

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Full Endowment General Equilibrium

po = (Yo

Y)

11minusσ

1βoΠo

Eo = φoYo = φopoQo

ESTIMATION

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

LITERATURE

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

THEORY

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Full Endowment General Equilibrium

po = (Yo

Y)

11minusσ

1βoΠo

Eo = φoYo = φopoQo

ESTIMATION

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

THEORY

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Full Endowment General Equilibrium

po = (Yo

Y)

11minusσ

1βoΠo

Eo = φoYo = φopoQo

ESTIMATION

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

THEORY

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Full Endowment General Equilibrium

po = (Yo

Y)

11minusσ

1βoΠo

Eo = φoYo = φopoQo

ESTIMATION

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

THEORY

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Full Endowment General Equilibrium

po = (Yo

Y)

11minusσ

1βoΠo

Eo = φoYo = φopoQo

ESTIMATION

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Literaturendash Gravity Model of International Trade provides theoretical

microfoundations to study the determinants of internationaltrade flows as a function trade costs (Tinbergen 1962Bergeijk amp Brakman 2010 Head amp Mayer 2014)

ndash Trade Costs increase with geographic distance restrictivetrade policy measures and other sources of transportationcosts (McCallum 1995 Anderson amp van Wincoop 2004)

ndash Transportation Costs decrease with infrastructuraldevelopments and adoption of new transportationtechnologies eg construction of railway networks canalopenings (closings) and introduction of steamships(Bougheas et al 1990 Pascali 2017 Donaldson 2018 Feyrer1999 2018)

ndash Institutions also affect trade costs through shared culturaladministrative and legal heritage (Head et al 2010 Ku ampZussman 2010 Francois amp Manchin 2013) and trade policy(Rose 2004 Baier amp Bergstrand 2007 Kohl et al 2016)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

THEORY

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Full Endowment General Equilibrium

po = (Yo

Y)

11minusσ

1βoΠo

Eo = φoYo = φopoQo

ESTIMATION

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

THEORY

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Full Endowment General Equilibrium

po = (Yo

Y)

11minusσ

1βoΠo

Eo = φoYo = φopoQo

ESTIMATION

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

THEORY

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Full Endowment General Equilibrium

po = (Yo

Y)

11minusσ

1βoΠo

Eo = φoYo = φopoQo

ESTIMATION

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Structural Gravity

ndash Quantitative trade models with counterfactual scenarios toestimate trade and welfare effects of changes in trade costs(Costinot amp Rodriguez-Clare 2014 Yotov et al 2016Anderson et al 2018)

ndash Recent applications consequences of abolishing bordersoptions for ldquoGlobal Britainrdquo after Brexit (Anderson amp Yotov2016 Brakman et al 2018)

ndash New development incorporating the role of supply-chaintrade (Caliendo amp Parro 2015 Aichele amp Heiland 2018Kaplan et al 2018)

THEORY

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Full Endowment General Equilibrium

po = (Yo

Y)

11minusσ

1βoΠo

Eo = φoYo = φopoQo

ESTIMATION

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

THEORY

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Full Endowment General Equilibrium

po = (Yo

Y)

11minusσ

1βoΠo

Eo = φoYo = φopoQo

ESTIMATION

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Full Endowment General Equilibrium

po = (Yo

Y)

11minusσ

1βoΠo

Eo = φoYo = φopoQo

ESTIMATION

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Full Endowment General Equilibrium

po = (Yo

Y)

11minusσ

1βoΠo

Eo = φoYo = φopoQo

ESTIMATION

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Full Endowment General Equilibrium

po = (Yo

Y)

11minusσ

1βoΠo

Eo = φoYo = φopoQo

ESTIMATION

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Full Endowment General Equilibrium

po = (Yo

Y)

11minusσ

1βoΠo

Eo = φoYo = φopoQo

ESTIMATION

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Intuition

ndash Estimate baseline gravity equation

ndash Predict baseline trade costs

ndash Introduce counterfactual change in trade costs 15 3050 reduction in bilateral distances or ldquoswitch onrdquo FTA binaryvariable

ndash Iteratively solve to obtain counterfactual multilateral resistanceterms output and trade based on last available year of data

ndash Calculate change in normalized trade with respect to thebaseline for trade and welfare (presented today)

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Full Endowment General Equilibrium

po = (Yo

Y)

11minusσ

1βoΠo

Eo = φoYo = φopoQo

ESTIMATION

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Full Endowment General Equilibrium

po = (Yo

Y)

11minusσ

1βoΠo

Eo = φoYo = φopoQo

ESTIMATION

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Solution (see Yotov et al 2016)

Partial Equilibrium for Given MRTs Production and Expenditure

Xod =YoEd

Y(

tCFLod

ΠoPd)1minusσ

Conditional General Equilibrium for Changed MRTs GivenProduction and Expenditure

Π1minusσo = Σd(

tCFLod

Pd)1minusσ Ed

Y

P1minusσd = Σo(

tCFLod

Πo)1minusσ Yo

Y

Solution (see Yotov et al 2016)

Full Endowment General Equilibrium

po = (Yo

Y)

11minusσ

1βoΠo

Eo = φoYo = φopoQo

ESTIMATION

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Solution (see Yotov et al 2016)

Full Endowment General Equilibrium

po = (Yo

Y)

11minusσ

1βoΠo

Eo = φoYo = φopoQo

ESTIMATION

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

ESTIMATION

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Procedure for Counterfactual Distances

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1ln(DSTod) + η2CTGod + η3BRDod + η4TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1ln(DSTod)

CFL + η2CTGod + η3BRDod + η4TAod ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Procedure for Counterfactual FTA

Step 1a Estimate Baseline

Xodt = exp[πot + pdt + ζod + η1TAodt ] + 983171odt

Step 1b Predict Baseline Trade Costs

tBSLNodt = exp[η1TAodt ] + 983171odt

Step 2 Predict Counterfactual Trade Costs

tCFLod = exp[η1TACFL

od ] + 983171od

Step 3 Solve the Baseline and Counterfactual Model

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Gross Exports v Value-Added Exports

ndash Gross export statistics suffer from ldquodouble countingrdquoproblematic for countries in supply-chain trade

ndash Value-added exports more closely linked to incomes ofcountries in supply-chain trade including (non-tradable)services used in global value chains (Johnson amp Noguera2012 Koopman et al 2012 Johnson 2014 Timmer et al2014)

ndash Several measures of ldquovalue-added exportsrdquo mostly focusingon the location of final demand for value-addedproductionmdashpotentially problematic for bilateral analyses(Los amp Timmer 2018)

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Gross Exports v Value-Added Exports

A rarr B rarr C rarr D rarr E

1 +1 +1 +1

From B C DTo C D E D E EGross exports 2 0 0 3 0 4Domestic value-added for consumption (VAX-C) 0 0 1 0 1 1 direct use (VAX-D) 1 0 0 1 0 1Note Based on Los amp Timmer (2018 Table 1)

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Data

ndash OECD Trade in Value Added (TiVA) Databasendash 63 countries + Rest-of-the-World (ROW) 1995-2011ndash GX (gross exports) and FV VA (value-added content of final

demand) series

ndash World Input-Output Database (Timmer et al 2015 Los ampTimmer 2018)

ndash 43 countries + Rest-of-the-World (ROW) 2000-2014ndash VAX-C (domestic value-added for consumption) and VAX-D

(domestic value-added for direct use)

ndash Distance common border (Mayer amp Zignano 2011) and FTAsKohl (2014) WTO

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

TiVA vs WIOD Country Coverage

Africa amp Middle East Americas Asia amp Pacific EuropeIsrael Argentina Australia Philippines EU28Morocco Brazil Brunei Russia IcelandSouth Africa Chile Cambodia Saudi Arabia NorwayTunisia Colombia China Singapore SwitzerlandTurkey Costa Rica Hong Kong South Korea

Peru India Taiwan ROWCanada Indonesia ThailandMexico Malaysia VietnamUSA New Zealand

Note Blue marks countries only in TiVA

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

RESULTS

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Cfrsquol Distance mdash Southbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Trade Effects

Welfare

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Cfrsquol Distance mdash BRI mdash Trade Effects

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Cfrsquol Distance mdash BRI mdash Welfare Effects

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash Summary mdash Trade Effects

BRIFTA RCEP TPP

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash Summary mdash Welfare Effects

BRIFTA RCEP TPP

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

CONCLUSION

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Conclusion

ndash Counterfactual analysis exploring trade and welfare effects ofBRI in general equilibrium Accounting for supply-chain tradematters

ndash BRI-based infrastructural improvements have mostpronounced effect on value-added trade for Russia Chinaand less so for EU The welfare effects tend to be small

ndash Trade-policy alternatives are not appealing for China (RCEPTPP) ldquoBRIFTArdquo marginally so due to access to Europeanmarket

ndash Extensions in future research design of trade agreements(Kohl et al 2016 Baier et al 2018) sectoral impact analysestrade cost reductions per mode of transport

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

TACK FOR YOUR ATTENTION

The Belt and Road Initiativersquos Effect on Supply-Chain TradeEvidence from Structural Gravity Equations

Tristan Kohl

tkohlrugnl

wwwtristankohlorg

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Cfrsquol Distance mdash Southbound mdash Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Cfrsquol Distance mdash Westbound mdash Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash RCEP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash TPP mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back

Cfrsquol FTA mdash BRI mdash Trade amp Welfare Effects

Back