Lectures_Spring 2015.pdf

Political Economics III, Spring 2015

The Political Economics of Development Clusters

Torsten PerssonIIES, Stockholm University

1

A. BackgroundHuge income disparities

Massive gap between rich and poor countries

a ratio of income per capita on the order of 200is a common starting point

Why are some countries rich and others poor?

classical question in economics, and in other social sciencesalso of paramount importance for donors in various formsof development assistance

But development not only about income

very clear in policy discussion about weak/fragile states

2

Fragile states — Figure 1.2

Central concept in development policy community

subject of various initiatives

What is a fragile (weak) state?

it can not support basic economic functions, raise anysubstantial revenues, deliver basic services, keep law and order, ...

Existing indexes

examples from Brookings (in book) and Polity IV (here),though definitions appear to mix up symptoms and causes

incidence depends on definition, but 20-30 states failed/very weakequally many fragile/weak, and others in risk zone

concentration in Sub-Saharan Africa, south/central Asia

3

tpersTypewritten Text

tpersTypewritten Text

tpersTypewritten Text

tpersTypewritten Text

tpersTypewritten Text Figure 1.2 Polity IV Index of Fragile States 2009

Development clusters

Income links not only with state institutions, but with violence

massive poverty associated with weak state institutionsand societies plagued by internal conflicts

developed countries: high income, institutions work,policies in good order, conflicts resolved peacefully, ...

strong clustering of outcomes in different dimensionsfew strong economies with weak states

Multidimensional problem — the development problem?

clustering of low income, violence, andof dysfunctional state institutions

4

Example of clustering — Figures

Three forms of state capacity

infrastructure to raise revenue by levying taxes onbroad bases as income (value added) — fiscal capacity

infrastructure to support markets by enforcingcontracts or protecting property rights — legal capacity

infrastructure to augment markets by supplyingpublic goods — collective capacity

Illustrate with specific measures

share of income taxes in total taxes (1999)index of contract enforcement (2006)index of school attainment and life expectancy (early 2000s)strongly correlated with each other, GDP/capita (2000),and index of civil war and repression (1975-2006)

5

6

7

How understand such patterns in the data?

Basically need to pose — and answer — three general questions

Question 1what forces drive building of different state capacities,and why do these capacities move together?

Question 2what forces drive different forms of political violence?

Question 3what explains clustering of institutions, income, and violence?

6

Scope of book

Some over-arching objectives

analyze the politics and economics of state building andpolitical violence in the process of development

try to understand the observed development clusters ofinstitutions, income, and violence

aim at constructing new theory and uncovering new evidencehope to bring these issues into mainstream of economics

Pool together four broad research agendas

determinants of long-run developmentdeterminants of different forms of political violenceimportance of history in explaining today’s patterns of developmentinteraction of economics and politics in shaping of societies

7

Background — earlier and newer publications

"Wars and state capacity", Journal of European EconomicAssociation, 2008

"Repression or civil war?", American Economic Review,Papers and Proceedings, 2009

"The origins of state capacity: Property rights, taxationand politics", American Eeconiomic Review, 2009

"State capacity, conflict and development", Econometrica, 2010

"Fragile states and development assistance", Journal ofEuropean Economic Association, 2011

10

"The logic of political violence", Quarterly Journalof Economics, 2011

"Taxation and development", Handbook ofPublic Economics, 2013

"Weak states and steady states: The dynamics of fiscal capacity",(with Ethan Ilzetzki), American Economic Journal Macro, 2013

"The causes and consequences of development clusters: Statecapacity,peace, and income", Annual Review Economics, 2014

"Why do developing countries tax so little?",Journal of Economic Perspectives, 2014

11

These lectures

Try to tell the major story

describe overall approach and main messages of bookuse the core, macroeconomic and macropolitical, modelvery brief on extensions, microfoundations, and referenceslook at data as motivation and in somewhat greater depth

Road map

A. BackgroundB. Fiscal and Legal CapacityC. Political ViolenceD. State SpacesE. Development AssistanceF. Political ReformG. Lessons Learned?

9

B. Fiscal and Legal CapacityFiscal capacity — Existing research

Ignored, or assumed, in mainstream economics

(macro) development economics sees income per capita,not fiscal institutions, as central outcome

capacity to raise revenue from certain tax bases basicallyassumed in development, public finance, political economics, ...

Extractive government important in political and economic history

fiscal powers important in themselves, for military success andstate development, more generally (Hintze, Schumpeter, Tilly)

war major motive to build fiscal capacity‘war made the state and the state made war’ (Tilly, 1990)

10

Expansion of taxation in rich countries — Figures

Last century — vast expansion of government size

1910: total taxes around 10% of GDP in Europe and US,while today’s figures are 30-50%

number of innovations and expansions of infrastructureunderpin the capacity to raise so much revenue

Investments in fiscal capacity over time

dating of reforms in 75 (mostly) rich countriesintroduction of income tax 1840s-1970s, income-taxwithholding later, VAT still not complete

concentration of reforms around world warsadditional data from Besley and Persson (2013)

11

0.2

.4.6

Ave

rage

sha

re o

f Inc

ome

Tax

0.1

.2.3

Ave

rage

sha

re o

f tax

in a

ggre

gate

inco

me

1900 1920 1940 1960 1980 2000Year

Tax revenue in aggregate income (left scale)Share of income tax in revenue (right scale)

Evolution of tax revenue and income tax for a sample of 18 Countries

Figure 1: Taxes and share of income tax over time

14

0.2

.4.6

.81

Pro

porti

on o

f Cou

ntrie

s

1800 1850 1900 1950 2000year

Income Tax VAT

Fiscal capacity in a sample of 75 countries

Figure 2.1 The historical evolution of fiscal capacity

Weak states in poor and low-tax countries — Figures

Stylized facts about tax take

rich countries collect a much larger share of their incomein taxes than do poor countries

rich countries rely to a much larger extent on income taxes,as opposed to trade taxes, than do poor countries

high-tax countries rely to a much larger extent on incometaxes, as opposed to trade taxes, than do low-tax countries

rich countries collect much higher tax revenue than poorcountries despite comparable statutory rates

Illustration of these stylized facts

these facts hold in the contemporary cross section and overthe last century, in a sample of 18 (currently rich) countries

13

Taxes and income in the cross section and time series

Income taxes and trade taxes in the cross section and time series

Income taxes and total taxes in the cross section and time series

Background — legal capacity

Also interested in productive role of government — legal capacity

government efforts to make private economy more productivefocus on legal protection, subject to legal infrastructure

will allow us to endogenize income

Two views of long-run causes of low productivity

it reflects lacking technology — the Solow traditionit reflects misallocated resources — the Lewis tradition

We will take the second view

poorly functioning economic institutions generate frictions incontracting or protection of property

potential for improvement by investing in legal infrastructure

14

Empirical motivation — Figures 1.3 and 3.1

Legal and fiscal capacity strongly correlated

both with each other and incometrue for different measurestotal tax take in GDP vs. index for protection ofproperty rights — Figure 1.3

income tax share in total taxes vs. index of contractenforcement — Figure 3.1

15

 

Figure 1.3 Legal and fiscal capacity conditional on income

0.2

.4.6

.8S

hare

of i

ncom

e ta

x in

tota

l tax

es 1

990-

2000

0 .2 .4 .6 .8 1Contract Enforcement (Normalized Country Rank)

High income in 1990 Middle income in 1990Low income in 1990 Fitted values

Income taxes and Contract Enforcement by GDP

 

Figure 3.1 Income taxes and contract enforcement conditional on GDP

Legal institutions — Existing research

Market-supporting institutions emphasized in economics and history

North-Weingast on economic institutions — crucial for growthand unleashed by events like Glorious Revolution

Acemoglu-Johnson-Robinson (Hall-Jones) on productive vs.extractive institutions, triggered by nature of colonial settlement

Engerman-Sokoloff on the how inequality (factor endowments)shaped more or less productive institutions across the Americas

Political and legal origins of financial institutions

political origins of weak institutions, due to rent-seeking,polarization, etc. — Svensson, Rajan-Zingales, Pagano-Volpin

institutions may have deep historical roots as inLa Porta, Silanes, Shleifer and Vishny on legal origins

Our approach: other mechanisms and legal + fiscal capacity16

Road map — Figure 1.7

1. Basic Structure of Core Model2. Policy3. Investments in State Capacity4. Comparative Statics5. Data and Partial Correlations6. Microfoundations of Fiscal Capacity7. Microfoundations of Legal Capacity8. Microfoundations — Predation and Corruption

17

Legal capacity Fiscal capacity

Common vs. redistributive interests

Cohesiveness of political institutions

Resource or (cash) aid independence

Income per capita

Figure 1.7 Scope of Chapter 3

Political stability

1. Basic structure of core model

Two time periods, = 1 2

(infinite horizon: extension in text ch. 2)

Two identical groups of individuals, =

each has share 12 of population size, normalized to 1(asymmetries: extension in text ch. 2)

Incumbents and opponents

at beginning of = 1 one group holds powerwe call this group the incumbent 1 ∈ {}

the other group is the opponent 1 ∈ {}with exogenous probability there is a peacefultransition of power until = 2

thus measures political instability (turnover)(to be endogenized in part C)

18

Private utility

Linear utility functions

linear utility buys us risk neutralityand a model that is recursive in policy and investments

= +

private consumption of group- member at no savings (one of extensions in text of ch. 3) utility from consumption of public goods, their value;think about as "defense", and "threat of external conflict"(curvature: extension in text ch. 2, back to this in part E.)

19

Value of public goods

Value of public goods stochastic

has two-point distribution ∈ { }where 2 1 and Prob[ = ] = (continuous distribution: extension in text ch. 2)

shocks to iid over timerealization of known when policy set in

20

Taxation and fiscal capacityGovernment has discretion over current taxation

taxes income at rate , but is constrained byexisting fiscal capacity, i.e., ≤

Microeconomic foundations (see Section 6 below)

individual can earn some income in informal (untaxed) sector,but incentives to hide depend on risk and cost of getting caught

Investments in fiscal capacity

e.g., tax authority, compliance structures, infrastructure toenforce income tax (or impose value-added tax)

initial stock 1 is given, but can be augmentedto achieve fiscal capacity 2 requires non-negative investment2 − 1 at = 1 (depreciation and reversibility in text ch. 2)

convex cost F(2 − 1) where F (0) = 021

Incomes and legal protection

Group 0s income depends on "legal protection"

= ( )

where is an increasing function

no tax distortions (extension in text ch. 2)think of as "legal protection of group contracts"or "legal protection of group property rights"

Alternative microfoundations in two-sector model

(i) credit-market model with partial enforcement of collateralizeddebt contracts: higher better enforcement (Section 7)

(ii) model of coercive theft of output by predatory activity: higher more clamp-down on predation (in text ch. 3)

22

Legal protection and capacity

Incumbent controls current legal protection

constrained by existing legal capacity, i.e., ≤

Investment in legal capacity

e.g., courts, educated judges, credit or property registriesinitial stock of legal capacity, 1, given, but can be augmentedby non-negative investment 2 − 1

convex costs L(2 − 1)where L(0) = 0

23

Government budget

Budget items at

{ }= and total investments

=

½F(2 − 1) + L(2 − 1) if = 1

0 if = 2

budget constraint is

+ () + (

)

2= + +

+

2

where is a non-negative targeted transfer to group is additional (constant) revenue source accruing to governmentinterpret as natural resource rents, or foreign (cash) aid is randomly distributed on support [ ]

24

Political institutions

Model as constraint on incumbent

incumbents must give fixed share to oppositionof any given unit of transfers to its own group

by the budget constraint

= [ +

() + ( )

2− −]

where = 2(1− ) and = 2 and where 0s share = 1+ ∈ [0

12] represents more cohesive institutions

the closer is to its maximum of 12interpret as more checks and balances on executive,or better representation of opposition(micropolitical foundations in text ch. 7)

25

Timing

1. Start out with state capacity {1 1} and incumbent group 1nature determines 1 and

2. 1 chooses a set of first-period policies {(1 ) (1 ) 1 1} andinvestments in period-2 state capacities 2 and 2.

3. 1 remains in power with probability 1− , nature determines 24. New incumbent 2 chooses current policy {(2 ) (2 ) 2 2}

goal is to solve for a subgame-perfect equilibrium in policy,and state-capacity investments — treat in that order

26

2. PolicyPolicymaking in period

Policy objective

linearity makes model recursive, so that we can studypolicy choice at stages 2 and 4 separately from investments

whoever holds power, choosesn( ) (

)

oto maximize

+ (1− ) () + subject to

≤ ≤ , ≥

and the government budget constraint

Optimal policy design?

can be described by four observations

27

Observation 1 — legal protection

Will legal protection be allocated in same way to each one ofthe groups — i.e., will there be rule of law?

For ∈ {1 2} any incumbent any and any regulation fully utilizes all legal capacity, = =

"Obvious" result in the current set up

relates to Diamond-Mirrlees production efficiencyand a Political Coase Theorem

this result can be violated, when there are rents(in text ch. 3)

does not mean that legal protection will be extensivethis depends on investment in legal capacity

28

Observation 2 — public goods

Equilibrium public-good provision

linear preferences give us a "bang-bang", corner solutionthe level of public goods provided is

( ) =

½ + ()− if ≥ 2 (1− )0 if 2 (1− )

depending on whether public goods is worth more to the incumbentthan transfers to her own group (1st row), or not (2nd row)

29

Observation 3 — taxes

Equilibrium tax rate =

Interpretation

always worthwhile to fully utilize all fiscal capacity, since gainof higher tax rate is, at least, 2 (1− ) () while loss is ()

30

Observation 4 — transfers

Equilibrium transfers to incumbent group

follow from

= [ + ()− ( )−]

Interpretation — recall = 2(1− ) and = 2higher value of the opposition’s share, reflects morecohesive political institutions

as stated earlier, this may reflect more minority protectionby constitutional checks and balances, or more representationthrough PR elections or parliamentary form of government

if = 12, transfers shared equally across the two groups

31

Indirect utility and value functions

Plug in optimal policy in utility at to get

( ) = ( ) + (1− )() +

[ + ()− ( )−]period utility of group

Define "value functions"

(2 2) = ( 2 2 0 ) + (1− ) ( 2 2 0 )

and

(2 2) = ( 2 2 0 ) + (1− ) ( 2 2 0 )

for being incumbent or opposition group in period 2depending on the two state variables

32

3. Investments in State CapacityPreliminaries

Investment objective is

(1 1 1F(2 − 1) + L(2 − 1) 2(1− ))+(1− )(2 2) + (2 2)

What’s the shadow cost of public funds for incumbent?

value realized in period 1

1 = max {1 2 (1− )}and value expected for period 2

(2) = + (1− )2where

2 =

½ if ≥ 2(1− )2[(1− )(1− ) + ] otherwise

33

Euler equations

First-order conditions

for fiscal and legal capacity are

(2)[((2)− 1] 0 1F (2 − 1)c.s. 2 − 1 > 0

(2)[1 + ((2)− 1)2] 0 1L (2 − 1)c.s. 2 − 1 > 0

Marginal cost of investment — RHS

period-1 foregone consumption of public or private goods

Marginal net benefit of investment — LHS

collect any direct effect on period-2 private income plusindirect effects via the government budget

34

When is investment positive?

Because F (0) = L(0) = 0 it is sufficient that(2)− 1 ≥ 0

expected value of public funds must to be large enoughthis depends on key parameters: { }

Immediate interim agenda

analyze optimal investmentunderstand how it depends on the model parameters

35

Two conditions

To pin down the type of equilibrium, define

Cohesiveness: ≥ 2 (1− )requires close enough to 12 or large enough i.e., strong enough common-interest vs. redistributive motives

guarantees that (2)− 1 ≥ 0

Stability: + (1− ) 2 [(1− ) (1− ) + ] ≥ 1relevant only when Cohesiveness fails — depends on e.g., holds as → 0 even if → 0

also guarantees that (2)− 1 ≥ 0

These conditions uniquely define three possible outcomes

36

Three types of state

Proposition 2.2 If Cohesiveness holds, then the outcome isa common-interest state (the same as chosen by a Pigouvianplanner). Public goods are provided for any and there ispositive investment in fiscal and legal capacity

Proposition 2.3 If Cohesiveness fails, while Stability holds, thestate is redistributive. Public revenues finance only transfers when = and the state invests in both fiscal and legal capacity

Proposition 2.4 If Cohesiveness and Stability fail, the state isweak with no investments in fiscal capacity and lower investmentsin legal capacity than in a common-interest or redistributive state

this is one dimension of our state-space matrix in part D.

37

Complementarity and supermodularity

Complementarity

a further consequence of (2)− 1 ≥ 0has two important implications

Substance

higher raises motives to invest in and vice versa

Analytical convenience — monotone comparative statics

supermodularity holds (by positive cross-partial)if reduced-form objective function (2 2;) supermodularin (2 2) then (2 2) monotonically increasing in if 2 (·) 2 ≥ 0 and 2 (·) 2 ≥ 0

very easy to derive effects of most parameter shifts

38

4. Comparative StaticsValue of public goods

Proposition 3.2 Higher expected demand for public goodsraises investments in state capacity in common-interestand redistributive states

(2)

= − 2 0

common interests make fiscal capacity more valuable

external conflict promotes fiscal capacity and legal capacityconsistent with historical work by Hintze, Tilly and others,but augmented prediction for productive side of government

39

Political instability and cohesiveness

Proposition 3.3 Investment in fiscal and legal capacity are promotedby lower political instability if institutions are not cohesive

lower raises the likelihood that Stability holds andincreases 2 if it does hold

this effect is stronger, the more non-cohesive political institutionscase study of England in 18th century: after Glorious Revolution(higher ), Whigs rule for many decades (low ), greatexpansion of tax capacity, and more independent andwell-paid judiciary (higher )

more cohesiveness has an uncertain effects on state capacity inredistributive state, but raise probability of common-interest state

40

Costs of investments

Proposition 3.4 Lower costs of either legal or fiscal capacity raiseinvestments in legal and fiscal capacity in common-interestand redistributive states

a downward multiplicative shift of L(·) or F(·) cutsthe RHS of investment FOCs for given 2 and 2

this gives a theoretical rationale for "legal origins" hypothesis,but with an auxiliary prediction for fiscal capacity

41

Exogenous growth and income

Exogenous productivity differences

= Λ³

´perhaps due to geography or Hicks-neutral technology

Proposition 3.5 More productive economies (higher Λ2) choosegreater investments in fiscal and legal capacity in common-interestand redistributive states.

higher Λ2 raises Λ2(2) and Λ2(2) for given 2 whichmakes both types of investments in the state more worthwhile

42

Resource or aid dependence

Define equilibrium GDP in period as

() = +Λ( () + ())

2

and consider variations in andΛ () that keep () constant

Corollary Higher resource or aid dependence, higher for given (2 ) means lower investments in legal and fiscal capacityin common-interest and redistributive states

clue why some aid or resource-dependent countries in Africaand South Asia may have weak incentives to build their states

consistent with idea of "rentier states"

43

Endogenous growth

The model also has "endogenous" growth

income grows due to investments in legal capacitywhatever the source of these investments

(2 )− (1 ) (1 )

growth driven by institutional deepening leading tomore efficient private markets, when 2 1

by complementarity, (expected) government size growstogether with legal capacity and income

44

Clustering of state capacity and income

Strong positive associations

recall correlations in Figures 1.3 and 3.1

Earlier results shed light on observed clustering

positive correlation can reflect higher (exogenous)income causing higher state capacity

but may also reflect other factors that lead to higherstate capacity, which — in turn — spills over intohigher (endogenous) income

45

Extension: Polarization/heterogeneity

Different valuations of public goods across groups

assume drawn from same two-point distribution { }n

operiod- realizations for groups and and

(1− ) = Prob { = | = } ≤ 1greater polarization/heterogeneity, higher , giveslower expected value of public funds

(2)

= −( − ) 0

Proposition 2.5 If Cohesiveness fails, more polarization (higher )decreases fiscal and legal capacity-investments in redistributive states,and raises the likelihood of a weak state. Both effects are largerwith greater political instability (higher )

46

5. Data and Partial CorrelationsMeasuring state capacity

Five proxies for fiscal capacity (IMF and World Bank data)

ratio of total tax revenue to GDP, at end of 1990sshare of income taxes in total revenue, at end of 1990sshare of non-trade taxes in revenue at end of 1990sdifference between income-tax and trade-tax share1− (share of informal economy in GDP around 2006)

Five proxies for legal capacity (ICRG and World Bank data)

index of government anti-diversion policy, end of 1990snormalized rank on Doing Business indicators, circa 2006normalized rank on ease of registering propertynormalized rank in the ease of access to creditnormalized rank on a measure of enforcing contracts

47

Table 2.1 Correlations between fiscal capacity measures  

Tax revenue share in GDP

Income tax share Non-trade tax share

Income tax bias Formal sector share

Tax revenue share in GDP

1.000

Income tax share

0.815 1.000

Non-trade tax share

0.729 0.693 1.000

Income tax bias

0.846 0.954 0.878 1.000

Formal sector share 0.564 0.587 0.580 0.624 1.000  

Table 3.1 Correlations between legal capacity measures

Government Anti-diversion Policy

Doing Business Registering Property

Obtaining Credit Contract Enforcement

Government Anti-diversion Policy

1.000

Doing Business

0.8010 1.000

Registering Property

0.5082 0.5670 1.000

Obtaining Credit

0.6680 0.7879 0.4360 1.000

Contract Enforcement

0.7277 0.7062 0.3851 0.4069 1.000

Measuring parameters of the model

Use various proxies

common interests: proportion years in external war from 1816(or independence) until 2000 (Correlates of War data)

polarization: 1− (degree of ethnic fractionalization)(Fearon 2003 data on (0,1))

cohesive institutions: average from 1800 (or independence)to 2000 of constraints on executive ("Xconst" in Polity IV data,1-7 scale normalized to (0,1))

political stability: same period average of non-open andnon-competitive recruitment of executive (normalized (0,1)score for "Xrcomp"+"Xropen" in Polity IV)

investment costs: legal origin indicators (La Porta et al 1998)

48

Partial correlations — Figures and tables

Compute partial correlations

regress measure of state capacity on suggested determinants;of course, no claim of causal interpretation

Basic correlations in line with theory

for different measures of fiscal as well as legal capacityin cross sectional, as well as in panel data (conditional oncountry and year fixed effects)

Auxiliary predictions of theory?

interaction effects are mixed successadditional measures implied by extensions (in text ch. 3):private investments, private credit, corruption — also correlatedwith basic determinants in line with model predictions

49

Table 2.2 Fiscal Capacity and Covariates: Simple Correlations (1) (2) (3) (4) (5) Tax revenue share

in GDP in 2000

Income tax share in 2000

Non-trade tax share in 2000

Income tax bias in 2000

Formal sector share around 2000

Prevalence external war before 2000

1.897* (1.142)

1.213 (0.952)

2.387** (0.915)

1.972** (0.965)

1.671** (0.690)

Average executive constraints before 2000

2.130*** (0.374

2.309*** (0.335)

1.135*** (0.312)

2.001*** (0.307)

1.768*** (0.356)

Average non-open executive recruitment before 2000

1.080** (0.432)

1.254*** (0.451)

0.541 (0.391)

1.054*** (0.392)

1.490*** (0.447)

Ethnic homogeneity (1 - ethnic fractionalization)

1.058*** (0.300)

0.438 (0.271)

0.656** (0.304)

0.606** (0.270)

0.709** (0.298)

Observations 104 104 103 103 109 R-squared 0.503 0.465 0.301 0.482 0.317 Notes: Robust standard errors in parentheses: (* significant at 10%; ** significant at 5%; *** significant at 1%)  

Table 2.3 Fiscal Capacity and Covariates: Interaction Terms (1) (2) (3) (4) (5)

Tax revenue share in GDP in 2000

Income tax share in 2000

Non-trade tax share in 2000

Income tax bias in 2000

Formal sector share around 2000

Prevalence external war before 2000 3.136 (2.928)

1.221 (3.076)

7.819*** (2.426)

4.604** (2.288)

- 1.029 (2.790)

External war*High executive constraints dummy

- 1.539 (3.030)

- 0.134 (3.180)

- 6.204** (2.449)

- 3.096 (2.421)

3.176* (2.880)

Average non-open executive recruitment before 2000

1.934* (1.167)

2.074*** (0.683)

1.053* (0.536)

1.834*** (0.635)

1.125** (0.562)

Non-open executive recruitment* Low executive constraints dummy

- 1.425 (1.140)

- 1.176 (0.774)

- 0.838 (0.636)

- 1.156 (0.701)

0.961 (0.630)

High executive constraints dummy 0.495 (0.388)

0.169 (0.371)

0.010 (0.516)

0.080 (0.365)

- 0.572 (0.460)

Average executive constraints before 2000

1.083* (0.596)

1.790*** (0.543)

1.078** (0.516)

1.679*** (0.501)

2.772*** (0.658)

Ethnic homogeneity (1 - ethnic fractionalization)

0.774** (0.312)

0.233 (0.285)

0.472 (0.358)

0.389 (0.291)

1.021*** (0.352)

Observations 104 104 103 103 109R-squared 0.550 0.490 0.337 0.503 0.352

Notes: Robust standard errors in parentheses: (* significant at 10%; ** significant at 5%; *** significant at 1%)

Table 2.4 Fiscal Capacity and Covariates: Additional Controls (1) (2) (3) (4) (5) (6) Tax revenue

share in GDP

Income tax share in total

revenue

Formal sector share

Tax revenue share in GDP

Income tax share in total

revenue

Formal sector share

Prevalence external war before 2000

1.536 (1.076)

0.884 (0.867)

1.203* (0.660)

0.819 (1.341)

0.583 (0.860)

1.484** (0.659)

Average executive constraints before 2000

1.595*** (0.415)

1.757*** (0.383)

0.891** (0.397)

1.163** (0.452)

1.240*** (0.402)

1.131** (0.429)

Average non-open executive recruitment before 2000

0.686* (0.408)

0.866** (0.410)

0.989** (0.428)

0.891* (0.474)

0.473 (0.396)

1.249** (0.475)

Ethnic homogeneity (1 - ethnic fractionalization)

0.718* (0.368)

0.085 (0.339)

- 0.010 (0.372)

0.423 (0.384)

0.024 (0.322)

0.084 (0.397)

Log(GDP per capita) in 2000

0.209** (0.105)

0.221** (0.099)

0.398*** (0.106)

0.350*** (0.112)

0.342*** (0.083)

0.378*** (0.117)

Low value of inequality

0.513* (0.297)

0.321** (0.151)

- 0.182 (0.191)

Observations 103 103 109 83 83 90 R-squared 0.531 0.496 0.404 0.591 0.570 0.480

Notes: Robust standard errors in parentheses: (* significant at 10%; ** significant at 5%; *** significant at 1%)  

 

Explaining fiscal capacity over time, 18 countries 1900-2000

 

(1)

(2) (3) (4)

War 0.085** (0.036)

0.103** (0.043)

0.075** (0.022)

0.067*** (0.020)

Political stability 0.063 (0.052)

0.141** (0.058)

0.052 (0.031)

0.080** (0.031)

Cohesive political institutions

0.086*** (0.023)

0.094*** (0.024)

0.045** (0.027)

0.058* (0.035)

Income Per Capita (thousands $US)

– 0.003 (0.010)

Country Fixed Effects

No No Yes Yes

Time Fixed Effects No Yes Yes Yes Number of observations 300 300 300 290 Adjusted R2 0.164 0.509 0.822 0.848

Notes: Dependent variable is ratio of tax to aggregate income aggregated into five-year periods from Mitchell (2007). Independent variables explained in the text. Fixed effects added as indicated. Standard errors in parentheses (clustered by country): * significant at 10%, ** significant at 5%, *** significant at 1%.

 

 

 

       

 

 

Figure 1.8 State capacity and external war  

 

 

 

     

 

 

Figure 1.9 State capacity and executive constraints  

Table 3.2 Legal Capacity and Covariates: Simple Correlations (1) (2) (3) (4) (5) Government Anti-

Diversion Policy

Doing Business Registering Property

Obtaining Credit Contract Enforcement

Prevalence external war before 2000

1.294** (0.580)

0.427** (0.185)

0.278 (0.441)

0.355* (0.203)

0.749*** (0.230)

Average executive constraints before 2000

2.085*** (0.291)

0.535*** (0.084)

0.222* (0.122)

0.358*** (0.092)

0.287*** (0.108)

Average non-open executive recruitment before 2000

1.467*** (0.303)

0.235** (0.109)

0.229 (0.152)

- 0.082 (0.114)

0.202* (0.09)

Ethnic homogeneity

1.079*** (0.259)

0.241*** (0.073)

0.257*** (0.091)

0.286*** (0.089)

0.104 (0.096)

English Legal Origin - 0.157 (0.189)

0.148*** (0.050)

0.106* (0.064)

0.062 (0.054)

0.103* (0.054)

Scandinavian Legal Origin 0.706***

(0.204) 0.276*** (0.067)

0.327*** (0.079)

0.127 (0.081)

0.452*** (0.069)

German Legal Origin 0.627***

(0.185) 0.280*** (0.054)

0.244*** (0.079)

0.219*** (0.051)

0.365*** (0.063)

Socialist Legal Origin 0.013

(0.153) 0.062

(0.050) 0.155** (0.059)

- 0.007 (0.059)

0.265*** (0.053)

Observations 122 147 147 147 147 R-squared 0.623 0.552 0.293 0.414 0.442 Notes: Robust standard errors in parentheses: (* significant at 10%; ** significant at 5%; *** significant at 1%). French legal origin is the omitted category.  

Table 3.3 Legal Capacity and Covariates: Interaction Terms (1) (2) (3) (4) (5) Government Anti-

diversion Policy

Doing Business Registering Property

Obtaining Credit Contract Enforcement

Prevalence external war before 2000 1.369 (1.918)

0.708 (0.518)

1.529*** (0.549)

0.515 (0.659)

1.052* (0.561)

External war*High executive constraints dummy

0.146 (2.062)

- 0.299 (0.554)

- 1.535** (0.680)

- 0.203 (0.676)

- 0.320 (0.596)

Average non-open executive recruitment before 2000

0.547 (0.630)

-0.030 (0.199)

0.151 (0.257)

- 0.059 (0.210)

- 0.038 (0.229)

Average non-open executive recruitment*Low ex constraints

1.097* (0.657)

0.245 (0.204)

0.028 (0.259)

0.010 (0.212)

0.254 (0.233)

Average executive constraints before 2000

2.147*** (0.561)

0.632*** (0.127)

0.235*** (0.175)

0.613*** (0.147)

0.216 (0.150)

Ethnic homogeneity

1.150** (0.296)

0.256*** (0.070)

0.241*** (0.091)

0.310***

(0.089)

0.102

(0.096) English Legal Origin 0.137

(0.171) 0.145*** (0.050)

0.112* (0.065)

0.074 (0.053)

0.092 (0.057)

Scandinavian Legal Origin 0.931***

(0.286) 0.340*** (0.101)

0.347*** (0.091)

0.171 (0.107)

0.492*** (0.097)

German Legal Origin 0.714***

(0.286) 0.305*** (0.054)

0.265*** (0.078)

0.227*** (0.059)

0.388*** (0.067)

Socialist Legal Origin - 0.037

(0.171) 0.061

(0.052)

0.132** (0.062)

0.005 (0.062)

0.254*** (0.056)

Observations 122 147 147 147 147 R-squared 0.635 0.566 0.318 0.441 0.450 Notes: Robust standard errors in parentheses: (* significant at 10%; ** significant at 5%; *** significant at 1%). French legal origin is the omitted category. 

Table 3.4 Other Outcomes and Covariates: Simple Correlations (1) (2) (3) (4) (5) (6)

Private Credit to GDP

Corruption Perceptions

Private Investment Rate

Tax Revenue Share in GDP

Income Tax Share in Total Revenue

Formal Sector Share

Prevalence external war before 2000

2.490*** (0.571)

2.130*** (0.495)

0.132 (0.659)

3.227*** (1.160)

2.056* (1.100)

2.159*** (0.807)

Average executive constraints before 2000

1.729*** (0.331)

1.799*** (0.275)

0.906*** (0.260)

1.491*** (0.420)

1.690*** (0.421)

1.485*** (0.375)

Average non-open executive recruitment before 2000

1.099** (0.429)

0.870*** (0.310)

0.751** (0.356)

0.640 (0.388)

0.849* (0.473)

1.249*** (0.471)

Ethnic homogeneity 0.489 (0.301)

0.693*** (0.254)

0.991*** (0.216)

0.650** (0.311)

0.171 (0.283)

0. 549(0.353)

English Legal Origin 0.131 (0.218)

0.078 (0.156)

0.298* (0.161)

0.047 (0.178)

0.225 (0.183)

0.089 (0.233)

Scandinavian Legal Origin - 0.346 (0.41)

1.719*** (0.212)

0.154 (0.212)

1.966*** (0.348)

1.114*** (0.293)

0.499** (0.215)

German Legal Origin 1.618*** (0.407)

1.117*** (0.231)

0.272 (0.232)

0.677* (0.359)

1.273*** (0.219)

0.892** (0.221)

Socialist Legal Origin N/A -0.376*** (0.120)

0.268* (0.146)

-1.027*** (0.171)

- 0.308 (0.450)

- 0.172 (0.239)

Observations 96 147 154 104 104 109R-squared 0.633 0.643 0.332 0.630 0.554 0.375

Notes: Robust standard errors in parentheses: (* significant at 10%; ** significant at 5%; *** significant at 1%). French legal origin is the omitted category.

6. Microfoundations of fiscal capacity (sketch)

Outside option for taxpayers

earn untaxed income of in informal sectorwith expected punishment of getting caught

Maximally enforceable tax rate

if agents engage in rational tax arbitrage

≤ =()− +

()

everybody avoids or complies

Costs of investments in fiscal capacity

various compliance structures to enforce income taxcosts F(2 − 1) associated with raising 2also dependent on economic structure

50

Partial tax complianceLess than full compliance

assume individual expected cost of noncompliance (κ) indexed by κ uniformly distributed κ ∈ [0 1]

type κ pays income tax if()− + (κ)

()≥

and proportion paying taxes is 1− bκ ( ) with bκ defined by()− + ( bκ ( ))

()=

tax-avoidance share bκ goes up with , down with Alternative interpretation of fiscal capacity

tax revenue raised is now (1− bκ ( )) and = ̂ () = max

{(1− bκ ( )) }

51

Rational tax evasion

Tax payments can be avoided by costly effort

payment from group to government given by

h −

iwhere is amount of tax owed but not paid, due to evasiveeffort at increasing and convex cost

³

´Alternative interpretation of fiscal capacity

let ³

´ 0 and

³

´≥ 0 such that investment in

higher fiscal capacity makes private tax evasion more costlyBesley and Persson (2012) show how such a framework,extended to many goods and elastic tax bases, changesthe conventional analysis of optimal taxation

52

Tax morale

Cost of cheating reflects social norm

suppose individual cost is (κ bκ) with bκ 0“tax culture” affects fiscal capacitycf. research by political scientists

Tax morale may also reflect trust in how revenue is used

willingness to cheat may be lower if higheri.e., government spends on public goods

“fiscal contract” between government and citizens,public-goods provision exchanged for tax compliance?

interaction between social norms and tax enforcement,are laws and norms substitutes or complements?

53

7. Microfoundations of legal capacityTwo-factor, two-sector model

Simple model in Lewis tradition

traditional vs. modern sector — possible misallocations

Traditional sector

uses only labor and provides outside option with wage

Advanced sector

uses capital and labor in (CRS) Cobb-Douglas production

() = (1−)

a fraction, of group- members can operate this technologylet =

+2 be the economy-wide fraction of "entrepreneurs"

54

Factor ownership and markets

Capital

each group- citizen owns exogenous amount of capital at date (endogenized by private investment in text ch. 3)

economy-wide stock is =P

2

Labor

each citizen owns one unit, so that = 1

Traded in factor markets

capital market may be frictionless or not

55

Allocations in frictionless necoclassical economy

Work with intensive form of advanced production

capital intensity in advanced sector, =

with factor prices ( ) an entrepreneur maximizes

[() − − ]and optimal capital demand ̂ solves

= (̂)−1

some capital borrowed from non-entrepreneurs

56

Assume capital is scarce

We postulate that (other cases treated in text of Ch. 3)

= ()−1

where is return on some outside option (US T-bills)then all capital is employed in advanced sector(hypothetical) wage when all labor is employedin the advanced sector

(1− ) ()

57

Two possible cases in necoclassical economy

Some traditional production: (1− ) () traditional-sector productivity high enough to attract some labor,but each entrepreneur operates capital in advanced sector

real wage pinned down by

No traditional production: (1− ) () economy like a one-sector model with factor rewards ( )each entrepreneur runs one advanced sector-firm with share of the economy’s capital and 1 of its workforce

Consider institutionally constrained economies instead

and analogs of these two cases to studymicrofoundations of core model (this section)and genius of taxation (next section), respectively

58

Capital-market imperfections and legal capacity

Frictions when borrower can walk away from her loan

collateral required of borrowers own capitalmaximum capital used by entrepreneur in group at

≤ (1 + ) ,where, by risk neutrality, ∈ [0 ] is the probability thata court enforces collateral if called upon to do so

note legal protection assumed excludable across groups

Legal capacity

constraint on by naturally interpretable as # ofcourts and qualified judges, or existence and quality ofa centralized credit registry

59

Are capital constraints binding?

Compare frictionless and constrained demand for capital

entrepreneurs in group constrained if

(1 + )

and economy institutionally constrained for = if

1 + min{

} , ∈ {}

assume small enough that neither group can access thesame capital as in the frictionless neoclassical economy

Implication for labor demand?

if constraint binds, labor demand by entrepreneurs in solves

= (1− ) ((1 + ) ̃ ) ∈ {}

60

Group income and legal protection

How do incomes depend on ?

have to consider two cases above, i.e., when we havesome traditional-sector production or not, so wageis or isn’t pinned down by traditional-sector wage

Consider Case 1: Some traditional-sector production

per-capita income of group member is

( ) = [((1 + )

)

(̃ )1− − ̃ ] +

= [

(1− )](1−1)(1 + )

+

first term is quasi-rent on capital, which is constraineddue to this quasi-rent, ( ) is increasing

61

Microfoundations for core model

Suppose equal ownership and entrepreneurship across groups

then, we have written down a microeconomic foundationfor function

³

´ used in Sections 1-4 so all our analysis

there applies to this symmetric casenote that we can then write total credit/GDP ratio as

(1 + )

()1−

which is monotonically increasing in

also easy to see how we can deal with asymmetries

62

4. The Genius of TaxationConsider Case 2: No traditional-sector production

Now, we have a market-determined wage

(

) = (

X

((1 + ) 2)

(1− ) .

increasing in each — more labor demand raises the wage

Average income in group

( ) =

[((1+ ) )

(̃ )1−−( )̃ ]+( )

again made up by quasi-rents plus wage income

32

How is the group 0s income affected by ?

Take derivative of ( )

( )

= (1− ̃)

(

)

S 0 as ̃ R 1

cross effect is negative (positive), if group is a net importer

(exporter) of labor, so that

2 ̃

12

2 ̃

in this case, analysis in Sections 1-2 may no longer apply

33

Revisit the period-1 incumbent’s policy problem

Rewrite period policy payoff

+ (1− ) ( ) + and budget constraint

+ (

) = + +

+

2

where ( ) is national (non-resource) income per capita

( ) =

P

( )

2=X

2((1 + )

)

note that always increasing in even though may

not be — the cross-wage effect is a pecuniary externalitysocially efficient to follow rule of law =

=

34

Equilibrium legal protection

Still optimal to tax and transfer as before

imposing = we have the policy objective

+(1− ) ( )+2 (1− ) [+ ( )− −]Proposition 3.6 If , there is ̂ () ̂ () ̂ ()such that, for all ≥ ̂ (), all legal capacity is fully utilized, i.e., = = But if ̂ (), then = and = 0

for a rich incumbent group, with higher or the fiscalgains from higher may not be high enough to compensate forcut in quasi-rents from higher wages, if fiscal capacity low enough

cutoff value for lower when value of public goods is highsuch rent-seeking leads to production inefficiency — a violationof Diamond-Mirrlees, and a failure of Political Coase Theorem

35

Can this situation persist when endogenous?

Answer is yes

may still have a weak state — (2) 1 as in Section 2if low and high — caught in a ‘non-investment trap’

a richer group has lower incentives to invest in fiscal capacitythan a poor group since it pays higher share of taxes (cf. Ch 2)

Motives to invest in legal capacity

with rent-seeking these are generally weaker as well,if ̂ () 2 the marginal benefit of investment includes (2 0) ((2)− 1)2 (2 2) ((2)− 1)2

36

How can weak fiscal capacity shape income and growth?

Simple illustration

two states: (for Weak) and (for Strong)same initial legal capacity 1 =

1 = 1 and

=

but 1 b () 1 so at opposite sides of fiscal-capacitythreshold of Proposition 3.6

Compare incomes in period 1 and 2

period-1 difference is given by

1 − 1 = (1 1)− (1 0) 0 has lower income, as legal protection of inefficientperiod-2 difference (if incumbent persists)

2 − 2 = (2 2 )− (2 0) (1 1)− (1 0)income gap grows, since 2

2

37

Further perspective on income/state-capacity clusters

Recall positive correlations in Figures 1.3 and 3.1

results in Section 2 suggest they may reflect other factors causinglow state capacity and hence low (endogenous) income,or low (exogenous) income causing low state capacity

results here suggest low state capacity may cause low(endogenous) income via production inefficiencies

Ways out of inefficiencies in investment trap?

circumstances: higher or may make it too costly to staywith low fiscal capacity and inefficient production

institutions: higher or lower , may pull the economy outof fiscal-capacity investment trap

38

Relation to debate about financial development?

Work on political origins of financial (under)development

a ruling elite may hold off creating financial institutionsso as to create or preserve its own rents

but that work generally considers financial sector alonewithout attention to the tax-transfer system

results may implicitly assume weak fiscal capacity

Need to ask Political Coase Theorem question

why doesn’t government maximize the size of the pie andthen carry out the desired redistribution

stressed by Acemoglu (2003, 2005)here the friction is the absence of a credible mechanismfor transferring efficiency gains, beyond the institutionalcommitment entailed in

39

8. Microfoundations — Predation and Corruption

Alternative important source of misallocation

look at economic costs (and political benefits) of predationpredation could be private, due to lacking legal protectioncould also be public, as corrupt bureaucrats abuse their powerlegal capacity allows for legal protection against predation

Adapt earlier two-factor, advanced-traditional sector model

assume predation is only an issue in advanced sectorworks as a tax and may hinder structural transformationstudy simple symmetric case where every citizen holdscapital and each group has same share of entrepreneurs

Also study the working of a predatory state

governed by rent-seeking elite that monopolizes predation64

The mechanics of predation

A given group of predators

share of members ∈ [0 1] from both groupssuch that + = 1

corruption is special case, where = 1− = 1Predation as an informal tax

predators capture a share of output dependingon their effort which has convex cost ()

can target predation across groups, depending on howwell groups are protected, i.e., depending on ∈ [0 ]

simple formulation where predatory tax rate on a group

( ) = (1− )falls in legal protection and rises in predatory effort

65

Expected incomes and returns

Expected output in advanced sector

for group in period

[1− ( )]1−

we focus again on scarce capital

[1− ( )] ()−1 Again we have two cases

w or w/o traditional-sector production, depending on

Predatory returns

all predators act jointly to maximize profits from group

(

)

1− − ( )and split these according to ownership shares

66

Case 1: Some traditional production

Labor demand ̃ by sector entrepreneurs

solves[1− ( )] (1− ) (

̃ ) =

Optimal predation rate ̂

given by condition

(1− ) () (̃ )1− = (̂ )Better protection of group 0s property rights

higher has two beneficial allocation effectspredation effect: lower ̂ like decreasing production taxreallocation effect: pulls more labor into the advanced sector

67

Case 2: No traditional production

Labor demand

given by ̃ =1 and advanced-sector production

net of predation is

(1− ( )) ()

Optimal predation rate ̂

(at interior solution) is now given by

(1− ) () = (̂ )

now has only a predation effect, no reallocation effect

Consider Case 1 in the following analysis

68

Different types of income

Define net income to group

from production in the advanced sector

[1− (̂ )]̃( )where ̃( ) = ()

(̃ )1− is gross production

Net income to group from predation

[(̂ )̃(

) + (̂

)̃(

)− (̂)− (̂ )]

Total income for incumbent group

add these and income from the traditional sector

( ) = [1− (̂ )]̃() + (̂ )̃( )

−(Σ(̂ ) + (1− ̃)income by predation of own members (̂

)̃(

) nets out

69

Normative benchmark

Total (non-resource) national income/capita

add ( ) and

( ) similarly defined

( ) =

P∈{} ̃(

)− (̂ ) + (1− ̃ )

2

terms in (̂ ) are pure transfers, which drop out

Proposition 3.8 Income per capita maximized when = = ,i.e., full and equal legal protection, given available legal capacity

gross production,P

∈{} ̃( ) + (1− ̃ ) maximized by

minimizing implicit taxes on advanced-sector productiondeadweight loss from predation, −

P∈{}(̂

), minimized by

deterring predation as much as possible

70

Political equilibriumIncumbent faces similar problem as in Section 4

maximize expression in ( )

assume that — so that most predation rentscaptured by incumbent group

Predation on group

mostly redistributes within the group but generatessubstantial deadweight costs, cf. term −(̂) in ( )

may be optimal to set = Predation on group

generates substantial income for group , cf. term(̂

)̃(

) in

( )

may be optimal to set = 0

Formal argument as in Section 4 — omitted here71

A predatory stateChange assumption about who obtains the rents

so far rents accrue within each group, and incumbents acton behalf of all group members — presumes Coasian bargain

Alternative, more realistic, assumption

all predatory rents go to an “elite”, a share ” 1 of anyincumbent group, and bears all costs of predation

political turnover is between the two elite groupsadd third political institutions parameter to and viz.governance ∈ [0 1] a transaction cost imposed on eliteperhaps reflecting the independence of the judiciary

Realized corruption rents per capita in the eliteP∈{} (̂

)̃(

)− (̂ )

(1− )

72

Policy objective of incumbent elite

Assumed to act selfishly

but considers membership of the elite as well asmembership in group P

∈{} (̂

)̃(

)− (̂ )

(1− ) +

+ (1− ) [1− (̂ )]̃³

´+

as 1 elite puts greater weight on itself than on its groupi.e., agency conflict within groups and conflict between groups and determined as before

73

Legal protection revisited

For assignment of legal protection, we get

Proposition 3.9 Protection of property rights depends on governance.We have two thresholds (

) ( )

1. If ≥ ( ), = = 2. If ∈ [( ) ( )], ≥ ≥ 03. If ≤ ( ), = = 0

New results

bad governance: both groups may be denied legal protectionbasically, the elite has to be small enough

intermediate governance: result like in Genius of taxation

74

Back to investments in legal capacity

Define

Bad governance: ≤ ( )

Proposition 3.10 If Bad Governance holds, no incentive to investin legal capacity, and reduced motive to invest in fiscal capacity

Intuition is simple

under bad governance 2 = 2 = 0 so the prospective benefits of

investment (2) = 0; no future incumbent uses legal capacityby complementarity, fiscal-capacity investment is lower

Legal-capacity investment trap under bad governance

a new possibility, which may be jointly present with theearlier fiscal-capacity investment trap in weak states

75

Taking stock

Implied effects of predation and corruption

usual static distortions of production, but also twoadditional margins, where predation distorts

incentives for governments to provide legal protectionto citizens, given existing legal capacity

disincentives to build effective legal institutions

Normative implications for institutional reform

in core model, focus on cohesive institutions: high in this model, focus on good governance: high in practice, the two may be closely related as both callfor imposing constraints on the executive discretionof incumbent governments

76

The story so far

Determinants of state capacity

we have developed a framework to analyze investments in theextractive and productive parts of the statefiscal and legal capacity and

Up to now, politics has been kept in the background

the nature of political institutions (cohesiveness) , and therate of political turnover (instability)

still these parameters crucially shape the motivesfor building the state

they are (partly) endogenized in this and following lecture

2

C. Political violence

Motivation

Risk of external violence

by argument in part B, can promote state buildingcommon interest vs. redistributive (group) interest

Risk of internal political violence — civil war, repression?

not common interests — rather, extreme redistributive strugglemay entail very different incentives to invest in state

one way to endogenize political instability, with high relevancefor many developing countries

of course, better understanding of political violence is alsoimportant in and of itself

3

Facts about civil war — Figure 1.10

Unfortunately, this is a common phenomenon

civil war has plagued many nations in postwar periodprevalence over all nations and years since 1950 above 10%,cumulated death toll exceeds 15 million

Two big facts

prevalence varies greatly over years,peaks above 15% in early 1990s

prevalence varies greatly over countries,civil war and poverty (low GDP/capita) strongly correlated

two leading interpretations of 2nd fact: (i) reflects lowopportunity costs of fighting (Collier-Hoeffler, 2004),(ii) reflects low state capacity (Fearon-Laitin, 2003)

4

Figure 1.10 Prevalence of civil war and repression

Facts about government repression — Figure 1.10

One-sided political violence

many governments use violent means to raise their probabilityof staying in power without civil war breaking out

such repression shows up in violations of human rights:executions, political murders, imprisonments, brutality, ...

Prevalence?

by strict measure, purges, about 8% of country-years since 1950by wider measure, human-rights violations, about 32%, 1976-2006

Relation to civil war facts

purges have opposite trend to civil wars until early 1990speaks among higher-income countries than civil war

hint of substitutability between the two5

Existing research

Theory of civil conflict

quite a few papers on group conflict, often modeled as a contest,but small role for institutions, including state capacities,and public finance

Empirical work on civil war and repression

weak connections to theory, so difficult to interpret resultstakes income as given, though violence and income likely havesimilar determinants — e.g., parallel ‘resource curse’ literatures

separate literatures on civil war and repression, though bothreflect that institutions fail to resolve conflicts of interest

6

Need for theoretical work

Political violence, income, and state capacity?

political violence clusters with income — cf. Fig 1.10as well as state capacity — recall Fig 1.4

two-way relations amongst these outcomessame economic and political determinants may cause all three

Complex relations in the data calls for explicit theory

existing theory does not take institutions well into accountneed explicit theory to build bridge to empirical workexplicit theory may also help us understand relation betweencivil war and repression — and their relation to state capacity

7

Analytical approach

First step — this part.

study a simple model of political violence, extending modelin part B, but treat legal and fiscal capacity decisions as given

(long) detour confront conflict model’s implications with data

Second step — next part

reintroduce state-capacity investments in new frameworkreturn briefly to the dataput pieces together

8

Road map — Figure 1.11

1. The Core Model with Political Violence

2. Extensions: Polarization and Predation

3. From Theory to Evidence

4. Data and Empirical Results

9

Repression

Civil war

Common vs. redistributive interests

Cohesiveness of political institutions

Resource or (cash) aid independence

Income per capita

Figure 1.11 Scope of Chapter 4

1. The Core Model with Political Violence

Modifications of earlier setup

start out from exactly the same model of policy andstate-capacity investments as in part B

replace earlier exogenous transition of power byoutcome of (potential) conflict, triggered byinvestment in violence

but treat state capacity at = 1 2 as given

10

Violence and transitions of powerIncumbent and opposition can simultaneously invest in violence

period 1 opposition group 1 can mount insurgency witharmy ≤ paid within group, at marginal cost of funds

incumbent group 1 can invest in army ≤

paid out of the public purse, at marginal cost 1no conscription: each soldier just paid the period-1wage (1)

Probability of opposition takeover — conflict technology

( ; ξ) increasing in decreasing in

winner becomes next period’s incumbent, 2 ∈ {}loser becomes new opposition, 2 ∈ {}

Peaceful transitions

if nobody arms, transition probability is (0 0; ξ)

11

New timing

1. Start with state capacity 1 1 and incumbent group 1nature determines 1

2. 1 chooses a set of first-period policies {(1 ) (1 ) 1 1} andinvestments in period-2 state capacities 2 and 2

3. At the same time as 2, 1 and 1 simultaneously investin violence and

4. 1 remains in power with probability 1− ( ξ)nature determines 2

5. New incumbent 2 chooses current policy {(2 ) (2 ) 2 2}

we will study subgame perfect equilibrium in investments inviolence and policy at stages 3 and 5

in part D below, we will revisit state-capacity investments2 and 2 at stage 2 — for now, take those as given

12

Stage 5 — New incumbent 2 policymaker

Period 2 budget and policy instruments

exactly as before with budget constraint

+[(2) + (

2 )]

2= 2 +

2 + 2

2Equilibrium policies

same outcome as in part B, also in period 1

Indirect payoff and value functions

in earlier notation, we have

( ) = ( ) + (1− )() +

[ + ()− ( )−]

(2 2) =[ ( 2 2 0

)+

(1− ) ( 2 2 0 )]13

Stage 3 — Define the investment objectives

Expected utilities of groups 1 and 1

(1 1 11 )

+(1− ( ξ)) (2 2) + ( ξ) (2 2)and

(1 1 11 )− (1)

+( ξ) (2 2) + [1− ( ξ)] (2 2)now, 1 includes violence investment by 1 i.e., (1)whereas investment by 1 deducted from period-1 payoff

14

Stage 3 — PreliminariesProspective trade-off

when incumbent and opposition decide how much to invest, theyweigh investment cost against higher probability of policy control

First-order conditions

−(̂ ̂ ξ)h (2 2)− (2 2)

i− 1 (1) ≤ 0

and

(̂ ̂ ξ)

h (2 2)− (2 2)

i− (1) ≤ 0

common first term can be written

(2 2)− (2 2) = (1) 2 (1− 2)where

= + 2(2)−((2 2))

(1)is the wage-adjusted, expected redistributive pie in period 2

15

Restrictions on conflict technology

Make following assumption

Assumption 4.1 For all ∈h0

i, we have:

a. if ∈ (0 1), 0 0, 0 0b. −(00;ξ)

(00;ξ)≥ and

c. ≥ ≥

consistent with commonly used contest functions withcertain assumptions on parameters (in text ch. 4)

this assumption allows us to pin down the Nash equilibriumassociated with the two first-order conditions

16

Peaceful resolution of conflict game

Suppose 2 = 2 ≥ 2(1− )then, ( 2) = + 2(2)⇒ = 0i.e., no transfers will be paid at stage 4

Suppose 2 = ≥ 2(1− )then, Cohesiveness holds, and we have a common-interest statei.e., = 0 and any residual revenue again spent on public goods

in both cases expected payoff for is decreasing in ,whichever group gets into power, so = 0 =

Proposition 4.1 If ≥ 2(1− ) or if → 1no group invests in violence

there is always peace in common-interest states,or in states with high risk of external violence

17

Prospectively violent solution to conflict game

Proposition 4.2 If Assumption 4.1 holds, 2(1− ) and 1there are two thresholds ( ; ξ) and ( ; ξ)

( ; ξ) = − 1 (0 0; ξ) (1− )2(1− 2)

( ; ξ) =

(0 0; ξ) (1− )2(1− 2)such that:

1. if ≤ there is peace with b = b = 02. if ∈

³

´, there is repression with b b = 0

3. if ≥ there is civil conflict with b b 0 .Moreover, b and b, whenever positive, increase in

18

Anatomy of three regimes

1. Peace:

wages 1 high, non-tax income low, opposition’sshare high; too expensive to fight, or not enough to fight over

2. Repression: ∈h

i1 lower/ higher/ lower, so more redistribution at stake,and incumbent’s arming threshold lower, by Assumption 4.1b.

3. Civil war:

even more at stake, so both parties invest in violence, andnobody stops fighting as goes up, by Assumption 4.1c;in fact, always fights more intensively

19

Parallels with state-capacity determinants

Common-interest states

never have violence; recall they always invest in state capacity

Redistributive states

sometimes have violence; variables that trigger more violencealso generate low state capacity

high resource-rent or cash-aid share, high gives high low cohesiveness of political institutions, low gives low

low demand for public goods, low gives low

low income (given and ) low 1 gives high

Weak states

often have violence; recall that weak states — in countrieswith low and low — do not invest in the state at all

20

Role of political stability

State capacity framework in part B

there, stability treated as parametric — a high value of implies weak motives to invest in state capacity

Political violence framework

here, is endogenous

How do the forces highlighted in the two frameworks interact?

a natural question — posed and answered in part D

... but first we quickly sketch a couple of extensionsand make a (long) detour into the empirics of political violence

21

2. Extensions: Polarization and PredationPolarization

Ethnic polarization and conflict

quite a large literature: e.g., Esteban-Ray (2008) on theoryMontalvo-Reynal-Querol (2005) on empirics

Polarization as conflict about public goods

adopt same formulation as in ch. 2can write crucial benefit term in first-order conditions as

(2 2)− (2 2) = [ − ] + (1) 2 (1− 2)i.e., benefit from holding office is larger, as it goes beyondthe control of redistributive transfers

22

Greed vs. Grievance as drivers of civil war

Distinction made by Collier and Hoeffler (2004)

has become common (although murky) ideabasic model emphasizes greed, 2nd term in expression,while polarization extension adds grievance, 1st term

Possible further extension

let groups pick leaders that are more or less polarizingsay = 0 or = 1

if hardliner allows commitment to repression (butnot civil war), this could be preferred action

cf. leadership changes and Israel-Palestine conflict

23

Predation

Consider groups that are run by small predatory elites

same formulation as in text ch. 3,where =

= 0 and rents earned by the ruling elite

Π̂0 =

P∈{} [ (̂0 0) ̃ (0)− (̂0)]

Value difference in and out of future power

term in first-order conditions becomes

(2 2)− (2 2) = Π̂0 [1− ] + (0) 2 (1− 2)where

= + 2((0))−((2 2))

(0)

24

Differences with basic model

Proposition 4.1 no longer holds

high demand for public goods no longer sufficient to ruleout conflict, as new rents term in − remains

this adds to the incentives to invest in violence,formally, thresholds and are shifted down

Wage may be lower (0) not () in

cheaper invest in violence, unless pinned down by

Income lower than in non-predatory state

(0) ≤ (), so smaller incentive to fightbut likely dominated by the rent effect

Possible extension

violence within groups to control the elite, coups d’etat25

3. From Theory to EvidencePreliminaries — observability

Back to basic model in Section 1

which parts of and observed for certain country, at time ?can measure, or find decent proxies for and but genuinely hard to measure (0 0; ξ) and (0 0; ξ)and cost parameters and

Unobserved randomness in determinants of violence

treat ( ) as given and write random variable − as

− =− −

where is a constant and an "error term" with c.d.f. ()

26

Preliminaries — observability (continued)

Similarly, we can write

− =− −

where error has c.d.f. ()

Incidence of violence ?

we do not directly observe and

but do observe if there is civil war, or repression, in and may observe = (if interpret as external conflict)

27

Conditional probability of civil war

By Proposition 4.2, civil war in some country at date if

− ≥ 0 ⇔ ≤ −

given the information available to us, the conditionalprobability — i.e., the likelihood — to observe this event is

( − )Prediction

higher or lower raises probability of observing civil warbut, by Proposition 4.1, no effect if close to 1 or ≥ 2(1− )can test this with time-varying measures of and

28

Conditional probability of other violence statesConditional probability of observing peace

but not civil war, at date

1− ( − )down with up with unless → 1 or ≥ 2(1− )

Conditional probability of observing repression

( − )− ( − )effects of shocks, now depend on densities

Alternative way of stating model predictions

higher or lower raise the probabilityof observing some form of political violence

states of peace, repression, and civil war ordered in calls for estimating ordered logit

29

Identification — what variation to use in data?

How clean inference from unobserved determinants?

using cross-sectional variation risks confounding variablesof interest, like and with nuisance parameters, like ξ

instead estimate panel regressions with fixed country effectsequivalent to estimating, e.g., for civil war

( − )−{( − )}Heterogeneity in incidence of violence over time

now driven by time variation in and add fixed year effects to allow for world-wide shocks,non-parametric trends in violence — recall Figure 1.10exploit only country-specific time variation in and

30

Identification — further issues

How take fact that predictions conditional on into account?

let Θ = 1 be cohesive political institutions ( ≥ 2(1− ))and Θ = 0 non-cohesive political institutions

represent index function, in country period as

− = (Θ) + (Θ) + (Θ) ewhere e are time-varying regressors proxying for and according to the theory (0) 0, while (1) = 0

Still need exogenous variation in ewithin-country variation no panacea, unless we can alsocredibly argue that variation in e is exogenous to violence

31

4. Data and Empirical ResultsPolitical violence data

Civil war

binary indicator from Uppsala/PRIO data set, 1950-2005alternative: COW data, but shorter series (end in 1997)

Repression

purges variable from Banks (2005) data set, 1950-2005alternative: PTS data, but shorter series (begin in 1976) anddoubts about US State Department’s coding during cold war

Construct ordered dependent variable

combine repression and civil war measures as followspeace = 0, repression/but not civil war = 1, civil war = 2

32

Political institutions data

Main indicator of weak and strong institutions

indicator for highest score (7 on 1-7 scale) forExecutive Constraints variable in the Polity IV data set

corresponds best to in the theoryset indicator for the whole panel Θ = 1 only if(i) positive prevalence pre-1950 and (ii) sample prevalence 0.6

conservative criterion: selects less than 20% of sample

Alternative measure

indicator based on parliamentary democracy takenfrom Polity IV and Persson-Tabellini data sets

analogous (i)-(ii) definition for Θ = 1

33

Two forms of shocks to eNatural disasters — negative shocks to or positive shocks to

from EM-DAT data set, 1950-2005indicator for having at least one out of four disaster events:heat-wave, flood, slide, or tidal wave — associated with2.5% lower level of GDP/capita

Cold-war, security-council membership — positive shocks to

agnostic about effect of membership, in generalbut insist members likely to get more aid due to geopoliticalimportance during cold war (Kuziemko—Werker 2006, for US)

34

An initial observation

By Prop 4.1 — no violence when ≥ 2(1− ) ?32 countries in our panel classified as Θ = 1

only 8 (25%) of those has some year with eithercivil war or repression from 1950 to 2005

125 countries classified as Θ = 0

97 (80%) of those has some year with eithercivil war or repression in same period

informative, but hazardous to draw causal inference fromsuch cross-sectional variation

35

Basic results — Table 4.4

Estimate ordered logits implied by the theory

columns (1)-(3)

fixed-effect ordered logits — implement as suggested byFerrrer-i-Carbonell and Frijters (2004)full sample, and interaction effects with indicators for cohesiveinstitutions and measured by constraints on executiveparliamentary democracy, respectively

Results in line with theoretical predictions

only significant effects on violence with expected sign in sampleswith low executive constraints or non-parliamentary democracies

statistically robust: results hold up when bootstrap standard errorsin column (8)

36

Table 4.4 Basic Results (1) (2) (3) (4) (5) (6) (7) (8)

Dependent variable Ordered variable

Ordered variable

Ordered variable

Political violence

Political violence

Civil war Civil war Ordered variable

Natural Disaster 0.263** (0.107)

0.317*** (0.110)

0.299*** (0.111)

0.278** (0.109)

0.327*** (0.112)

0.370** (0.152)

0.431*** (0.155)

0.263** (0.111)

Security council member

−1.048*** (0.399)

−1.194*** (0.417)

−1.382*** (0.456)

−1.110*** (0.412)

−1.269*** (0.430)

−1.360** (0.545)

−1.383** (0.547)

−1.048** (0.413)

Security council member in cold war

1.275*** (0.439)

1.461*** (0.458)

1.657*** (0.495)

1.267*** (0.453)

1.465*** (0.472)

1.074* (0.633)

1.105* (0.635)

1.275** (0.504)

Natural disaster x strong institutions

−0.701* (0.374)

−0.333 (0.318)

−0.618* (0.376)

−1.233** (0.595)

Security council member x strong institutions

1.975* (1.173)

2.940*** (1.123)

2.186* (1.178)

Security council member in cold war x strong institutions

−2.577* (1.375)

−3.379*** (1.247)

−2.746** (1.381)

Strong institutions measure

High executive

constraints 1950-2005

Parliamentary Democracy 1950-2005

High executive

constraints 1950-2005

High

executive constraints 1950-2005

Estimation method FE Ordered Logit

FE Ordered Logit

FE Ordered Logit

FE Logit FE Logit FE Logit FE Logit FE Ordered Logit

Significance of interactions (p-value)

0.61 0.49 0.66 0.17

Observations 4251 4251 4251 4251 4251 2061 2061 4251Number of Countries 97 97 97 97 97 49 49 97 Notes: The time period covered is 1950 to 2006. For definitions of variables refer to the text. Standard errors are in parentheses: * significant at 10%; ** significant at 5%; *** significant at 1%. Robust standard errors are in columns (1)-(7) with bootstrapped standard errors in column (8). The p-value refers to the significance of a test of the hypothesis that coeff_[natural disaster x strong institutions] = -coeff_[natural distaster] & coeff_[security council member x strong institutions]=-coeff_[security council member] & coeff_[security council member in the cold war x strong institutions] = - coeff_[security council member in the cold war], where coeff_ is the estimated coefficient on the variable in question. The reduced sample size in columns (6) and (7) is due to all countries which never had a civil war during this period being dropped.

Look at alternative violence margins — Table 4.4

Estimate conditional logits implied by the theory — columns (4)-(7)

conditional (fixed effect) logit for two margins where theoryhas bite: peace vs. violence, and non-civil war vs. civil war

full sample and interaction effects with high executive constraints

Results again, basically, in line with theoretical predictions

only see significant effects on both forms of violencewith low executive constraints

37

Table 4.4 Basic econometric results

(1) (2) (3) (4) (5) (6) (7) (8) Dependent variable Ordered

variable Ordered variable

Ordered variable

Political violence

Political violence

Civil war Civil war Ordered variable

Natural Disaster 0.263** (0.107)

0.317*** (0.110)

0.299*** (0.111)

0.278** (0.109)

0.327*** (0.112)

0.370** (0.152)

0.431*** (0.155)

0.263** (0.111)

Security council member

−1.048*** (0.399)

−1.194*** (0.417)

−1.382*** (0.456)

−1.110*** (0.412)

−1.269*** (0.430)

−1.360** (0.545)

−1.383** (0.547)

−1.048** (0.413)

Security council member in cold war

1.275*** (0.439)

1.461*** (0.458)

1.657*** (0.495)

1.267*** (0.453)

1.465*** (0.472)

1.074* (0.633)

1.105* (0.635)

1.275** (0.504)

Natural disaster x strong institutions

−0.701* (0.374)

−0.333 (0.318)

−0.618* (0.376)

−1.233** (0.595)

Security council member x strong institutions

1.975* (1.173)

2.940*** (1.123)

2.186* (1.178)

Security council member in cold war x strong institutions

−2.577* (1.375)

−3.379*** (1.247)

−2.746** (1.381)

Strong institutions measure

High

executive constraints 1950-2005

Parliamentary

Democracy 1950-2005

High

executive constraints 1950-2005

High

executive constraints 1950-2005

Estimation method FE Ordered Logit

FE Ordered Logit

FE Ordered Logit

FE Logit FE Logit FE Logit FE Logit FE Ordered Logit

Significance of interactions (p-value)

0.61 0.49 0.66 0.17

Observations 4251 4251 4251 4251 4251 2061 2061 4251 Number of Countries 97 97 97 97 97 49 49 97 Notes: The time period covered is 1950 to 2006. For definitions of variables refer to the text. Standard errors are in parentheses: * significant at 10%; ** significant at 5%; *** significant at 1%. Robust standard errors are in columns (1)-(7) with bootstrapped standard errors in column (8). The p-value refers to the significance of a test of the hypothesis that coeff_[natural disaster x strong institutions] = -coeff_[natural distaster] & coeff_[security council member x strong institutions]=-coeff_[security council member] & coeff_[security council member in the cold war x strong institutions] = - coeff_[security council member in the cold war], where coeff_ is the estimated coefficient on the variable in question. The reduced sample size in columns (6) and (7) is due to all countries which never had a civil war during this period being dropped.

Inspecting the mechanism — Table 4.5

Go further than the reduced forms in earlier tables?

columns (1)-(4)fixed-effect OLS (linear probability model); useful check onrobustness of cols (4)-(7) in earlier table, and results easier tointerpret in quantitative terms

columns (5)-(6)"first stage" effects on total aid (OECD data) and GDP percapita (PWTdata) of natural disasters andUNSecurity Council

columns 7-8"second stage" of fixed-effects IV; at best a diagnostic, as theexclusion restrictions not necessarily satisfied

Mechanism?

appears to run mainly through higher aid flows38

Table 4.5 Extended econometric results (1) (2) (3) (4) (5) (6) (7) (8)

Dependent Variable Political violence

Political Violence

Civil War Civil War Log GDP per capita

Log Aid Disbursements

Political Violence

Civil War

Natural Disaster 0.024* (0.013)

0.029* (0.017)

0.029** (0.013)

0.043*** (0.016)

−0.005 (0.003)

0.105** (0.043)

Security council member

−0.066** (0.027)

−0.092*** (0.029)

−0.051** (0.023)

−0.053** (0.023)

0.009 (0.008)

−0.269*** (0.092)

Security council member in cold war

0.090** (0.040)

0.129*** (0.045)

0.034 (0.029)

0.036 (0.029)

−0.004 (0.010)

0.434*** (0.113)

Natural disaster x strong institutions

−0.024 (0.037)

−0.079*** (0.024)

Security council member x strong institutions

0.148***(0.054)

Security council member in cold war x strong institutions

−0.205*** (0.068)

2-year lagged log GDP per capita

0.905***(0.013)

Log GDP per capita

0.062 (0.039)

0.046 (0.040)

Log Aid Disbursements

0.191*** (0.046)

0.161*** (0.050)

Observations 5880 5880 5880 5880 6300 5067 3914 3914Number of Countries 158 158 158 158 178 150 R-squared 0.030 0.031 0.056 0.059 0.914 0.136

Notes: The time period covered is 1950 to 2006. For definitions of variables refer to the text. Robust standard errors adjusted for clustering by country in parentheses (* significant at 10%; ** significant at 5%; *** significant at 1%). The specification in columns (1) through (6) is OLS. The results in columns (7) and (8) are IV specifications in which natural disaster, security-council member and security-council member in the cold war and 2-year lagged log income per capita are used as instruments for Log GDP per capita and Log Aid Disbursements.

D. State Spaces — Figure 1.12Preliminaries — Endogenous turnover

Return to state-capacity investments

political-violence model endogenizes political turnoverstructure of model gives convenient recursive structure, whereviolence shapes incentives only via political instability

Equilibrium turnover

define the endogenous takeover probability as

Γ ( ) =

⎧⎪⎪⎨⎪⎪⎩³̂ ̂

´ ( )

³0 ̂

´(; ) ≥ ( 1 )

(0 0 ) ( 1 ) ≥ assume parameter raises (lowers) incumbent’s (opposition’s)marginal return to fighting − (0 0 ) 0 ( (0 0 ) 0)

39

Legal capacity Fiscal capacity

Common vs. redistributive interests

Cohesiveness of political institutions

Resource or (cash) aid independence

Income per capita

Figure 1.12 Scope of Chapter 5

Repression Civil war

Preliminaries — Comparative statics of

Proposition 5.1 The probability that the incumbent loses officevaries with ( ) as follows:

1. higher reduces the probability that the incumbent loses office,when there is either repression or civil war.

2. higher reduces the probability that the incumbent loses office,when there is civil war.

3. higher reduces the probability that the incumbent loses office,when there is either repression or civil war.

these comparative statics follow from Assumption 4.1

40

Implications for investment

State capacity problem is recursive

Euler equations for legal and fiscal capacity become

(2)[1 + ((2; )− 1)2] 0 1L (2 − 1)c.s. 2 − 1 > 0

(2)[((2; )− 1] 0 1F (2 − 1)c.s. 2 − 1 > 0

where

(2; ) = + (1− )(2|; )is