Macroeconomic and Distributional Implications of Shocks and Policies

Affecting the Provision of Environmental Goods and Services: A Modeling

Approach.

Leonardo Garrido

Senior Research Consultant for the World Bank Group

Prepared for Workshop on Green Growth Modeling.

Advanced Systems Analysis (ASA) and Risk and Resilience (RISK),

International Institute for Applied System Analysis (IIASA).

Schloss Laxemburg, Austria. July 18th, 2017

Agenda

Conceptualization

Ecological Models

Macroeconomic Models

Micro, Distributional Impacts Models

Demographics

Model Examples

Thinking About Models for Linking

Natural Capital and the Socio-Economic

Natural Capital Model

Macroeconomic Model

Distributional Impacts

Demographics Dynamics

NC-MA MI-MA

Policy, Intervention

DE-MADE-MI

Conceptualization (1/6)

Linking Natural Capital and the Macro

Economy

NatCap Macro : Impacts of Natural Capital Depletion,

Pollution, Emissions on Factor Productivity, Input and

factor supply (e.g. Jorgeson IGEM)

Macro NatCap : Impacts of Economic Activity on

Pollution, Emissions, Natural Capital Depletion

NatCap Macro: Integrated Assessment Methods (e.g.

Nordhaus DICE, RICE Models; World Bank’s ENVISAGE)

Conceptualization (1/6)

Feed Backs: Economy and Climate

EmissionsAtmospheric

Concentration

Climate

TemperatureDamage

FunctionProductivity

Economy

Carbon

Cycle

Policy

Conceptualization (1/6)

High Schematic Diagram for DICE Model

Social UtilityDiscounted

Utility Rate

PopulationNet growth rate

population

Index of

ProductivityGrowth rate of

Productivity

Gross Output

Neoclassical Production

Function

Consumption per

Capita

Capital StockGross Capital

FormationDepreciation

rate

Labor

Consumption

Saving

Ratio

Climate

Damage

Abatement

Costs

Global Temperature

Change

Fossil Fuels

Atmosphere

Land Biota

Soil

Ocean Surface

Deep Ocean

Conceptualization (1/6)

Emissions Control

Ecological Model

Bio-economic modeling framework

Vibrant Oceans Initiative

https://www.bloomberg.org/program/environme

nt/vibrant-oceans/

https://www.rare.org/our-work#.WWnv-0u5uUc

Kobe Process. Philippines, 2014

Ecological Models (2/6)

Ecological Model Carbon

Coastal Vulnerability

Crop Pollination

Fisheries

Habitat Quality

Habitat Risk Assessment

Malaria

Marine Fish Aquaculture

Marine water quality

Offshore wind energy

Water purification

https://www.naturalcapitalproject.org/invest/

Ecological Models (2/6)

Macro Model

Simple Complex

Orthodox • SAM, I-O

Multiplier

• Neoclassical

Growth Model

• CGE Model (Static,

Dynamic)

• Large-scale macro-

econometric model

Heterodox • Simple SysDyn

• Simple ABM

• System Dynamics

• Agent-based,

computational

models

Macroeconomic Models (3/6)

Continued Reliance on Social Accounting

Matrices for Registering Economic

Transactions Across Sectors, Institutions and

Factors of Production

At one point in time (Does not pick up changing technologies, supply /

demand relationships)

Not all factors of production included (Labor, Capital, Sometimes Land)

Retains all faults and weaknesses attributed to national accounts to deal with

natural phenomena

Macroeconomic Models (3/6)

Households

Factor

Markets

Productive

Activities

Factor earnings

(Value Added)

Sales Income

Commodity

Markets

Intermediate

Demand

Rest of the

World

Exports Imports

Private

Consumption

Investment

Investment

Demand

Government

Government

Consumption

Household Savings

Remittances Grants, Loans FDI

Direct Taxes

Subsidies

Fiscal BalanceIndirect Taxes

Circular Flow Diagram

Macroeconomic Models (3/6)

N = Total population

z = Poverty line ($$)

yi = Income of HH or individual

H = Number of poor (number of people with yi<z)

a=0, FTG = poverty headcount: Fraction

of population in poverty

z

a=1, FTG = poverty gap: Area up to z

divided by total area

a=2, FTG = poverty gap, squared: Higher

weight for those farther below poverty line

Micro, Distributional Impacts Models (4/6)

Welfare Considerations as Ultimate

Goals for Modeling Process

Poverty and Distributional Impacts.

Top-Down Approach: ADePT SimulationMicro data

(e.g. HH Survey)

Price Data

• Labor Force Status Module

• Earnings Equation

• Remittances

Estimate

• Income and Consumption

• Individuals and Households

Adjust

Macro Projections

Population Growth

• Change in Labor Force Status

• Change in Real Earnings

• Change in Remittances

Predict

Rule:

Best fit to micro data

Rule: Replicate macro-

proportional changes at

micro level

• Income and Consumption distributions

• Poverty, inequality measures

Results

http://www.ifc.org/wps/w

cm/connect/Topics_Ext_C

ontent/IFC_External_Corp

orate_Site/Development+I

mpact

Micro, Distributional Impacts Models (4/6)

An ongoing Demographic, affects the age structure of population and

dependency ratios has important implications about the supply of labor and

employment dynamics:

- Demographic Dividend in Developing Countries

- Post Transition - Aging in Advanced Economies

Demographics (Example from the Philippines, based on UN Estimates)

Demographics (5/6)

A Model for Socio-Economic Analysis of

Policies for Sustainability of Fisheries in

the Philippines

Ecological Model:Vibrant Oceans

Initiative

Sustainable Fisheries

Group, UC-SB - RARE

Exogenous Demographics:United Nations, DESA

Population Projections

Medium Variant Scenario

Macro Model:Unconstrained SAM

Multiplier

RARE Organization

(VENSIM)

Micro Model:Distributional Impacts

RARE Organization

(ADePT Simulation)

Policies for Sustainability, VOI:

Financing, CRM

RARE Organization

(VENSIM)

User EnginePolicy Analysis

RARE Organization

(FORIO)

Model Example 1 (6/6)

The Model in VENSIM

Model Example 1 (6/6)

SAM SAM Coefficients

Ident Matrix

Identity minus

SAM Coeff

Ident minus Coeff

Inverse

Matrix number rows

Zt = (I – M)-1*E0

VOI Revenues

Reg Artisanal VOI Revenues Reg

Commercial

Reduce Revenues

Reg Artisanal Reduce Revenues

Reg Commercial

Manage Revenues

Reg Artisanal Manage Revenues

Reg Commercial

BAU Revenues

Reg ArtisanalBAU Revenues Reg

Commercial

Revenues

Artisanal Reg by

Scenario

Revenues

Commercial Reg by

Scenario

Revenues Reg by

Scenario and fleet

Revenues by Scenario

Revenues by

Scenario and Fleet

Revenues by

Scenario and Region

Change Revenues by Scen

and Fleet Relative to Initial in

Percent

Chg Revenues by

Scen Relative to

Initial in Percent

Initial Revenues by

Scenario and Fleet

Initial Revenues

by Scenario

Share Revenues by

Scenario and Fleet

Initial Share Revenues

by Scenario and Fleet

Switch all fleet

is 0 Artisanal

Only is 1

Commercial

only is 2

Share Regions in Revenues

by Scenario and Fleet

Efficiency Factor that

increases Revenue from

Supply Chain Intervention

Efficiency Factor Time

Series Scenario VOI by

fleet from data

Built in

Efficiency Factor

SwitchEfficiency

Factor equals1 if uses

preloadeddata

Initial Efficiency

FactorTarget Efficiency

Factor

Slope of changes in

efficiency per period

Target Efficiency factor

from Decreased Spoilage

Target Efficiency factor

from Quality Related

Price Increase

Target Efficiency Factor

that increase Net Revenue

from lower costs

<Time>

Period when

Efficiency factors

kick in

Price elasticity of

demand

Price index in

fisheriesChange in Price

Index Fisheries

Initial Price Index

of Fisheries

Max annual increase

in prices of fisheries

Smoothed price

index fisheries

<Smoothed price

index fisheries>

<Aggregate employment

Fish by Fleet>

Revenues per worker

by scen and fleet

Revenues per worker

fisheries by scen

Value Added by

Scenario

Change in Value Added by

Scen relative to previous

period

CummulativeStock of Fish

extractedChange quantity fish

extracted Relative to previous

by period in Millions

Quantity Fish by

scenario

<VOI Revenues Reg

Commercial>

<Reduce Revenues

Reg Commercial>

<Manage Revenues

Reg Commercial>

<BAU Revenues

Reg Commercial>

<Quantity Fish by

scenario>

Ratio Qcheck

change quantity fish

relative to initial

<Quantity Fish by

scenario>

Initial Value

Added by Scen

Change in Value Added

by Scen relative to Initial

Quantity Fish in

Billions LCU

Revenues by Scen in

Billion LCU

Non food CPI

weight

Food non fish CPIweight

Fish CPI weight

Food Price Index

Non food CPI

CPI

Revenues by Scen and

Fleet in Billions LCU

Initial Real GDP

by Sector

Shock Ocean Fishing and

Financing by Year and

Scenario

Matrix Multiplic is

MInverse x Shock by

Scenario and year

<Ident minus

Coeff Inverse>

Summary Effects

Output

Summary Effect

Changes GDP

Summary Effects

Income

Summary Effects

Government

Summary Effects

Savings Investment

Summary Effects

Rest of the World

Target GDP by sector

year and scenario

Shares sectors in

Real GDP

Speed of adjustment of macro

variables to shocks

<Time>

periodvar

Summary Effect Chg

GDP period panel

Target GDP by

Scenario

Initial Real GDP by

sector year and scenario

Real GDP by sector

year periodset and

scenario

<Time>

<INITIAL TIME>

<periodvar>Variation Real GDP

Variation real

GDP total

Real GDP by

Comm and Scen

Variation real GDP AggInitial Real GDP by

Sector in LCU

Financing in

support of

Sustainability

in USD

Exchange rate

2017 Peso per

USDInflation 2017

versus 2000

Percent of

Expenditure

Allocated by

Year

Government

Financing in Support

of Sustainab in LCU

of 2000 by year

PercentFinancing

fromForeignDonors

ForeignFinancing in

LCU of 2000 byyear

percent allocation

gov financing to

commodities

percent allocation

foreign financing

by commodities

Summ Effect Chg

GDP by Comm

Size of Shock to GDP

GDP path in Billions

LCU of 2017

Fraction

Govt

financing via

taxes Rest is

Loans

Interest rate in foreign

loans to government

Grace period in foreign

loans to government

Initial Total

Real GDP

percent reduction

comm consumption

from taxes

Government Shock for

financing Sustainability

Policies

Effect on

Governement

finances

<Summary

Effects

Government>

Effect on

Governemnt

Finances as

percent of GDP

<Real GDP by

Comm and Scen>

Real GDP Fisheries by

Scenario in LCU of 2017

Change real GDP

relative to initial

<Initial Real GDP

by Sector>

Total Government

Shock

<Fraction of laid off in

Commercial fisheries tha is

employed in other primary activ>

<Initial Average

Product of Labor>

Additional Output in Agric from laid

off Comm fishermen

<Empl by fleet and sector

keeping constant Empl in

ArtisFisheries at Init Level>

Change in Employment

Comm fish

Allocate Additional Output in

RestAg as Demand only in RestAg

Fraction of Avg Product of

Labor that is attained by

displaced fishermen

Total Financing

LCU by year

<Inflation 2017

versus 2000>

Real GDP by Comm and

Scen in Billion LCU of

2017

Summary effect on

Income LCU of 2017

<Inflation 2017

versus 2000>

<Change in Value

Added by Scen

relative to previous

period>

<Value Added by

Scenario>

Change Real GDP

Fisheries relative to

initial

Initial Real GDP

fisheries by Scen in

LCU of 2017Cumm Effect Gov

Finances MM

LCU of 2017Effect Gov Finances

MM LCU of 2017

Cumm Effect Gov

Finances as Ratio of

Initial GDP

Total Financing

LCU of 2017

by year

Outputs from Socio Economic Model

BaU: Do Nothing, Over-fishing

VOI Pessimistic: Sustainable fishing, no VOI Financing, no Efficiencies

VOI Optimistic, Sustainable fishing, VOI financing, Efficiencies

Model Example 1 (6/6)

Distributional Impacts Model

UN Population Projections,

through 2050, by Age Groups

and Gender, Rural and Urban

Household Survey Data, 2011,

with Population Weights, by

Age, Gender, Area, and Regions

Recalibrated population by

categories, Entropy Method:

Pop. by years, areas, age groups

Welfare Measures:

Consumption,

Income, Distribution

World Bank, WDI; PHL Labor

Force Survey Data: Labor Force,

Participation Rate, Employment

Macro Model: Multiplier

Analysis Results (Value Added,

Consumption…)

Household Survey: Simulated Changes

in Mean and Distribution (Poverty and

Inequality)

Model Example 1 (6/6)

Some Outputs from ADePT Simulation

Model Example 1 (6/6)

User Friendly Interface for Policy

Analysis

Model Example 1 (6/6)

https://forio.com/epicenter/builder/#/agagern/neda-rare-cba

A Model for Ex-Ante Analysis of Shocks

and Development Policies in Ethiopia

CGE Macro Model, Static. GAMS

• Social Accounting

Matrix

• Initial Conditions

• Model parameters

Calibration

System Dynamics Macro-Micro Model. VENSIM

• Disaggregated sectors

• Market disequilibrium:

Delays, Sticky prices,

Stock accumulation,

non-linearity

• Different closures

Extension, Changes

in model structure

Household Level Data

Household Heterogeneity in

Utility Maximization Problem

Policies and Shocks

User Engine

FORIO

Interface

Model Example 2 (6/6)

Main Features of Model

Static CGE Model

HH Utility Maximization (Stone Geary)

Firms take demand as given, Maximize

Profit

12 activities, commodities

Open Economy, Armington, CET

Government collects taxes, spends

Different closures: Savings – Investment;

RER – Foreign Savings

Model calibrated to SAM values

Micro-Macro linkages from SAM, HH data

(Replicate Consumption by Commodity,

Income)

System Dynamics Model (Changes CGE)

Sticky Prices, elasticities

Stock accumulation (inventories by

tradable commodities)

Full developed sectors

(Exogenous) Changes in Price of X, M

Debt accumulation

Additional closure to Model: Restricted vs

unrestricted Government Expenditure

Dynamic calibration to historical period

Top-Down, Bottom-Up model results

emerge from feed-back structures,

calibration rules, closures and scenarios

Model Example 2 (6/6)

Using Ethiopia Model for ex-ante

assessment of policies as shocks

For assessing expected impacts of policies included in Ethiopia Long Term Development Plan

Model versatile to “connect” to fully fledged sectors (Including agriculture), modules linked to provision of environmental goods and services

Convenience of carrying on policy, shock analysis with Tod-Down, Bottom-Up structures, under a consistent Macro-Micro framework and given “rules” or “closures”

Model amenable to compute several indicators included among Sustainable Development Goals

Model Example 2 (6/6)

Model Example 2 (6/6)

Total Real GDP Expenditures

1000

500

0

2005 2010 2015 2020 2025 2030

Time (year)

Historic RGDPT : Baseline

"Total Real GDP (RGDPT)" : Baseline

"Total Real GDP (RGDPT)" : Scenario1

Aggregate Demand Components: Model

1

.5

0

2005 2010 2015 2020 2025 2030

Time (year)

Total Priv Cons to Agg Dem Ratio : boostTotal Gov Exp to Agg Dem Ratio : boostAgg Invest to Agg Dem Ratio : boostTotal Export to Agg Dem Ratio : boost

Log Total Real Exports: Hist vs Model

5

4.25

3.5

2.75

2

2005 2010 2015 2020 2025 2030

Time (year)

Log Hist Tot Exports : Baseline

Log Tot Exports : Baseline

Log Tot Exports : Scenario1

Quantity of Labor by Education Type

1

.9

.8

2005 2010 2015 2020 2025 2030

Time (year)

Share of QLabor by Educ Type[labpri] : boost

Share of QLabor by Educ Type[labsec] : boost

Share of QLabor by Educ Type[labter] : boost

Headcount Poverty from Consumption (Based on 1.9$/day Poverty Line)

40

30

20

10

0

2005 2010 2015 2020 2025 2030

Time (year)Headcount poverty rate based on HH Consumption : boost"Historic Poverty headcount ratio (Equivalent to $1.90 a day (2005 PPP)" : boost

Some Preliminary Charts

Thank You!