GEM: A New International Macroeconomic Model

O C C A S I O N A L PA P E R

GEMA New International

Macroeconomic Model

Tamim BayoumiWith assistance from Douglas Laxton, Hamid Faruqee, Benjamin Hunt,

Philippe Karam, Jaewoo Lee,Alessandro Rebucci, and Ivan Tchakarov

239

INTERNATIONAL MONETARY FUND

Washington DC

2004

239

GEMA New International Macroeconomic Model

O C C A S I O N A L PA P E R 239

GEMA New International

Macroeconomic Model

Tamim BayoumiWith assistance from Douglas Laxton, Hamid Faruqee, Benjamin Hunt,

Philippe Karam, Jaewoo Lee,Alessandro Rebucci, and Ivan Tchakarov

INTERNATIONAL MONETARY FUND

Washington DC

2004

© 2004 International Monetary Fund

Production: IMF Multimedia Services DivisionFigures: Theodore F. Peters, Jr.

Typesetting: Alicia Etchebarne-Bourdin

Cataloging-in-Publication Data

Price: US$25.00(US$22.00 to full-time faculty members and

students at universities and colleges)

Please send orders to:International Monetary Fund, Publication Services

700 19th Street, N.W., Washington, D.C. 20431, U.S.A.Tel.: (202) 623-7430 Telefax: (202) 623-7201

E-mail: [email protected]: http://www.imf.org

recycled paper

Bayoumi, Tamim A.GEM: a new international macroeconomic model/Tamim Bayoumi; with

assistance from Douglas Laxton . . . [et al.]—Washington, D.C.: InternationalMonetary Fund, 2004.

p. cm.—(Occasional paper, 0251-6363); 239

Includes bibliographical references.

ISBN 1-58906-375-9

1. Economic policy—Econometric models. 2. Macroeconomics—Econo-metric models. I. Laxton, Douglas. II. Occasional paper (International Mone-tary Fund); no. 239.

HD87.B39 2004

Preface v

1 Overview 1

II Philosophy and Approach 2

Why a New Model? 2Structure of GEM 3Strengths and Weaknesses of GEM 7

III How Has GEM Been Used? 11

Measuring the Benefits of Raising Euro Area Labor and ProductMarket Competition 11

Should Monetary Policy Rules Differ Between Industrial Countries andEmerging Market Countries? 14

The Impact of Higher Oil Prices 17

IV Current Development Work 20

Fiscal Policy 20International Asset Markets 22

V The Road Ahead 24

References 26

Boxes

2.1. Estimating Parameter Values 63.1. GEM Simulations of the Benefits of Greater Euro Area

Competition 123.2. Using GEM to Analyze Monetary Policy Rules 16

Tables

2.1. Stylized View of the Strengths and Weaknesses of SuccessiveGenerations of Macroeconomic Models 4

3.1. GEM Estimates of the Long-Run Effects of More Competition-Friendly Policies in the Euro Area 13

3.2. MULTIMOD: Impact of a Permanent $5 a Barrel Increase inOil Prices After One Year 17

Figures

2.1. Stylized View of Model Development 32.2. Simple GEM Structure 52.3. More Complicated GEM Structure 5

Contents

iii

CONTENTS

2.4. Dynamic Responses to a One Percentage Point Hike in Interest Rates for One Year: GEM Compared with Large Forecasting Models 8

2.5. Dynamic Responses to a Hike in Interest Rates: GEM Compared with a VAR 9

3.1. Dynamic Effects of More Competition-Friendly Policies in the Euro Area (Anticipated and Perfectly Credible) 14

3.2. Taylor Trade-Off in Monetary Policy Analysis 153.3. GEM: Impact of a Permanent 20 Percent Oil Price Hike After

One Year 183.4. GEM: Impact of a Permanent and Temporary 20 Percent Oil

Price Hike After One Year 184.1. Structural Fiscal Balances in the Major Economic Regions 204.2. Sum of International Assets and Liabilities in Major Advanced

Economies 214.3. Capital Constraints for Emerging Markets 22

iv

The following symbols have been used throughout this paper:

. . . to indicate that data are not available;

— to indicate that the figure is zero or less than half the final digit shown, or that the itemdoes not exist;

– between years or months (e.g., 2003–04 or January–June) to indicate the years ormonths covered, including the beginning and ending years or months;

/ between years (e.g., 2003/04) to indicate a fiscal (financial) year.

“n.a.” means not applicable.

“Billion” means a thousand million.

Minor discrepancies between constituent figures and totals are due to rounding.

The term “country,” as used in this paper, does not in all cases refer to a territorial entity thatis a state as understood by international law and practice; the term also covers some territorialentities that are not states, but for which statistical data are maintained and provided interna-tionally on a separate and independent basis.

This paper provides a nontechnical overview of the Global Economic Model(GEM), a new multicountry model based on strong microeconomic underpinningsdeveloped in the Research Department of the International Monetary Fund.

GEM was, above all, a collaborative effort. The idea came from then EconomicCounsellor Kenneth Rogoff, and the main development was accomplished by Douglas Laxton of the IMF’s Research Department, and Paolo Pesenti, who came tothe Fund on a six-month assignment from the Federal Reserve Bank of New York.They were ably assisted by Susanna Mursula, also of the Research Department. With-out the energy and drive of these three individuals the project would not have got offthe ground. Since then, many others have also made important contributions to differ-ent aspects of the project that range from providing useful comments and helping tobuild the toolbox to support GEM development to using or extending the model to ad-dress real world policy issues. Principal among them have been Stéphane Adjemian,Dennis Botman, Selim Elekdag, Hamid Faruqee, Benjamin Hunt, Tore AndersHusebo, Michel Juillard, Sarma Jayanthi, Philippe Karam, Heesun Kiem, Jaewoo Lee,Gian Maria Milesi-Ferretti, Alin Mirestean, Dirk Muir, Kjetil Olsen, Alessandro Rebucci, Luca Ricci, Oistein Roisland, Tommy Sveen, Ivan Tchakarov, and RanjithVarma. In addition, the team has had valuable input from a range of distinguished vis-iting scholars including Paul Bergin, Fabio Canova, Lawrence Christiano, GiancarloCorsetti, Michael Devereau, Chris Erceg, Jordi Galí, Fabio Ghironi, Chris Gust, LucaGuerrieri, Frank Schorfheide, Chrisopher Sims, and Raf Wouters. Useful commentshave also been received at presentations at workshops both inside and outside theFund, most notably by Jarle Bergo, Ralph Bryant, Richard Harrison, Klaus Schmidt-Hebbel, Dale Henderson, Tiff Macklem, David Reifschneider, Thomas Sargent, FrankSmets, and Michael Woodford. Finally, Victoria Ashiru, Alfred Go, and Laura Leonprovided able assistance in producing this manuscript, and Esha Ray of the ExternalRelations Department coordinated production of the publication.

The contributions of all of these people have helped ensure this small pebble hasbeen thrown into the sea of knowledge.

Preface

v

Over the last two years, IMF staff has been devel-oping a new multicountry macroeconomic

model called the Global Economy Model (GEM). Anobvious question is why such a model is needed,given that the Fund’s existing model, MULTIMOD,also focuses on interdependence across countries.This paper answers this question by explaining howGEM differs from its predecessor and outlining howthese new features can improve the Fund’s policyanalysis. The paper is aimed at a general audience andavoids technical detail. Following this overview, Sec-tion II outlines the motivation, structure, strengths,and limitations of the model. Section III presents amore detailed discussion of three simulation exercisesthat have been completed. Section IV looks at areas inwhich development of the model is currently underway, while the final section provides a more generaldiscussion of GEM’s future path.

GEM is an early example of a large internationalmacroeconomic model built using recent economicresearch based on an explicit microeconomic frame-work in which consumers maximize utility and pro-ducers do the same with profits. In particular, the in-tegration of domestic supply, demand, trade, andinternational asset markets in a single theoreticalstructure allows transmission mechanisms to be fullyarticulated, providing a range of new insights not ob-tainable from earlier models more loosely linked totheory. Being on the forefront of this work has manyadvantages, including ensuring that Fund analysisincorporates recent theoretical advances in interna-tional macroeconomics and finance. There is a paral-lel with the introduction of rational expectations intopolicy models in the 1980s (of which MULTIMODwas a pioneering effort for multicountry work)whose more sophisticated dynamic responses to ag-gregate demand disturbances also led to a range ofnew insights. Indeed, the Fund is again on the cut-ting edge of developing international policy models

with GEM. While still very much a project in devel-opment, GEM is already the model of choice formost policy simulations, although MULTIMOD re-mains useful for some types of analysis.

An example of the advantages of this more theo-retically integrated approach concerns the treat-ment of international trade. In MULTIMOD, as inother models of its type, trade was modeled usingreduced-form dynamic equations based on the de-mand for goods that, in practice, resulted in limitedinternational spillovers in goods markets. In GEM,by contrast, the impact of changes in activity andthe exchange rate on trade depends on the interac-tion of consumer preferences, technology, andcosts of adjustment of volumes and prices. As a re-sult, the responses of exports and imports dependon a wide range of factors, most notably the type ofgood being traded (for example, components forproducers or final goods) and the type of distur-bance (for example, temporary or permanent).

A range of GEM simulations have already beenused in Fund work to assess issues such as the do-mestic and international consequences of policies toincrease competition in markets, the impact of oilprice hikes, the effects of exchange rate volatilityacross industrial countries on emerging marketeconomies, and appropriate monetary policy rules foremerging market countries. To take a specific exam-ple, GEM was used to identify the benefits of greatercompetition in the euro area in a manner that was notpossible in earlier models. This was achievable be-cause the level of competition across firms is explic-itly modeled in GEM, so that the impact of changingit can be analyzed directly. In addition, the simulationindicates that greater competition in labor and prod-uct markets also fosters greater wage and price flexi-bility, an intuitive result that again reflected the mi-croeconomic foundations of the model, this time withrespect to the setting of prices and wages.

I Overview

1

Why a New Model?

The development and rationale for GEM can bebest explained through a brief description of the roleof large macroeconomic policy models. Academicwork in macroeconomics tends to focus on specificissues, such as the consumption function or a newtheoretical insight. Large macroeconomic policymodels, on the other hand, are used to quantify theimpact of a range of issues within a unified structure,most notably countercyclical macroeconomic poli-cies. A stylized way of thinking about the interactionbetween academic work and large policy models isprovided in Figure 2.1. A new theoretical insight(such as rational expectations) with strong policy im-plications is developed in academia in response toevolving policy challenges and the limitations of ex-isting models. Once these ideas have been distilled tothe point where they are able to fit the data reason-ably, they form the basis for large policy models,starting with single-country versions and then ex-tending to a multicountry setting. Subsequently, theacademic and policy communities refine these ideasand the paradigm becomes increasingly dominant. Atsome point, a new insight emerges and the leadingedge of academic work switches to this new para-digm. However, large policy models do not followbecause the ideas are not yet able to provide theneeded quantitative insights, and academic interest inlarge macroeconomic models wanes. In short, the“production cycle” of policy models tends to lag thatof their academic brethren, given the greater need forpolicy models to fit the stylized facts of the cycle.

One such major overhaul occurred with the adop-tion of rational expectations (Table 2.1 summarizessuccessive generations of policy models). In the1960s and 1970s large policy models using adaptiveexpectations and a Keynesian aggregate demandframework quantified the impact of macroeconomicpolicies. However, in the wake of the great inflationof the 1970s, the implication that output could beraised permanently by injecting aggregate demandthrough monetary and fiscal policy was recognizedas a flaw. Rational expectations fixed this and pro-vided a new range of insights, such as the impor-

tance of rules in macroeconomic analysis (Taylor,1993), exchange rate overshooting (Dornbusch,1976), and the “random-walk” model of consump-tion (Hall, 1978). Such models were gradually de-veloped to the point where they could be used in pol-icy circles. Indeed, MULTIMOD, created in the late1980s, was an early example of a large internationalversion of such a model (see Masson, Symansky,and Meredith, 1990, for a description).

These rational expectations models, however, weresusceptible to the “Lucas critique.” This was that pol-icy analysis using reduced-form equations that fit thedata but were loosely tied to theory—such as thoseused in large macroeconomic policy models—wasfraught with danger, as such models could not ade-quately account for resulting shifts in behavior(Lucas, 1976). The focus in much of academia in the1980s and early 1990s was on developing rational ex-pectations models incorporating the explicit micro-economic structure advocated by Lucas. Initially thistook the form of “real-business-cycle” models inwhich prices were assumed to be fully flexible (seeKydland and Prescott, 1982, for a closed economyversion and Mendoza, 1991, written in the ResearchDepartment of the IMF for an open economy one).However, the assumption of flexible prices largelyobviated the impact of macroeconomic policies onreal activity, making these models of little value inanalyzing such policies. Consequently, large policymodels generally remained in the reduced-form Key-nesian framework, although with an increased focuson adding supply-side linkages.

Over time, it became increasingly clear that theshort-term dynamics of real-business-cycle modelscould be improved by introducing some form ofnominal inertia. Theoretical developments in the mi-croeconomics of wage and price setting with imper-fect competition led to single-country monetarymodels that combined the explicit microeconomicfoundations typical in real business cycle modelswith price stickiness (Christiano, Eichenbaum, andEvans, 2001). The leap to multicountry models ofthis type was accomplished in the mid-1990s (Obst-feld and Rogoff, 1995). The new models merged themicroeconomic foundations (of the type advocated

II Philosophy and Approach

2

Structure of GEM

by Lucas) with sticky prices, combining production,consumption, nominal rigidities, trade, and interna-tional financial markets in a coherent theoreticalstructure. Work on such models has exploded in re-cent years (Lane, 2001, provides a survey).

This deep paradigm shift is transforming the studyof international finance and international macroeco-nomics, reevaluating the Mundell-Flemming analy-sis developed at the Fund in the early 1960s, in thesame way that the traditional IS/LM/Phillips curveanalysis was recast by recent monetary models (seeObstfeld, 2001). Major insights from these modelsinclude that macroeconomic policies, such as the ex-change rate regime, can have long-term effects onthe level of consumption, labor effort, and the capitalstock, in contrast to earlier views generally held inthe profession. In addition, policies can be analyzedin terms of their impact on economic welfare of con-sumers—which includes, for example, the disutilityof working harder and having less leisure—ratherthan their effect on less accurate proxies for welfaresuch as output and inflation. This has reignited inter-est in the impact of alternative exchange rateregimes, the benefits from international macroeco-nomic cooperation, and the role of asset markets inthe international business cycle.

Structure of GEM

GEM is a large-scale version of such a micro-founded open economy model. It integrates and

builds upon the results in the existing literature—mostly devoted to exploring small and relativelytractable apparatuses—to create a unifying frame-work for the analysis of international interdependen-cies. GEM has a modular structure, allowing themodel to treat issues in a flexible manner. In additionto GEM, such models have been developed by theFederal Reserve Board (Erceg, Guerrieri, and Gust,2003), and are areas of active research in severalother institutions (such as the central banks ofCanada, Finland, Italy, Norway, Spain, and theUnited Kingdom), and are being considered in someemerging market countries, such as by the centralbanks in Brazil, Chile, and the Czech Republic. TheEuropean Central Bank (ECB) has developed a sin-gle-country model (Smets and Wouters, 2002) and isplanning a multicountry extension.

The model comprises firms that produce goods,households that consume and provide labor and cap-ital to firms, and a government that taxes and spends(Laxton and Pesenti, 2003). The microeconomicstructure of GEM uses standard functional formsthat allow firms and consumers to be aggregated asif they were a single entity. On the production side,for example, many small firms produce differenti-ated goods made using identical constant elasticityof substitution (CES) production functions usinglabor, capital, and (in some cases) intermediategoods such as components or commodities. Becausethe goods are differentiated, firms have marketpower and restrict output to create excess profits.Capital and intermediate goods can be produced and

3

Paradigm dominates

Paradigm dominates

Paradigm dominates

Paradigm wanes

Paradigm wanes

New ideasdeveloped

New ideasdeveloped

Policy models

Policy models

Academic work

Time

Academic work

Academicwork

Paradigm 1

Paradigm 2

Paradigm 3

Policymodels

Figure 2.1. Stylized View of Model Development

II PHILOSOPHY AND APPROACH

traded while the labor force in each country is fixed,with workers choosing how much to work versus en-joying leisure. Workers also have market power andhence restrict their labor to raise their real wage. Theworkers own the firms in their country, and hence re-ceive their revenues (net of investment) in the formof wages and profits. This income is spent on homeand foreign goods based on a CES utility function.Given the focus on trade and macroeconomic inter-dependencies, the fiscal and financial sides of themodel are currently relatively simple. The govern-ment spends on government consumption fundedthrough lump-sum taxes less transfers, domestic fi-nancial sectors are not modeled explicitly, whilecountries pay (receive) a small premium for interna-tional borrowing (lending). These sectors are allareas of active development (see Section IV).

To generate realistic dynamics, the model includesjudicious use of adjustment costs on real and nominalvariables, thereby elongating the responses to shocksand ensuring that consumption and production do notimmediately jump to a new long-term equilibrium.On the real side, such costs prolong the adjustment ofthe capital stock and the level of imports, while“habit persistence” plays a similar role in elongatingthe responses of consumption and hours worked.Sticky prices are also modeled using adjustmentcosts, with the prices of domestic goods and imports,as well as wages, displaying inertia. These costs aremodeled parsimoniously with only one or two para-meters determining the speed of response, and arefully integrated into the theoretical structure.

An innovative feature of GEM compared to mostpolicy models is that it has a flexible structure, sothat one can include or exclude features such as non-traded goods, a distribution sector, or trade in com-modities or other intermediate goods. In addition,the model can be created with any number of coun-tries, although work to date has involved either twoor three countries. Figure 2.2 illustrates the simplestpossible version of the two-country model, in whichlabor and capital are combined to produce a singletype of tradable good that can be used for consump-tion or investment. Given the preferences of con-sumers, firms, and governments, these goods arethen distributed across countries.

Figure 2.3 shows the same two-country modelwith three major additional features incorporated.The first is that production is split into two stages.In the first stage, labor, capital, and (possibly) landare used to create intermediate goods that can betraded, such as oil or components for manufactur-ing. These intermediate goods are then combinedwith additional labor and capital at home andabroad to produce final goods. The addition of in-termediate goods allows the model to examine is-sues that are particularly important for developingcountries. These include the policy challengesfaced by economies that supply either low value-added components (such as textiles) to industrialcountries, assemble higher-technology componentsfrom such countries into final products (for exam-ple, assembling computers), or are commodity pro-ducers and exporters.

4

Table 2.1. Stylized View of the Strengths and Weaknesses of Successive Generations ofMacroeconomic Models

Keynesian Keynesian Stochastic Dynamic Model Type Adaptive Expectations Rational Expectations Real Business Cycle General Equilibrium

Strengths Allowed researchers to Generated more realistic Strong theoretical founda- Integrates aggregate assess the impact of policies dynamic responses to tions improved supply supply and demand and other cyclical shocks in cyclical disturbances. side and allowed direct responses through a unified manner. calculation of welfare. microeconomic theory.

Weaknesses Adaptive expectations Absence of strong theoret- Assumption of flexible Models are in early allowed policymakers to ical foundations made it prices left little room for stages of development consistently mislead the difficult to assess effects of analysis of macroeconomic and are difficult to build public, creating a bias policies on aggregate policies. and run.toward expansionary supply.macroeconomic policies.

Major IMF con- Mundell-Flemming model, MULTIMOD Mendoza (1991) GEMtributions to Mundell (1963) and international Flemming (1962)analysis

Structure of GEM

Another feature shown in Figure 2.3 is that finalgoods are split into those that can be traded andthose that cannot. Differentiating between tradedand nontraded goods is central to a number of issues

in international macroeconomics. Most notably,rapid productivity increases in traded goods relativeto nontraded goods help explain why real exchangerates tend to appreciate in countries that are growing

5

Final goods

Investment InvestmentConsumption Consumption

Final goods

Capital CapitalLabor Labor

Private Private Private PrivatePublic Public Public Public

Home Country Foreign Country

TradeTrade

ProductionProduction

Figure 2.2. Simple GEM Structure

Investment InvestmentConsumption Consumption

Distribution Final goods

Intermediate goods

LandLabor LaborCapital CapitalLand

Intermediate goods

Final goods

Private

NontradedDistributionNontradedNontraded NontradedTraded Traded

Private Private PrivatePublic Public Public Public

Home Country Foreign Country

Trade

Production

Trade

Production

Figure 2.3. More Complicated GEM Structure


rapidly—generally referred to as the Balassa-Samuelson effect (Balassa, 1964; and Samuelson,1964). Including nontraded goods is also useful formany industrial country issues, such as the degree towhich actual (and anticipated) productivity increasesin information technology goods help explain thestrong appreciation of the U.S. dollar over the late1990s (see Hunt and Rebucci, 2003, for an analysisof this issue using GEM).

Finally, a distribution sector is included. There isstrong evidence from microeconomic studies that thesame goods are sold at different prices across coun-tries. One way of incorporating this observation is toinclude a distribution sector in the model (Corsettiand Dedola, 2002). All domestic and foreign goodsneed to go through this sector before they can bebought. As the distribution sector is assumed to con-sist of nontraded goods, this means that the finalprices of all goods are an amalgam of the cost ofproducing these goods and domestic distributioncosts, so prices of imported tradable goods do not

fully reflect changes in the real exchange rate evenin the long run.

GEM’s flexible modular structure provides a num-ber of advantages. Given the size and complexity ofthe model, it is often useful to ignore factors that arenot of central interest to the issue at hand. For exam-ple, while one would wish to include commoditieswhen analyzing a major commodity producer, or theimpact of oil price shocks, it is essentially a distrac-tion when looking at countries specialized in manu-facturing. Similarly, distribution costs matter whenpass-through of exchange rates into prices is impor-tant, but is an unnecessary complication for manyother issues. In addition, the transmission mecha-nisms become simpler and more transparent insmaller versions of the model, allowing the conse-quences of the theoretical structure to be more easilyascertained, making the model less of a “black box.”Simpler models are also essential for some forms ofsimulation and estimation that are particularly com-puter intensive.

6

Box 2.1. Estimating Parameter Values

The deep parameters in micro-founded models likeGEM—such as the degrees to which changes in realwages affect the desire to work, consumption respondsto changes in real interest rates, firms can substitutelabor and capital in response to changing conditions,and home and foreign goods are substitutes—have gen-erally been calibrated using estimates from microeco-nomic studies combined with an assessment of the waythe model fits the overall properties of the data. Unfor-tunately, parameterization is time consuming, as it in-volves experimenting across a wide range of potentialvalues. Ideally, it would be better to simplify thisprocess by estimating the parameters from macroeco-nomic data.

The major constraint to estimating deep parametersis that their impact on the model’s short-term dynamicsis often subtle. As the information in the data mainlypertains to these dynamic responses, it is difficult to ob-tain accurate estimates of crucial elasticities using clas-sical estimation techniques. The standard errors onthese coefficients are generally large, and the estimatedvalues often deviate a long way from values regardedas “reasonable” based on theory and microeconomicestimates. Hence the preference for the more time-consuming approach of calibration based on microeco-nomic evidence.

Providing parameters that fit the facts and are rela-tively simple to estimate for different countries is par-ticularly important for a policy model such as GEM, asthere is a premium on making the model both fit thefacts and look at a range of issues across many differ-ent countries. Accordingly, the GEM team has beenworking on using a Bayesian approach to estimate the

model’s parameters. Bayesian estimation is the mainalternative to the classical approach. The key differ-ence between the two is that in the Bayesian frame-work the analyst specifies his or her initial view aboutthe value of each parameter and the certainty withwhich this view is held before estimating the model.This prior information is combined with the evidencecontained in the data to obtain final estimates of themodel’s parameters. Initial views about parameters canthus be used to improve the accuracy of estimates ofkey elasticities on which the underlying data have lim-ited information, and the resulting estimates can beused to create parameterizations for simulations.

In many respects, Bayesian estimation is simply amechanized version of the approach currently used tocalibrate models, with its combination of using priorinformation obtained from outside sources, such as mi-croeconomic studies, with judgments about how wellthe model fits the data. The great advantage is that,once the Bayesian routine has been established, it canbe used to rapidly provide plausible parameterizationsacross countries. For example, priors can be used to en-sure that deep underlying elasticities across a range ofcountries could be very similar (on the basis thathuman nature is relatively invariant across countries),while parameters associated with the speed of adjust-ment, which are more dependent on the particular insti-tutional arrangements in a country, can be derivedmainly from the data. Indeed, Bayesian techniqueshave already been used to generate plausible, albeitpreliminary, parameter estimates for a simple closedeconomy model of the United States, but have not yetbeen extended to the full GEM.

Strengths and Weaknesses of GEM

Once the structure of the model has been deter-mined, the parameters are selected. For most deepparameters that define long-term responses of firmsand consumers, such as the responsiveness of hoursworked to changes in real wages or the substi-tutability of different types of goods, estimatesfrom microeconomic studies are used to determineplausible values (Box 2.1 discusses issues associ-ated with estimating parameters). Next, more de-tailed coefficients are selected to mimic key char-acteristics of the economic environment, such asthe relative size of the countries, their levels oftrade, and their capital-output ratios. Finally, thecoefficients on costs of adjustment and habit per-sistence are chosen to generate realistic dynamicresponses.

An important feature of any policy model such asGEM is that it fits the dynamics seen in the data. Tothis end, the adjustment cost parameters are cali-brated to fit existing evidence from policy modelsand estimated vector autoregressions (VARs). Figure2.4 provides a comparison of the responses to a one-year hike in short-term interest rates in the euro areaand the United States in GEM to those from policymodels used in central banks—the Area-WideModel (AWM) of the ECB and the FRB-US modelof the Board of Governors of the Federal ReserveSystem, respectively. These models were chosen asthey are primarily designed to fit the dynamics in thedata. An alternative is to compare the model to theresponses to an interest rate hike found using an esti-mated vector autoregression (Figure 2.5). In additionto having no imposed theoretical structure, vectorautoregressions provide statistical confidence inter-vals, thereby giving a better sense of the plausibilityof the responses produced by GEM.1 In both cases,GEM reproduces the typically hump-shaped path ofvariables, although the GEM responses tend to besomewhat faster, particularly for investment, possi-bly reflecting the absence of lags coming from thetime it takes to complete a project once initiated. Inshort, even with its strong theoretical underpinnings,GEM’s structure is rich enough to mimic short-termdynamics.


One of the great advantages of GEM comparedwith earlier types of models is that it can provideevaluations of policies in a general equilibrium set-ting, thus taking account of the full range of effects

across equations. As the model is built from explicitmicroeconomic foundations, a change in one of thedeep parameters in the model can have effects acrossa wide range of relationships. These complex inter-relationships help to identify economic linkagesmore precisely, providing a stronger framework foranalysis that can generate new insights as well as en-couraging closer links between IMF researchers andthe academic community. This is particularly impor-tant at the early stage of creation of a paradigm,when these insights have not been fully incorporatedinto mainstream analysis, and help explain the en-thusiasm GEM has created in academia.

As an example of unveiling linkages, consider apolicy that increases competition in the labor market(discussed in more detail in Section III). The mostobvious effect is that the market power of workersdiminishes, increasing output and, as more goodsneed to be sold to the rest of the world, depreciatingthe real exchange rate. The effects on domestic out-put depend crucially on the response of hoursworked to a change in real wages, while the interna-tional effects depend on the degree to which homeand foreign goods are substitutes. As these parame-ters are explicitly identified in GEM, the conse-quences of different assumptions about them can beeasily qualified, while in MULTIMOD these elastic-ities were combined with others in reduced-form re-lationships. In addition, an important new insightcoming from GEM is that more labor market compe-tition reduces nominal inertia. This is because it in-creases the costs firms incur when wages deviatefrom their flexible price level. Such an effect couldnot have been captured in models such as MULTI-MOD with a fixed Phillips curve relationship.

Another example is the impact of industrial coun-try exchange rate volatility on emerging marketcountries (see IMF, 2003a). The flexibility of GEMmeans that the effects of structural differences in theemerging market on the impact of such volatility canbe explored, including alternative exchange regimes,levels of openness, bilateral trade patterns, levels ofdebt, and exchange rate pass-through. Insights fromthis exercise include the importance of the exchangerate regime and degree of domestic exchange ratepass-through on the associated output volatility, ef-fects that depend crucially on the integration of sup-ply, demand, trade, and international asset marketsin GEM.

A second advantage of GEM is that the costs andbenefits of a policy can be evaluated in a more so-phisticated manner. As the model is derived from ex-plicit maximization of profit and utility, one canevaluate policies in terms of their effects on con-sumer welfare. The advantage of welfare is that itmeasures the gain to consumers, the ultimate objec-tive of economic activity, and provides a measure of

7

1That said, not all of the results using VARs are plausible, suchas the positive short-term response of prices to a hike in interestrates.


8

–0.4

–0.3

–0.2

–0.1

0

8765Quarters Quarters

Quarters Quarters

Quarters Quarters

Quarters Quarters

4321–0.4

–0.3

–0.2

–0.1

0

87654321

–1.6

–1.2

–0.8

–0.4

0

87654321–0.16

–0.12

–0.08

–0.04

0

87654321

–0.6

–0.4

–0.2

0

87654321–0.6

–0.4

–0.2

0

87654321

–1.2

–0.8

–0.4

0

87654321–0.5

–0.3

–0.1

0.1

87654321

ECB model

GEM GEM

GEM

GEM GEM

GEM

GEM

GEM

ECB model

FRB-US modelFRB-US model

FRB-US model

FRB-US model

ECB model

ECB model

Real GDP

Real GDP

Investment

Investment

Consumption

Euro Area

United States

Consumption

Consumer Price Index


Figure 2.4. Dynamic Responses to a One Percentage Point Hike inInterest Rates for One Year: GEM Compared with Large ForecastingModels(Percent deviation from baseline)

Source: IMF staff calculations based on information provided by the U.S. Federal Reserve Board and theEuropean Central Bank.


the “best” policy as well as a way of comparing theeffects of different policy options. By contrast, inMULTIMOD and similar models benefits were ana-lyzed using more ad hoc measures. For example,monetary policy was generally evaluated in terms ofthe variability of real output and inflation.

The large size of GEM is an advantage in this re-spect. Broadly speaking, policies are most usefulwhen they can help reduce economic distortions. AsGEM includes a relatively large number of such dis-tortions—such as monopolistic competition, stickyprices, and sluggish adjustment of trade volumes—the potential role for policies to improve welfare iscommensurately strengthened. That said, to date ithas only been possible to evaluate welfare benefits insmall models with simple structures often compris-ing only one country. This is because of the largecomputing power needed to solve for a dynamic pathwith the nonlinear functions needed to calculate wel-fare. Given the rapid progress in both computingpower and solution techniques, however, it is reason-able to expect that full welfare calculations of dy-namic simulations will be possible with GEM in thenear future.

Even without full welfare analysis, GEM can stillprovide new insights. As discussed further in SectionIII, the appropriate monetary rule depends impor-tantly on how potential output is evaluated. A modelsuch as GEM can provide a more accurate calcula-tion of the output gap by incorporating the impact ofshocks on the path of aggregate supply. Policy rulesusing this output gap can then be compared withthose using a more traditional approach in which po-tential output is assumed to change slowly over time,providing insights into the degree to which monetarypolicymakers should focus on identifying the sup-ply-side implications of disturbances in assessingmonetary conditions.

It is important to recognize that moving to amodel with a tight theoretical structure also imposeslimitations, at least in the short term. Accordingly,MULTIMOD remains a useful tool of analysis, al-though future development will cease and its usewill presumably diminish over time. There are theusual growing pains associated with any new pro-ject, such as the need to gain more experience withversions of GEM comprising three or more coun-tries. In addition, the need to create a large inter-linked structure constrains theoretical specificationsand hence model properties. For example, the use ofa representative consumer means that the model isnot currently suitable for analysis of income distrib-ution. The need for theoretical consistency can alsocomplicate the addition of new features. For exam-ple, the current version of the model does not gener-ate realistic short-term tax multipliers, although thisis an area of active development. As discussed fur-

9

–0.5

0

0.5

1.0

1917151311Quarters97531

–0.4

–0.2

0

0.2

0.4

191715131197531

191715131197531

191715131197531

191715131197531

–0.4

–0.2

0

0.2

0.4

–2

–1

0

1

–1.0

–0.5

0

0.5

95% confidence interval

Interest Rate

Quarters

Output

Quarters

Consumption

Quarters

Investment

Quarters


Result from GEM Result from VAR

Figure 2.5. Dynamic Responses to a Hike inInterest Rates: GEM Compared with a VAR

Sources: Altig and others (2003) for VAR; and IMF staffcalculations for GEM.


ther in Section IV, one of the main theoretical ap-proaches to creating such multipliers is to assumethat consumers have finite lives. Another implicationof finite lives is that consumers’ behavior depends onage and hence cannot be summarized by a single“representative” individual, which creates signifi-cant theoretical complications elsewhere.

Finally, calibration of GEM is currently time con-suming. This is partly because the concepts in themodel often do not dovetail with existing data. Forexample, it is not easy to split output into traded andnontraded goods or to determine the role of com-

modities and semifinished components in produc-tion. In addition, changes in a coefficient generallyaffect several equations, with a complex effect onmodel properties. To date, calibrations have onlybeen completed for three economies—the UnitedStates, the euro area, and the Czech Republic. Datasets that will help with calibration have been ob-tained for over 20 countries and regions, including awide range of emerging market countries as well asadvanced countries, while experience with earliercalibrations are helping to make the process lesstime consuming.

10

GEM simulations have already been used to pro-vide insights on a range of issues. In particular,

they have been incorporated into the IMF staff’sanalysis in the World Economic Outlook and otherIMF work examining the impact of entry into Euro-pean Monetary Union (EMU) on European Union(EU) accession candidates (Laxton and Pesenti,2003) and energy issues in the United States (IMF,2003b), as well as in contributions to academic con-ferences and journals. In many cases these effortshave been combined, so that work originally usedfor (say) the World Economic Outlook has generatedacademic papers, while work originally prepared foracademic conferences has provided the basis foranalysis to assist the staff. Indeed, such dual uses en-sure policy relevance and professional rigor. On theother hand, like MULTIMOD, GEM is not used togenerate the Fund’s forecasts. Rather, the WorldEconomic Outlook exercise uses the expertise avail-able on countries by aggregating projections fromindividual country desks.

This section discusses three such simulations,chosen to illustrate the value of GEM across a rangeof questions. Accordingly, each simulation focuseson a different type of shock and demonstrates a dif-ferent model strength. The first explores the impactof structural reforms that are assumed to raise com-petition in euro area labor and product markets toU.S. levels and illustrates how the microeconomicfoundations of the model allow GEM to tackle issueson which earlier models provided little insight. Thesecond asks how monetary policy rules should differbetween industrial countries and emerging marketcountries and illustrates how GEM can be used toevaluate the impact of policies in more sophisticatedways. The third examines the consequences of oilprice hikes on industrial countries and illustrateshow GEM’s flexible structure can be used to providedeeper analysis of an issue. These simulations do notexhaust the work using GEM. Other exercises usingGEM, which are not discussed in this section, in-clude an analysis of external developments in theUnited States in the late 1990s (Hunt and Rebucci,2003) and the impact of industrial country exchangerate instability on emerging markets (IMF, 2003a).

Measuring the Benefits of Raising Euro Area Labor and Product MarketCompetition

Structural reforms are becoming an increasinglyimportant element in Fund policy advice, as it hasbecome clear that there is a close connection be-tween macroeconomic policies and the underlyingeconomic environment. While the issues differacross regions, in western Europe the focus has beenon ways to increase competition in domestic mar-kets. Indeed, European leaders at a number of recentsummits have embraced this objective, most notablyat Lisbon in 2001.

Greater competition in product and labor marketsclearly benefits an economy through ensuring moreefficient allocation of resources and making marketsmore flexible in the face of shocks. However, it hasproved difficult to use macroeconomic models toprovide quantitative assessments of these gains. Thispartly reflects difficulties in measuring such con-cepts as levels of competition or institutional quality,as well as the issue of how reduced-form relation-ships identified by regression analysis might changeif policies are directed at improving them—anotherexample of the Lucas critique.

Work using policy models loosely tied to theory hasrelied heavily on microeconomic studies, which pro-vided estimates of how regulatory changes would af-fect a range of variables that were then incorporatedexogenously in model simulations. A good exampleof such an approach is the 1997 OECD Report onRegulatory Reform (OECD, 1997). Detailed micro-economic studies were performed that estimated thebenefits of comprehensive reform in five highly regu-lated industries representing around 20 percent of out-put in the United States, Japan, Germany, France, andthe United Kingdom. The studies measured how suchreforms would reduce prices of these goods, raise pro-ductivity in the sector, and affect employment. Theaggregate effects across all five sectors were then fedinto a macroeconomic model by exogenously chang-ing labor efficiency, employment, prices, and (throughthe erosion of rent sharing) wages. The main lessonfrom these simulations was that if the gains in labor

III How Has GEM Been Used?

11

III HOW HAS GEM BEEN USED?

efficiency were not fully captured by higher wages inthe affected sectors, modest additional increases inoutput could occur elsewhere in the economy. Focus-ing on the euro area, the results suggest that the initialincrease in real GDP found in the microeconomicstudies of 4 percent could rise by a further percentage

point. However, the bulk of the effects derive from theinitial microeconomic studies, with limited additionalinformation coming from the macroeconomic modelsimulations.

In GEM, lowering a single parameter, namely theequilibrium markup of prices over marginal cost,

12

Box 3.1. GEM Simulations of the Benefits of Greater Euro Area Competition

GEM simulations increasing competition in euroarea product and labor markets to U.S. levels reportedin Bayoumi, Laxton, and Pesenti (2004) imply thatsuch reforms would provide a wide range of benefits tothe euro area and the rest of the world (see Table 3.1).In particular:

• Euro area real GDP rises by 12!/2 percent, fueledby a 20 percent increase in the capital stock and an8 percent increase in hours worked. About two-thirds of these benefits to output are attributable toproduct market reforms, and the remainder to labormarkets.

• The percentage rise in euro area consumption isonly about two-thirds of the percentage increase inGDP, reflecting both the large rise in investmentand the real depreciation of the euro as higher out-put in the euro area lowers the real exchange rate.

• The rest of the world benefits, as the real deprecia-tion of the euro makes consumers elsewhere richerand higher euro area demand increases importsfrom the rest of the world. In particular, foreignconsumption rises by 1!/4 percent, about one-sixthof the increase in the euro area.

• The increases in welfare are substantial. Welfarerises by the equivalent of 2!/2 percent and 1!/4 percentof steady-state consumption in the euro area and therest of the world, respectively. Welfare rises less thanconsumption in the euro area because of higherhours worked, emphasizing the value of using a wel-fare-based measure of benefits as opposed to using(say) the increase in output or consumption.

• Euro area wages and prices become more flexible.As an illustration, the sacrifice ratio—defined asthe output cost of permanently lowering inflationby 1 percentage point—falls from two to underone-and-a-half. This makes it easier for the Euro-pean Central Bank to use monetary policy to stabi-lize the cycle.

• The increase in domestic output is relatively invari-ant when key parameters are changed (except thatthe impact of labor market reforms depends on theresponse of hours worked to movements in realwages), but the size of the spillovers to the rest ofthe world is quite sensitive to the chosen values fora number of important parameters, most notablythe substitutability of home and foreign goods.

• Dynamic simulations indicate that reforms increaseinvestment rapidly, but the benefits to consumptionsare more delayed (see Figure 3.1). After an initialboom funded from abroad, consumption fallsbelow baseline for a time as investment booms andreal interest rates rise before increasing as capacityrises. If the reforms are not fully credible, the con-sumption response becomes further elongated.

To explain the key linkages behind these results, it isuseful to focus initially on product markets. Increasedcompetition across firms raises desired output, therebyincreasing the demand for investment and labor. Thereis an investment boom as it is less costly in the long runfor firms to buy more capital than hire more labor, ascapital can be produced while the supply of labor isless flexible. Higher output at home reduces localprices compared with their foreign counterparts, andthis real depreciation boosts consumption abroad bymaking foreign consumers wealthier. In addition, do-mestic price stickiness is reduced as greater competi-tion increases the loss to the firm from allowing pricesto deviate from their desired level if prices were fullyflexible.

More competitive labor markets produce similar ef-fects, but with a larger impact on hours worked and asmaller effect on domestic output and internationalspillovers due to different microeconomic linkages.These differences emanate from the fact that labor reforms work by lowering real wages and raisinghours worked while product market reforms affectoutput and final prices directly. Because the main im-pact is on labor markets, the effect on output is mutedwhile lower real wages further diminish the interna-tional spillovers as the home country remains morecompetitive.

The simulation provides a number of policy conclu-sions. In particular, there are large domestic gains fromreforms to increase competition in both product andlabor markets, including by reducing the sacrifice ratio,and the rest of the world also benefits. However, the re-sults indicate that advantages for local consumers comeon stream only after a delay, particularly if the reformprogram is not fully credible. Such a pattern, whichseems consistent with experience, may explain part ofwhy such programs are difficult to initiate politically.These GEM simulations are simple to run, produceplausible benefits compared with earlier work, andavoid the need for expensive and complex microeco-nomic studies of the impact of specific reforms.

Measuring the Benefits of Raising Euro Area Labor and Product Market Competition

can simulate the impact of increasing product mar-ket competition, with a parallel structure in labormarkets. This illustrates the advantages of thestrong microeconomic foundations of the model. Inparticular, GEM assumes an explicit industrial orga-nization structure, namely, monopolistic competi-tion. Firms use their monopolistic power to restrictoutput and increase profits, reducing welfare andgenerating a markup of prices over marginal cost.By decreasing producers’ market power, greatercompetition in goods markets reduces markups andboosts output as lower goods prices raise real wagesand increase demand for products. The setup in thelabor market is analogous except the boost to outputcomes from lower costs to firms. Increasing compe-tition is thus equivalent to lowering the markup onprices and wages. Furthermore, it is possible to cali-brate this parameter relatively easily as thesemarkups have been estimated in the microeconomicliterature.

Accordingly, a two-country version of GEMusing a relatively streamlined structure (see Figure2.2) was created comprising the euro area and therest of the world, parameterized as the UnitedStates (Bayoumi, Laxton, and Pesenti, 2004).Based on cross-country empirical evidence for the1980s and 1990s, economy-wide price and wagemarkups were set at 35 percent and 30 percent inthe euro area, respectively, and 23 percent and 16percent in the United States. Simulations lowering

euro area markups to U.S. levels produced the fol-lowing results (for more details, see Box 3.1):

• Euro area output and welfare rises significantly(Table 3.1). Euro area real GDP rises by 12!/2percent and hours worked by 8 percent. The in-crease in welfare is smaller but still notable,equivalent to a 2!/2 percent rise in steady-stateconsumption as some of the benefits fromgreater consumption are partly offset by morework, demonstrating the value to using welfare-based criteria to evaluate policy changes. Abouttwo-thirds of these increases in output and wel-fare are attributable to product market reforms,and the remainder to more competitive labormarkets. These benefits are relatively invariantto alternative values of deep parameters.

• Reforms increase investment rapidly, but thebenefits to consumption are more delayed (Fig-ure 3.1). By diverting resources, the investmentboom generates a significant lag between the im-plementation of reforms and a sustained increasein consumption, particularly if the reforms arenot fully credible.

• There are positive spillovers to the rest of theworld. The need to sell the increased output leadsto a real depreciation of the euro, improving theterms of trade of the rest of the world. In the basecase simulation, consumption in the rest of the

13

Table 3.1. GEM Estimates of the Long-Run Effects of More Competition-Friendly Policies in the Euro Area(Percent deviations from baseline)

Product Market Labor Market Reforms Reforms Both Reforms

Euro areaGDP 8.6 3.5 12.4Consumption 4.9 3.3 8.3Investment 17.0 3.5 21.2Labor effect 4.5 3.6 8.3Real exchange rate 4.2 1.1 5.3

Utility1 1.9 0.9 2.4Sacrifice ratio 2.0 ⇒1.7 2.0 ⇒1.7 2.0 ⇒1.4

Rest of worldGDP 0.7 0.2 0.8Consumption 1.0 0.3 1.3Investment 0.5 0.1 0.7Labor effort 0.1 0.0 0.2Utility1 0.9 0.3 1.2

Source: Bayoumi, Laxton, and Pesenti (2004).1Percentage increase in terms of steady-state consumption.


world rises by about 1!/4 percent. As labor effort islargely unaffected, welfare also rises by the equiv-alent of 1!/4 percent of steady-state consumption.The size of these spillovers is sensitive to the as-sumed values of a range of deep parameters.

• Structural reforms ease the task of monetary poli-cymakers in the euro area. Greater competition,particularly in labor markets, reduces nominalrigidities in the euro area. This greater nominalflexibility reduces the inflationary costs of stabi-lizing output thereby making it easier to use mon-etary policy in a countercyclical manner.

Should Monetary Policy Rules DifferBetween Industrial Countries andEmerging Market Countries?

Ever since the rational expectations revolution,monetary policy has been analyzed in terms of the im-

pact of alternative rules rather than the impact of dis-cretionary responses to particular circumstances. Re-flecting the importance attached to the public’s as-sumptions about future policies, rational expectationmodels create an important distinction between un-derlying policy rules and discretionary deviationsfrom this path. The long-term impact of monetarypolicy is best summarized by comparing alternativerules that are fully understood by the public.

The most famous of these policy rules was intro-duced in the 1990s by John Taylor, who argued thata reaction function in which the short-term interestrate responded to movements of inflation from a de-sired value and to changes in the output gap (that is,the difference between actual output and its underly-ing trend) was a good summary of how U.S. mone-tary policy had been conducted (Taylor, 1993). This“Taylor rule” and variations that (for example) re-place current inflation by expected future inflationand/or add a term to take account of the fact thatcentral banks appear to smooth interest rate changes

14

–5

0

5

10

15

–5

0

5

10

0

10

20

30

–6

–4

–2

0

2

4

Euro area

Euro area

Euro areaEuro area

Other industrial countries Other industrial countries

Other industrial countries

Other industrial countries

Real GDP Consumption

Investment Real Aggregate Trade Balance/GDP

0 10 20 30 40 50Quarters

60 70 80 90 100

0 10 20 30 40 50Quarters

60 70 80 90 100

0 10 20 30 40 50Quarters

60 70 80 90 100

0 10 20 30 40 50Quarters

60 70 80 90 100

Figure 3.1. Dynamic Effects of More Competition-Friendly Policies in theEuro Area (Anticipated and Perfectly Credible)(Percent deviation from baseline)

Source: Bayoumi, Laxton, and Pesenti (2004).

Should Monetary Policy Rules Differ?

continue to form the basis for most analysis of mon-etary policy under flexible exchange rates.

Taylor also provided the framework for assess-ing alternative monetary rules based on the volatil-ity of inflation and output. Recognizing that theprimary objective of monetary policy is to providean anchor for inflation expectations, he argued it iscrucial for the central bank to react sufficientlystrongly to inflationary developments so as to raisereal interest rates and deflate the economy. A sec-ondary objective was to provide support to the realeconomy by reacting to cyclical developments.Within these constraints, the job of the monetarypolicymakers was twofold. First, to identify themost efficient monetary rules, in the sense that itprovided the lowest level of instability in (say) inflation for any given level of output volatility—the so-called Taylor efficiency frontier (Figure3.2). Second, choosing the preferred and most ro-bust rule from this frontier given their own prefer-ences between these two sources of macroeco-nomic instability.

This framework has been the workhorse for thelarge literature on monetary analysis for the lastdecade. The vast majority of the analysis has beenon industrial countries, and even within this mostwork has focused on closed economy models of theUnited States. It has encompassed everything fromthree-equation representations of monetary policy tolarge models such as the Federal Reserve’s FRB-US.In addition to examining the best options for a givenmodel, there has also been work on which rules pro-vide robust outcomes when parameters are uncertainor the analysis incorporates models with differenttheoretical structures. One general conclusion fromthis literature is that in large countries the monetaryauthorities should respond to both inflation and theoutput gap. Another is that the monetary authoritiesshould put more weight on inflation and less weighton the output gap as the country becomes more opento trade, but that there is little benefit from includingthe exchange rate in the policy rule. Finally, thereappear to be significant gains from smoothing inter-est rate changes over time, consistent with empiricalevidence that policymakers indulge in this practice.

GEM was used to examine how monetary rules insmall open emerging market countries might differfrom those for large industrial countries (Laxton andPesenti, 2003). That paper illustrates how the modelcan extend existing analyses through better measure-ment of concepts and benefits.2 A two-country ver-

sion of GEM was created consisting of the euro areaand the Czech Republic, with the euro area generat-ing 95 percent of total GDP. To capture the subtletiesof the Czech Republic’s relationship with the euroarea, which include importing components and reex-porting the finished product plus a trend apprecia-tion of the real exchange rate attributed at least inpart to Balassa-Samuelson effects, the model in-cluded trade in intermediate goods, traded and non-traded sectors, and distribution (see Figure 2.3).

Simulations to examine how these rules per-formed in the face of random shocks broadly corre-sponding to the historical record were then per-formed under a range of alternative monetary policyrules for the Czech Republic and the euro area. Theresults (discussed in more detail in Box 3.2) are asfollows:

• GEM broadly reproduces earlier results for theeuro area using a Taylor framework. This iscomforting as it implies that despite its strongtheoretical structure GEM is able to fit existingstylized facts for industrial countries in this well-researched field.

• The optimal monetary policy for the Czech Re-public depends crucially on how potential outputis measured. Using the conventional assumptionthat potential output is a slow-moving series,policymakers should only respond to inflation,and should ignore the output gap. This is mainlybecause of the greater importance of aggregatesupply shocks in emerging market countries,

15

2As in earlier models, monetary policy in GEM affects theeconomy through the impact of changes in the real interest rate onconsumers and producers, and depends on the interest sensitivityof spending as well as the impact on the exchange rate.

Policy rule

Inefficientoutcomes

Efficient frontier

Standard deviation of output

Standard deviation of inflation

Central bank preferences

Figure 3.2. Taylor Trade-Off in MonetaryPolicy Analysis


16

Box 3.2 Using GEM to Analyze Monetary Policy Rules

A two-country version of GEM was created compris-ing the euro area and the Czech Republic, with the euroarea generating 95 percent of total GDP (see Laxtonand Pesenti, 2003). To capture the subtleties of theCzech Republic’s relationship with the euro area,which include importing components and reexportingthe finished product as well as a trend appreciation at-tributed at least in part to Balassa-Samuelson effects,the model included trade in intermediate goods, tradedand nontraded sectors, and distribution (see Figure2.3). For the euro area, which approximates a closedeconomy, the trade-offs between alternative policyrules corresponded closely to those found by othersusing earlier large models (the figure illustrates thetrade-off coming from GEM). This is comforting, as itimplies that despite its strong theoretical structureGEM is able to fit existing stylized facts for industrialcountries in this well-researched field.

The main focus of the GEM work, however, was ex-amining policy rules for small emerging market coun-tries that are extremely open to trade, such as the CzechRepublic. This is an underresearched area that GEM isparticularly well designed to examine, as differences inunderlying economic structure can affect monetary re-sponses through the unified theoretical structure. Ascan be seen from the figure below, a rule that is robustfor a large and relatively closed economy such as theeuro area produces a high level of inflation variabilityin such an emerging market country. Further analysisindicates that a more robust rule for the emerging mar-ket country involves a greater focus on inflation.Hence, the Czech National Bank should put a muchhigher weight in its rule to responding to inflation and anegligible weight on the output gap (and, it turns out,the exchange rate). This work was recently extended toexamine the macroeconomic effects of EU accessioncountries entering EMU (see Schadler and others,2004). The paper concludes that adoption of the euromight be expected to have some macroeconomic costscompared with a well-designed monetary framework,but these effects have to be set against the microeco-nomic benefits associated with a single currency.

The results for the Czech Republic represent a magni-fied version of the conclusion from the existing literaturethat more open industrial economies should have rulesthat focus more on inflation. This amplification in GEMcomes from two important differences in economicstructure between the Czech Republic and industrialcountries. First, emerging markets such as the Czech Re-public have high levels of wage-price flexibility, so thatthere is less need for monetary policy to respond to ag-gregate demand shocks that move output temporarilyfrom potential. Second, such countries are subject to ahigher proportion of aggregate supply shocks. As theseshocks have consequences for the long-term path of out-

put, inflation gives a better signal to the monetary au-thorities than conventional measures of the output gapthat assume potential output adjusts slowly.

Further analysis using GEM indicates that Czechmonetary policy can be improved by using a more so-phisticated measure of the output gap. The theoreticalstructure in GEM allows the replacement of a conven-tional slow-moving measure of potential output by thelevel of output that would obtain if prices were fullyflexible but adjustment costs remain on real variablessuch as the capital stock (this would be much moredifficult in older models, as their structure does notprovide a clear distinction between real and nominalrigidities). GEM simulations indicate that if the Czechmonetary authorities could calculate this more sophis-ticated measure of the output gap then they should in-clude the output gap in the monetary rule. Two policymessages come out of this exercise. The first is theimportance of monetary policymakers rapidly taking aview on the sources of disturbances to the economy,particularly in emerging market countries that aremore subject to frequent supply disturbances. Second,because estimating the consequences of shocks on po-tential output is inherently relatively uncertain, small,open economies subject to large supply shocks shouldgenerally have monetary rules that focus more onchanges in inflation.

00

0.5

1.0

1.5

2.0

2.5

3.0

0.5 1.0 1.5Standard deviation of output

Standard deviation of CPI inflation

2.0 2.5 3.0

Large industrial country

Robust rule

Robust emerging market country rule

Robust large industrial country rule

Small emerging market country

Policy Rules for Large Industrial andEmerging Market Countries

Source: Laxton and Pesenti (2003).

The Impact of Higher Oil Prices

disturbances to which such a measure of the out-put gap provides a perverse signal for monetarypolicy. The output gap, however, becomes a muchmore useful indicator if the impact of aggregatesupply shocks on underlying supply potential isimmediately incorporated into the measured out-put gap (as can be done in GEM but not in earliermodels).

These results emphasize the importance of mone-tary policymakers rapidly taking a view on thesources of disturbances to the economy, particularlyin emerging market countries that are often more sub-ject to supply disturbances. That said, because esti-mating the consequences of shocks on potential out-put is inherently uncertain, small, open economiessubject to large supply shocks should generally put ahigher weight on inflation in evaluating the monetarystance than larger, more closed economies where ag-gregate demand disturbances are more prevalent.

Subsequent work using GEM has examined thewelfare-maximizing monetary rule. Unfortunately,the computation burdens implied by the needed so-lution techniques have constrained this analysis to asingle country that is closed to trade. Preliminary re-sults suggest that consumers’ utility is maximizedwhen the monetary authority responds to inflationand real activity, as in the conventional Taylor rule.However, they also suggest that policymakers shouldfocus mainly on the rate of change of the output gaprather than its level because of the uncertainties as-sociated with measuring the level of potential out-put. Hence, this approach has significant implica-tions for the form of the monetary rule. The resultsalso indicate that there are significant welfare bene-fits to adopting a sound monetary framework, incontrast to earlier work using models with fewertypes of distortions, although the gains are smallerthan those typically found from substantive changesin structural policies (see also Galí, Gertler, andLópez-Salido, 2002).


Oil prices continue to be a significant source ofvolatility for the global economy. Sustained move-ments in the dollar oil price of 10 percent or more,which trigger a new baseline for Fund analysis, re-main relatively common over short periods of time.New baselines have resulted in last-minute changesto the forecasts contained in the World EconomicOutlook on several occasions in recent years, aswell as affecting many other aspects of the Fund’swork, including program design. Oil is the onlycommodity to have such a systemic impact, al-though other commodities are important for indi-vidual countries. Indeed, while the dependence ofindustrial countries on oil has diminished some-what as manufacturing sectors have shrunk in pro-portion to the rest of the economy, the opposite hasoccurred in many emerging market countries. Inshort, understanding the impact of changes in oilprices on activity is a key input into multilateralsurveillance.

The Research Department published a study in2000 of the impact of oil prices on the global econ-omy, which included ready reckoners of the impactof a permanent $5 a barrel hike in oil prices on ac-tivity across a range of industrial and developingcountries after a year (IMF, 2000). The results foremerging markets, poor countries, and oil produc-ers came from the inputs of country desks, whilethose for the industrial countries were based onMULTIMOD simulations. These simulations incor-porated the effects of oil price hikes through anumber of channels. The impact on external bal-ances was fully integrated in the model, beingbased on data on oil trade, which is identified sepa-rately in the MULTIMOD database. The impact onpotential output, however, was implementedthrough changes to total factor productivity. Asshown in Table 3.2, the results suggested that a

17

Table 3.2. MULTIMOD: Impact of a Permanent $5 a Barrel Increase in Oil Prices After One Year(In percent)

Real GDP CPI Inflation Trade Balance1

United States –0.4 0.5 –0.1Euro area –0.4 0.5 –0.1Japan –0.2 0.2 –0.2Other industrial countries –0.2 — 0.2

Source: IMF (2000).1Percentage points of GDP.


18

–0.5

–0.4

–0.3

–0.2

–0.1

0

0

0.05

0.10

0.15

0.20

–0.4

0

0.4

0.8

UnitedStates

Euro area Japan UnitedKingdom

Canada

UnitedStates


Canada

UnitedStates


Canada

Real GDP (In percent)

CPI Inflation (In percent, year-on-year)

Current Account (In percent of GDP)

Figure 3.3. GEM: Impact of a Permanent20 Percent Oil Price Hike After One Year

Source: IMF staff calculations.

–0.5

–0.4

–0.3

–0.2

–0.1

0

0

0.05

0.10

0.15

0.20

–0.4

–0.2

0

0.2

0.4

0.6

0.8

UnitedStates


Canada

UnitedStates


Canada

UnitedStates


Canada

CPI Inflation (In percent, year-on-year)

Trade Balance (In percent of GDP)

Real GDP (In percent)

Permanent Temporary

Figure 3.4. GEM: Impact of a Permanentand Temporary 20 Percent Oil Price HikeAfter One Year(Temporary hike assumes half of oil price increase dissipates afterone year)

Source: IMF staff calculations.


permanent $5 hike in oil prices would lower outputafter a year by 0.4 percent in the United States andeuro area and 0.2 percent in Japan and other indus-trial countries, and would be accompanied by sig-nificant effects on inflation and the trade balance.

The GEM simulations of oil price hikes illustratethe advantages of having a flexible model structure.Rather than approximate the effect of an oil price hikethrough altering the level of productivity, a commod-ity submodel was constructed that is fully integratedinto the rest of GEM but can be turned on and off de-pending on the issue at hand. The commodity (here-after assumed to be oil) is produced using labor, capi-tal, and land, a separate factor of production thatexplains why production occurs in some places andnot others. Oil is then traded between countries andconsumed by firms and individuals, so oil distur-bances affect producers and consumers. Particular at-tention was placed on incorporating important ele-ments of the global oil market into the commoditymodel. The market power of the Organization of thePetroleum Exporting Countries (OPEC) was takeninto account by assuming that oil producers are mo-nopolistic competitors, so that oil price hikes can betriggered by an increase in market power owing togreater compliance of individual OPEC memberswith production quotas. The limited pass-through ofworld oil prices to domestic prices due to specifictaxes and other costs is modeled by assuming that oilpasses through a distribution sector before being usedby firms or consumers. Finally, while the long-run de-mand for oil and gas is quite sensitive to the real pricein the long term, it is assumed to be extremely costlyfor firms to adjust their oil use in the short run. Cali-bration used earlier analysis of the oil market, in par-ticular the long-term consequences of the oil pricehikes of 1974 and 1979.

The commodity model has been used to repeatthe earlier ready-reckoner exercise using GEM’smore integrated theoretical structure, as well as toproduce simulations of the impact of oil price hikeson U.S. growth reported in last year’s U.S. ArticleIV “Selected Issues” paper (IMF, 2003b). Figure3.3 suggests that a permanent oil price hike of 20percent (approximately equal to the $5 hike usedearlier) would reduce real output by some 0.4 per-cent for the United States and euro area and abouthalf of this for Japan in a version of GEM involving

the rest of the world and the United States, euroarea, Japan, the United Kingdom, and Canada, re-spectively. These differences across countries re-flect variations in oil and gas production, trade, andthe oil and gas intensity of production. In particu-lar, the relatively large output losses in NorthAmerica and small output losses in Japan reflectthe latter phenomenon. These output effects aresimilar to those produced using extensive judgmen-tal changes in MULTIMOD, but other responsesare quite different. For example, the inflationary re-sponse is lower in the GEM simulations, consistentwith the limited impact of oil price hikes on infla-tion in recent years, while the impact on the currentaccount is larger, reflecting firms’ inability to sub-stitute away from oil in the short run.

Additional analysis using GEM indicates that theimpact on output decreases rapidly as the priceshock becomes more temporary, while the impact onresponses of the current account and inflation is lessmarked. More specifically, when it is assumed thathalf of the initial shock to oil prices is eliminatedafter a year, the impact on output is about one-fifththat of a permanent disturbance, reflecting the richtheoretical structure. In particular, producers andconsumers feel less pressure to adjust knowing thatthe impact is not as permanent (Figure 3.4). The im-pact on inflation, however, is similar across the twoexperiments, as it is dominated by the pass-throughof the initial shock to oil prices into the consumerprice index (CPI). The deterioration in the currentaccount is somewhere between these two extremes,reflecting the fact that oil prices are significantlylower at the end of the first year in the temporarydisturbance compared with the permanent one. Thesmaller impact on real GDP helps to explain whytemporary spikes in oil prices, such as those that occurred over the 1990 Gulf war, appear to have hadrelatively little impact on global activity.

Thus far the model has been used to assess the im-pact of oil price hikes on industrial countries, in partbecause these effects have been the most heavily ex-amined in other work. However, the framework canclearly be used to examine the consequences ofchanges in oil or other commodity prices for othertypes of countries. In particular, the model can beused to look at the impact of oil or other commoditymarket disturbances on developing country producers.

19

GEM continues to be developed. Primarily, thisreflects the enormous amount of work cur-

rently being done in academia using new open econ-omy models to reexamine a range of issues, whoseinsights can provide ideas for how to improve GEM.In addition, as discussed earlier, the strong theoreti-cal structure makes it complicated to add new fea-tures. This section discusses active developmentwork on fiscal policy and international asset mar-kets, including emerging market financing con-straints. These areas have been chosen because oftheir relevance to the work of the IMF, in line withthe overall philosophy underlying GEM.

Fiscal Policy

Fiscal policy remains an important macroeco-nomic lever for stimulating the economy and hasbeen used actively in recent decades (see Figure4.1). The main issue associated with this is the de-gree of Ricardian equivalence. Full Ricardianequivalence implies that changes in taxes and trans-fers have no impact on aggregate demand (as is cur-rently the case in GEM). This is because consumersdiscount the future using the interest rates on gov-ernment paper, so the value of tax cuts and subse-quent tax increases offset each other, and peoplewill fully offset a tax cut by higher saving. There aretwo main ways of creating more realistic short-termtax multipliers. One is to assume that some individ-uals act as if they do not have access to financialmarkets, but rather vary their consumption in linewith their disposable income. Such rule-of-thumbconsumers can be easily incorporated into existingmodels and provide a way of examining income dis-tribution issues, but their behavior is highly me-chanical, responding as much to a temporary tax cutas to a long-term one. The alternative is to assumethat consumers have finite lives, adding a life-cycledimension to consumption. This provides more real-istic consumption dynamics, with spending re-sponding less to a temporary tax cut than a long-term one as predicted by the permanent incomehypothesis, but at the cost of adding considerable

theoretical complexity. In addition, the supply-sideeffects of fiscal policy can be incorporated byadding distorting taxes. The explicit modeling oflabor and product markets makes this an easy addi-tion in GEM, in contrast to earlier models.

The plan is to adopt a two-track approach to incor-porating fiscal policy into GEM. The main modelwill be altered to include distorting taxes and rule-of-thumb consumers, but not those with finite lives.Such a framework provides a reasonable way ofdealing with changes in fiscal balances resultingfrom disturbances elsewhere in the model by takingaccount of the impact of automatic stabilizers. Thesecond track involves developing an alternative ver-sion of the model that can be used to examine fiscalissues in more detail by incorporating finite-livedconsumers with a simplified version of the rest of theeconomy based on the existing GEM framework.

IV Current Development Work

20

–8

–6

–4

–2

0

2

1979 83 87 91 95 99 2003

United StatesJapan

Euro area

Figure 4.1. Structural Fiscal Balances in theMajor Economic Regions(General government as a percentage of potential output)

Source: IMF,World Economic Outlook.

Fiscal Policy

21

0100200300400500600700800

0

60

120

180

240

0

60

120

180

240

0

100

200

300

400

0

60

120

180

240

300

360

0

60

120

180

0

60

120

180

1980 84 88 92 96 2000 1980 84 88 92 96 2000

1980 84 88 92 96 2000 1980 84 88 92 96 2000

1980 84 88 92 96 2000

1980 84 88 92 96 2000

1980 84 88 92 96 2000

Portfolio investment Foreign direct investment Other

United States Japan

Germany France

Italy

United Kingdom

Canada

Figure 4.2. Sum of International Assets and Liabilities in Major AdvancedEconomies(In percent of GDP)

Source: IMF, Balance of Payments Statistics Yearbook.

IV CURRENT DEVELOPMENT WORK

Adding such consumers with higher discount ratesinto the current GEM is extremely difficult, as itslife-cycle implications are inconsistent with the as-sumption that consumer behavior can be calculatedfrom the actions of a single representative individ-ual, greatly complicating the theoretical structure.Early prototypes of both models have been created.

International Asset Markets

Gross holdings of other countries’ assets and lia-bilities have been rising rapidly across industrialcountries in recent years as financial deregulationhas reduced barriers to such transactions (Figure4.2).3 This provides a new mechanism for transmis-sion of the international cycle, as disturbances to fu-ture prosperity of domestic firms affect other coun-tries through equity prices, reducing idiosyncraticshocks across countries and increasing the synchro-nization of the global business cycle.

International models with strong theoreticalfoundations have generally assumed either that in-ternational asset markets are complete or that theyare limited to transactions in a single bond (Lane,2001). Complete markets imply that movements inconsumption across countries should be highly cor-related, responding little to country-specificchanges in domestic output, predictions that are sodifferent from the existing evidence that few policymodels have adopted this approach. Rather, theyhave tended to assume that one bond is the onlyasset traded across countries, which eliminates theneed to model demand across different assets. Thiswas the structure in MULTIMOD and in the currentversion of GEM. However, globalization of finan-cial markets is making this structure increasinglyproblematic.

The staff has been developing a prototype theoret-ical model in which all countries issue domestic debtand equity that can be traded in international mar-kets (similar work is also being undertaken by oth-ers). Different payment profiles create demand foreach asset and, it turns out, a higher rate of return forequities than bonds. Home bias in holding assets ismodeled by assuming that there is a cost to holdingeach asset, and these costs are higher for foreign as-sets than their domestic equivalents. These costs canalso be used to explain the inability of emergingmarkets to issue debt in domestic currency by as-suming that the costs of foreigners holding such debtis extremely high. Initial simulations indicate that

the addition of a wider range of assets produce morerealistic cross-country correlations of countries’ con-sumption. Once the properties of this prototypemodel have been more fully investigated, the nextstage will be to transfer the approach to GEM.

Another important issue in international assetmarkets is that for most emerging market countriesaccess is constrained, costly, and volatile (Figure4.3). In addition, access for specific countries oftenbecomes expensive or constrained just when they would normally want to borrow because ofshort-term domestic difficulties. Among other conse-quences, constrained access means that fiscal poli-cies become procyclical in economies where govern-ments are highly dependent on foreign borrowing.

22

3That said, typically only 5–10 percent of net wealth in themajor advanced economies is held in foreign assets (see IMF,2001).

0

400

800

1200

1600

–50

0

50

100

150

200

250

1998

1995 97 99 2001 03

99 2000 01 02 03

Total

EMBI Global EMBI Latin America

EMBI Asia

Asia

Latin America

Capital Flows (In billions of U.S. dollars)

Interest Rate Spreads (In percentage points)

Figure 4.3. Capital Constraints forEmerging Markets

Sources: J.P. Morgan; and IMF,World Economic Outlook.

International Asset Markets

There has been a large amount of recent work inthe new open economy macroeconomic and relatedliterature examining how to best characterize emerg-ing market borrowing constraints. Much of this workhas used the concept of the financial accelerator(Bernanke, Gertler, and Gilchrist, 1999). The cost ofborrowing for a firm is inversely related to its networth, so borrowing is increasingly difficult whenthe firm faces adverse shocks. 4 This provides a rea-

son why balance sheets matter for monetary policytransmission. It has been used to model domestic fi-nancial markets and banking systems, and has alsobeen transferred to the analysis of foreign borrowingby emerging market countries, with the main issuebeing how to define net worth. Work on adding a fi-nancial accelerator to GEM is proceeding, with aparticular focus on the consequences of limited ac-cess of emerging market countries to international fi-nancial assets.

23

4Other work has used the same basic structure, but has focusedon sudden stops in funding by assuming that countries can bor-row up to the value of its collateral but not beyond, so that there is

a change in behavior when the country hits its borrowing limit(Hart and Moore, 1994).

GEM is part of a burgeoning new open economymacroeconomics literature that is rapidly trans-

forming work in international macroeconomics andfinance. The strong microeconomic foundations ofthe new models provide an integrated way of com-bining aggregate supply, aggregate demand, nominalrigidities, trade, and asset markets in a single unifiedtheoretical framework. Using such models, policiescan be evaluated more satisfactorily by analyzingchanges in consumer welfare. In addition, becauseof the inclusion of a range of economic distortions,aggregate demand policies can have permanent effects on consumption, labor participation, and investment.

This approach is providing new insights on well-established policy issues such as the desirability ofalternative monetary policy rules. Even more inter-esting, simulations have examined questions such asthe impact of increasing competition in product andlabor markets that could previously only be analyzedin conjunction with costly and time-consuming mi-croeconomic studies. These new insights explainwhy a number of policy institutions are developingmodels based on strong microeconomic foundations.GEM is in the vanguard of multicountry policy mod-els being built using explicit microeconomic founda-tions, just as MULTIMOD was one of the first suchpolicy models to be built using rational expectations.

GEM is already generating useful simulationsacross a range of issues, but remains a project underdevelopment. In particular, the extensions discussedin Section IV of this paper will provide further depthwith regard to the modeling of fiscal policy and in-ternational asset markets, including financial fric-tions that have such an effect on emerging marketcountries. In addition, priority will be given to build-ing a three-country model comprising one large in-dustrial country plus some combination of additionalindustrial countries and/or emerging market coun-tries within GEM’s flexible structure. Such a modelwould allow a wide range of issues to be examinedin a unified underlying framework. At the same time,MULTIMOD will remain a useful tool for some pol-icy work, although its use will presumably declineover time.

GEM will continue to have a highly flexible struc-ture, in which the model will be adapted to the na-ture of the issue at hand. This flexibility is importantto help provide insights as to the underlying theoret-ical connections and avoid the model becoming toomuch of a “black box.” Indeed, one of the strengthsof a model with strong microeconomic foundationsis that it helps clarify the policy debate by ensuringthat the mechanisms at work are well articulated. Inaddition, such flexibility also makes it easier to addnew features to the model, ensuring that the projectremains relevant, up-to-date, and continues to garnerconsiderable interest from the rest of the world. Asan example, it would be useful to follow recent ad-vances in academia and move away from the as-sumption that goods are either always traded or notto a structure in which the choice of whether to ex-port a product depends on transportation costs andother characteristics of the good. In particular, suchan extension would allow the examination of macro-economic issues associated with trade reforms. Moregenerally, GEM refinements will continue to be dri-ven by developments in the broader literature and is-sues confronting the IMF.

A concerted effort is being made to make themodel accessible to those inside and outside of theFund. GEM is a large and complex model, and in-evitably it will require an effort for people to learnhow to use it. Significant resources have been putinto simplifying the software used to generate GEMsimulations. The modeling group has already pro-vided training to a group of IMF economists, in-cluding several in area departments, which will helpintegrate the model with bilateral surveillance activ-ities. Encouragingly, a number of the economistsbeing hired at the IMF have used new open econ-omy models in their doctoral work, and hence havea relatively strong background in model use. Out-side of the Fund, once the GEM’s structure has sta-bilized and its properties been more fully investi-gated, the code and programs to run it will beprovided free to those who wish to use the model.This approach was used successfully with MULTI-MOD, in that a range of outside groups used themodel for analysis.

V The Road Ahead

24

The Road Ahead

Given that GEM remains a project under devel-opment, it is difficult to look too far into the future.Two things, however, can be said about the futurewith some certainty. First, the new open economymacroeconomics literature will continue to expand

rapidly. Second, GEM provides a vehicle for ab-sorbing relevant insights from this literature intothe Fund, as well as encouraging mutually benefi-cial interactions between the academic and policycommunities.

25

Altig, D., L. Christiano, M. Eichenbaum, and J. Linde,2003, “Technology Shocks and Aggregate Fluctua-tions” (unpublished).

Balassa, Bela, 1964, “The Purchasing-Power Parity Doc-trine: A Reappraisal,” Journal of Political Economy,Vol. 72, pp. 584–96.

Bayoumi, Tamim, Douglas Laxton, and Paolo Pesenti,2004, “Benefits and Spillovers of Greater Competi-tion in Europe: A Macroeconomic Assessment,” ECBWorking Paper No. 341 (Frankfurt: European CentralBank).

Bernanke, Ben, Mark Gertler, and Simon Gilchrist, 1999,“The Financial Accelerator in a Quantitative BusinessCycle,” in Handbook of Macroeconomics, ed. by JohnB. Taylor and Michael Woodford (Amsterdam, NewYork, and Oxford: Elsevier Science, North-Holland).

Christiano, Lawrence J., Martin Eichenbaum, and CharlesEvans, 2001, “Nominal Rigidities and the DynamicEffects of a Shock to Monetary Policy,” NBER Work-ing Paper No. 8403 (Cambridge, Massachusetts: Na-tional Bureau of Economic Research).

Corsetti, Giancarlo, and Luca Dedola, 2002, “Macroeco-nomics of International Price Discrimination,” Temide Discussione, No. 461 (Rome: Bank of Italy).

Dornbusch, Rudiger, 1976, “Expectations and ExchangeRate Dynamics,” Journal of Political Economy, Vol.84, No. 6, pp. 1161–76.

Erceg, C., L. Guerrieri, and C. Gust, 2003, “SIGMA: ANew Open Economy Model for Policy Analysis,”draft paper prepared for the Annual Central BankModelers’ Workshop in Amsterdam.

Fleming, J. Marcus, 1962, “Domestic Financial PoliciesUnder Fixed and Under Floating Exchange Rates,”Staff Papers, International Monetary Fund, Vol. 9(November) , pp. 369–79.

Galí, J., M. Gertler, and J. D. López-Salido, 2002,“Markups, Gaps, and the Welfare Costs of BusinessFluctuations,” NBER Working Paper No. 8850 (Cam-bridge, Massachusetts: National Bureau of EconomicResearch).

Hall, Robert E., 1978, “Stochastic Implications of the LifeCycle-Permanent Income Hypothesis: Theory andEvidence,” Journal of Political Economy, Vol. 86, No.6, pp. 971–87.

Hart, Oliver, and John Moore, 1994, “A Theory of DebtBased on the Inalienability of Human Capital,” Quar-terly Journal of Economics, Vol. 109, No. 4, pp.841–79.

Hunt, Benjamin L., and Alessandro Rebucci, 2003, “TheU.S. Dollar and the Trade Deficit: What Accounts forthe Late 1990s?” IMF Working Paper No. 03/194(Washington: International Monetary Fund).

International Monetary Fund, 2000, “The Impact ofHigher Oil Prices on the Global Economy;” availablevia the Internet at http://www.imf.org/external/pubs/ft/oil/2000/oilrep.pdf.

———, 2001, “Business Cycle Linkages Among MajorAdvanced Economies,” World Economic Outlook,October 2001, World Economic and Financial Sur-veys (Washington: International Monetary Fund).

———, 2003a, “How Concerned Should DevelopingCountries Be About G-3 Exchange Rate Volatility?”World Economic Outlook, September 2003, WorldEconomic and Financial Surveys (Washington: Inter-national Monetary Fund).

———, 2003b, “United States: Selected Issues,” IMFCountry Report No. 03/245 (Washington: Interna-tional Monetary Fund).

Kollman, Robert, 2002, “Monetary Policy Rules in theOpen Economy: Effects on Welfare and Business Cy-cles,” CEPR Discussion Paper No. 3279 (London:Centre for Economic Policy Research).

Kydland, Finn E., and Edward C. Prescott, 1982, “Time toBuild and Aggregate Fluctuations,” Econometrica,Vol. 50 (November), pp. 1345–70.

Lane, Philip R., 2001, “The New Open Economy Macro-economics: A Survey,” Journal of International Eco-nomics, Vol. 54 (August), pp. 235–66.

Laxton, Douglas, and Paolo Pesenti, 2003, “MonetaryRules for Small, Open, Emerging Economies,”Journal of Monetary Economics, Vol. 50, No. 5,pp. 1109–46.

Lucas, Robert E., Jr., 1976, “Econometric Policy Evalua-tion: A Critique,” Journal of Monetary Economics,Vol. 1, No. 2, pp. 19–46.

Masson, Paul R., Steven A. Symansky, and Guy Meredith,1990, MULTIMOD Mark II: A Revised and ExtendedModel, IMF Occasional Paper No. 71 (Washington:International Monetary Fund).

Mendoza, Enrique G., 1991, “Real Business Cycles in aSmall Open Economy,” American Economic Review,Vol. 81 (September), pp. 797–818.

Mundell, Robert A., 1963, “Capital Mobility and Stabi-lization Policy Under Fixed and Flexible ExchangeRates,” Canadian Journal of Economics and PoliticalScience, Vol. 29, pp. 475–85.

References

26

References

Muth, J.F., 1961, “Rational Expectations and the Theory of Price Movements,” Econometrica, Vol. 29,pp. 315–35.

Obstfeld, Maurice, 2001, “International Macroeconomics:Beyond the Mundell-Fleming Model,” IMF Staff Pa-pers, Special Issue, Vol. 47, pp. 1–39.

———, and Kenneth Rogoff, 1995, “Exchange Rate Dy-namics Redux,” Journal of Political Economy, Vol.103, pp. 624–60.

Organization for Economic Cooperation and Develop-ment, 1997, The OECD Report on Regulatory Re-form, Vol. II: Thematic Studies (Paris: Organizationfor Economic Cooperation and Development).

Samuelson, Paul A., 1964, “Theoretical Notes on TradeProblems,” Review of Economics and Statistics, Vol.46, pp. 145–54.

Schadler, Susan M., Paulo F. Drummond, Louis Kuijs,Zuzana Murgasova, and Rachel N. van Elkan, 2004,Adopting the Euro in Central Europe: Challenges ofthe Next Step in European Integration, IMF Oca-sional Paper No. 234 (Washington: InternationalMonetary Fund; forthcoming).

Smets, Frank, and Raf Wouters, 2002, “An Estimated Sto-chastic Dynamic General Equilibrium Model of theEuro Area,” ECB Working Paper No. 171 (Frankfurt:European Central Bank).

Taylor, John, 1993, “Discretion Versus Policy Rules inPractice,” Carnegie-Rochester Conference Series onPublic Policy, Vol. 39 (December), pp. 195–214.

27

OCCASIONAL PAPERS

Recent Occasional Papers of the International Monetary Fund

239. GEM: A New International Macroeconomic Model, by Tamim Bayoumi, with assistance from DouglasLaxton, Hamid Faruqee, Benjamin Hunt, Philippe Karam, Jaewoo Lee, Alessandro Rebucci, and IvanTchakarov. 2004.

238. Stabilization and Reforms in Latin America: A Macroeconomic Perspective on the Experience Since theEarly 1990s, by Anoop Singh, Agnès Belaisch, Charles Collyns, Paula De Masi, Reva Krieger, GuyMeredith, and Robert Rennhack. 2004.

237. Sovereign Debt Structure for Crisis Prevention, by Eduardo Borensztein, Marcos Chamon, Olivier Jeanne,Paolo Mauro, and Jeromin Zettelmeyer. 2004.

236. Lessons from the Crisis in Argentina, by Christina Daseking, Atish R. Ghosh, Alun Thomas, and TimothyLane. 2004.

235. A New Look at Exchange Rate Volatility and Trade Flows, by Peter B. Clark, Natalia Tamirisa, andShang-Jin Wei, with Azim Sadikov and Li Zeng. 2004.

234. Adopting the Euro in Central Europe: Challenges of the Next Step in European Integration, by SusanSchadler, Paulo Drummond, Louis Kuijs, Zuzana Murgasova, and Rachel van Elkan. 2004.

233. Germany’s Three-Pillar Banking System: Cross-Country Perspectives in Europe, by Allan Brunner, JörgDecressin, Daniel Hardy, and Beata Kudela. 2004.

232. China’s Growth and Integration into the World Economy: Prospects and Challenges, edited by EswarPrasad. 2004.

231. Chile: Policies and Institutions Underpinning Stability and Growth, by Eliot Kalter, Steven Phillips,Marco A. Espinosa-Vega, Rodolfo Luzio, Mauricio Villafuerte, and Manmohan Singh. 2004.

230. Financial Stability in Dollarized Countries, by Anne-Marie Gulde, David Hoelscher, Alain Ize, DavidMarston, and Gianni De Nicoló. 2004.

229. Evolution and Performance of Exchange Rate Regimes, by Kenneth S. Rogoff, Aasim M. Husain, AshokaMody, Robin Brooks, and Nienke Oomes. 2004.

228. Capital Markets and Financial Intermediation in The Baltics, by Alfred Schipke, Christian Beddies, SusanM. George, and Niamh Sheridan. 2004.

227. U.S. Fiscal Policies and Priorities for Long-Run Sustainability, edited by Martin Mühleisen and Christo-pher Towe. 2004.

226. Hong Kong SAR: Meeting the Challenges of Integration with the Mainland, edited by Eswar Prasad, withcontributions from Jorge Chan-Lau, Dora Iakova, William Lee, Hong Liang, Ida Liu, Papa N’Diaye,and Tao Wang. 2004.

225. Rules-Based Fiscal Policy in France, Germany, Italy, and Spain, by Teresa Dában, Enrica Detragiache,Gabriel di Bella, Gian Maria Milesi-Ferretti, and Steven Symansky. 2003.

224. Managing Systemic Banking Crises, by a staff team led by David S. Hoelscher and Marc Quintyn. 2003.

223. Monetary Union Among Member Countries of the Gulf Cooperation Council, by a staff team led by UgoFasano. 2003.

222. Informal Funds Transfer Systems: An Analysis of the Informal Hawala System, by Mohammed El Qorchi,Samuel Munzele Maimbo, and John F. Wilson. 2003.

221. Deflation: Determinants, Risks, and Policy Options, by Manmohan S. Kumar. 2003.

220. Effects of Financial Globalization on Developing Countries: Some Empirical Evidence, by Eswar S.Prasad, Kenneth Rogoff, Shang-Jin Wei, and Ayhan Kose. 2003.

219. Economic Policy in a Highly Dollarized Economy: The Case of Cambodia, by Mario de Zamaroczy andSopanha Sa. 2003.

218. Fiscal Vulnerability and Financial Crises in Emerging Market Economies, by Richard Hemming, MichaelKell, and Axel Schimmelpfennig. 2003.

217. Managing Financial Crises: Recent Experience and Lessons for Latin America, edited by Charles Collynsand G. Russell Kincaid. 2003.

28

Occasional Papers

216. Is the PRGF Living Up to Expectations?—An Assessment of Program Design, by Sanjeev Gupta, MarkPlant, Benedict Clements, Thomas Dorsey, Emanuele Baldacci, Gabriela Inchauste, Shamsuddin Tareq,and Nita Thacker. 2002.

215. Improving Large Taxpayers’ Compliance: A Review of Country Experience, by Katherine Baer. 2002.

214. Advanced Country Experiences with Capital Account Liberalization, by Age Bakker and Bryan Chapple.2002.

213. The Baltic Countries: Medium-Term Fiscal Issues Related to EU and NATO Accession, by JohannesMueller, Christian Beddies, Robert Burgess, Vitali Kramarenko, and Joannes Mongardini. 2002.

212. Financial Soundness Indicators: Analytical Aspects and Country Practices, by V. Sundararajan, CharlesEnoch, Armida San José, Paul Hilbers, Russell Krueger, Marina Moretti, and Graham Slack. 2002.

211. Capital Account Liberalization and Financial Sector Stability, by a staff team led by Shogo Ishii and KarlHabermeier. 2002.

210. IMF-Supported Programs in Capital Account Crises, by Atish Ghosh, Timothy Lane, Marianne Schulze-Ghattas, Alesv Bulírv, Javier Hamann, and Alex Mourmouras. 2002.

209. Methodology for Current Account and Exchange Rate Assessments, by Peter Isard, Hamid Faruqee,G. Russell Kincaid, and Martin Fetherston. 2001.

208. Yemen in the 1990s: From Unification to Economic Reform, by Klaus Enders, Sherwyn Williams, NadaChoueiri, Yuri Sobolev, and Jan Walliser. 2001.

207. Malaysia: From Crisis to Recovery, by Kanitta Meesook, Il Houng Lee, Olin Liu, Yougesh Khatri, NataliaTamirisa, Michael Moore, and Mark H. Krysl. 2001.

206. The Dominican Republic: Stabilization, Structural Reform, and Economic Growth, by a staff team led byPhilip Young comprising Alessandro Giustiniani, Werner C. Keller, and Randa E. Sab and others. 2001.

205. Stabilization and Savings Funds for Nonrenewable Resources, by Jeffrey Davis, Rolando Ossowski,James Daniel, and Steven Barnett. 2001.

204. Monetary Union in West Africa (ECOWAS): Is It Desirable and How Could It Be Achieved? by PaulMasson and Catherine Pattillo. 2001.

203. Modern Banking and OTC Derivatives Markets: The Transformation of Global Finance and Its Implicationsfor Systemic Risk, by Garry J. Schinasi, R. Sean Craig, Burkhard Drees, and Charles Kramer. 2000.

202. Adopting Inflation Targeting: Practical Issues for Emerging Market Countries, by Andrea Schaechter,Mark R. Stone, and Mark Zelmer. 2000.

201. Developments and Challenges in the Caribbean Region, by Samuel Itam, Simon Cueva, Erik Lundback,Janet Stotsky, and Stephen Tokarick. 2000.

200. Pension Reform in the Baltics: Issues and Prospects, by Jerald Schiff, Niko Hobdari, Axel Schimmel-pfennig, and Roman Zytek. 2000.

199. Ghana: Economic Development in a Democratic Environment, by Sérgio Pereira Leite, AnthonyPellechio, Luisa Zanforlin, Girma Begashaw, Stefania Fabrizio, and Joachim Harnack. 2000.

198. Setting Up Treasuries in the Baltics, Russia, and Other Countries of the Former Soviet Union: An Assess-ment of IMF Technical Assistance, by Barry H. Potter and Jack Diamond. 2000.

197. Deposit Insurance: Actual and Good Practices, by Gillian G.H. Garcia. 2000.

196. Trade and Trade Policies in Eastern and Southern Africa, by a staff team led by Arvind Subramanian, withEnrique Gelbard, Richard Harmsen, Katrin Elborgh-Woytek, and Piroska Nagy. 2000.

195. The Eastern Caribbean Currency Union—Institutions, Performance, and Policy Issues, by Frits van Beek,José Roberto Rosales, Mayra Zermeño, Ruby Randall, and Jorge Shepherd. 2000.

194. Fiscal and Macroeconomic Impact of Privatization, by Jeffrey Davis, Rolando Ossowski, ThomasRichardson, and Steven Barnett. 2000.

193. Exchange Rate Regimes in an Increasingly Integrated World Economy, by Michael Mussa, Paul Masson,Alexander Swoboda, Esteban Jadresic, Paolo Mauro, and Andy Berg. 2000.

Note: For information on the titles and availability of Occasional Papers not listed, please consult the IMF’s Publications Catalog or contact IMFPublication Services.

29

GEM: A New International Macroeconomic Model

Documents

Transcript of GEM: A New International Macroeconomic Model