Global Trade Analysis: introduction to the GTAP modelling framework Frank van Tongeren Trade and development division, LEI.

Global Trade Analysis: introduction to the GTAP modelling framework

Frank van TongerenTrade and development division, LEI

Outline

Introduction: motivation for GTAP Database Model Some economics GTAP organisation

Motivation for GTAP

Increasing demand for quantitative analysis of global trade issues

E.g., WTO Doha round, Kyoto agreement, EU integration, Regional integration

Historically analysis has been done “in-house” in a few agencies: USDA, FAO, World Bank

Methodology is often poorly documented. Rarely subjected to peer review. Data base is treated as a strategic asset

and therefore not shared with others. Projects rarely survive change of

administration. Unfavorable findings result in termination.

Limitations of Agency-based Projects

Limitations of Projects Based in Academia

Limited financial support. Little incentive to invest in data base. Tendency to keep model and data until

it has outlived its usefulness. Limited contact with decision makers

who are the ultimate “clientele” for much of this work.

The GTAP Approach

Publicly funded project, based in academia.

Data base and standard model: fully documented publicly available at modest cost easy to use with regular courses accessible to non-modelers.

Key Resources: Data Base

Philosophy: Find the best person in the world to do the job GTAP as an assembler of information

Reconciled bilateral trade data and shipping margins: USDA

Protection data: WTO/WBank/UNCTAD, AMAD National data bases: national collaborators

e.g Version 5 EU-15 Input-Output tables, LEI with funding from DG-ENVIRON

Production cycle database: 18-24 Months

GTAP v6 commoditiesPrimary agriculture Paddy rice

WheatCereal grains necVegetables, fruit, nutsOil seedsSugar cane, sugar beetPlant-based fibersCrops necCattle,sheep,goats,horsesAnimal products necRaw milkWool, silk-worm cocoons

Natural resource based activitiesForestryFishingCoalOilGasMinerals nec

Processing agriculture and foodMeat: cattle,sheep,goats,horseMeat products necVegetable oils and fatsDairy productsProcessed riceSugarFood products necBeverages and tobacco products

Manufacturing TextilesWearing apparelLeather productsWood productsPaper products, publishingPetroleum, coal products

Chemical,rubber,plastic prods Mineral products necFerrous metalsMetals necMetal productsMotor vehicles and partsTransport equipment necElectronic equipmentMachinery and equipment necManufactures nec

Services ElectricityGas manufacture, distributionWaterConstructionTradeTransport necSea transportAir transportCommunicationFinancial services necInsuranceBusiness services necRecreation and other servicesPubAdmin/Defence/Health/

EducatDwellings

Additional data on energy use:

Coal, oil, gas, petroleum products, electricity

Volumes (Mtoe) and prices

Harmonized and consistent

Biomass small share -> not (yet) separate

GTAP v6 regions

87 regions (built from 226 members) Economy-wide coverage Input-Output tables Bilateral trade flows Policy data Benchmarked to year 2001

Key Resource: a standard multi-region CGE model

profit maximizing behaviour of producers (input restraint)

utility maximizing behaviour of consumers(budget constraint)

Armington approach: bilateral trade Computable General Equilibrium (CGE): all

markets clear simultaneously Product markets Factor markets (capital, labour, land, natural

res.) Prices and quantities determined by model

Model structure for one GTAP region

Economics of production: Cost Minimizing Behavior

Behavioural equations are derived from an economic decision problem:

Choose cost minimizing input mix, given the technical characteristics of production, and given prices:

Cost function: C(w, q) = min wi xi st.: q = f(x)

C(w, q)Factor demand: ----------- = xi(q, w) (Shephard’s lemma)

wi

Pricing

Constant returns to scale production function means:average cost = marginal cost

Competitive pricing at marginal cost. Therefore revenue = costs

Zero Pure Profits:

e trptrptrperpe

rprp

QFPFQFEPFE

QOPS

,,,,,,,,

,,

Market clearing

On all markets we have:Qd = Qs

Product markets: 56 commodities Factor markets: Land, SkLabour,

UnskLabour, Capital, Nat resources Savings = Investment ; Global bank

Small example: EU sugar reforms

-20.0% -15.0% -10.0% -5.0% 0.0% 5.0% 10.0%

European Union 27

ACP sugar protocol countries

EBA (non ACP) countries

Brazil

Sugar dev countries no prefs

Thailand &Australia

All other countries

EBA scenario EU July 2004 reform proposal + EBA

• EBA: LDCs get quota & duty free access to EU

• EU reforms:

Beet quota down,Internal support price down

% change in sugar output under EBA and EU reform

EU sugar reforms: welfare effectsNational income effects, EV

-1000

-800

-600

-400

-200

0

200

400

600

800

1000

European Union27

ACP sugarprotocol countries

EBA (non ACP)countries

Brazil Sugar devcountries no prefs

Thailand&Australia

All other countries WORLD

mil

lio

n U

S$

EBA scenario

EU July 2004 reform proposal + EBA

4600

Trade diversion

effect

Modifications to standard GTAP model

Network members add new model developments:

For example: imperfect competition and increasing returns to

scale international technology spillovers Modelling specific policies, e.g EU CAP Projections in time Dynamic extensions (GTAPDyn) Modelling of agricultural supply GTAP-E: energy model

GTAP organization

Core funding from international consortium, 18 international and national agencies, e.g:

• WTO, World Bank, UNCTAD, USDA/ERS, OECD,

MIT, USEPA ...LEI (member since Nov. 96)

• Each represented on the advisory board

• Headquarters Purdue Univ., Tom Hertel

Growth of GTAP Consortium

02468

101214161820

#

Five international: OECD, WB, WTO, UNCTAD

11 national: Australia, Denmark, Germany, Netherlands, Japan, US

Two non-profit

GTAP organization (II)

Open network of contributors

• more than 350 course alumni

• annual global conference

• Linked through WWW for distribution of database, software and publications: www.gtap.org

Summary

GTAP relatively unique in economics, but similar to consortia in sciences

Success of project traced back to courses Consortium key to long term viability: provides

$, but also key data inputs and much needed direction

Modest contribution by many agencies provides long run stability

Network externalities

There are many ways to get involved

Short course attendance = usually first step Purchase database and Gempack software Attend annual GTAP conference (approx

150 specialists in gobal economic analysis meet) 2003: The Netherlands, The Hague 2004: USA, Washington DC 2005: Germany, Lübeck

END

make or buy? make or buy? cooperatecooperate