Andres Velasco Economia

8/17/2019 Andres Velasco Economia

1/257

Volume 5 Number 2

200Spring

Journal of the Latin American and Caribbean Economic Association

Lora and Olivera on Electoral

Consequences of the WashingtonConsensus

Suominen and Estevadeordal

on Rules of Origin in Trade Agreements

Majnoni and Powell on Basel II

and Bank Capital Requirements

Echeverry, Ibáñez, Moya,

and Hillón on Reforming Public

Transport in Bogotá

Kaplan, Martínez González,

and Robertson on Wages afterDisplacement


2/257

E D I T O R

Andrés Velasco

LATIN AMERICAN AND CARIBBEAN

ECONOMIC ASSOCIATION

BROOKINGS INSTITUTION PRESS

Washington, D.C.

2005

Volume 5 Number 2 economíaSpring



3/257

Articles in this publication were developed by the authors for the biannual Economía meetings.

In all cases the papers are the product of the authors’ thinking alone and do not imply endorse-

ment by the staff members, officers, or trustees of the Brookings Institution or of LACEA, or of

those institutions with which the authors are affiliated.

Copyright © 2005

LATIN AMERICAN AND CARIBBEAN ECONOMIC ASSOCIATION

www.lacea.org

Published by

BROOKINGS INSTITUTION PRESS

1775 Massachusetts Avenue, N.W., Washington, DC 20036

www.brookings.edu

ISSN 1529-7470

ISBN 0-8157-2079-3

For information on subscriptions, standing orders, and individual copies, contact Brookings Insti-

tution Press, P.O. Box 465, Hanover, PA 17331-0465. Or call 866-698-0010. E-mail brookings

@tsp.sheridan.com. Visit Brookings online at http://bookstore.brookings.edu.

Brookings periodicals are available online through Online Computer Library Center (contact the

OCLC subscriptions department at 800/848-5878, ext. 6251) and Project Muse (http://muse.jhu.edu).

Authorization to photocopy items for internal or personal use or the internal or personal use of specific clientsis granted by the Brookings Institution for libraries and other users registered with the Copyright ClearanceCenter Transactional Reporting Service, provided that the basic fee is paid to the Copyright Clearance Center,222 Rosewood Drive, Danvers, MA 01923. For more information, please contact CCC at 508/750-8400. Thisauthorization does not extend to other kinds of copying, such as copying for general distribution, or for creatingnew collective works, or for sale. Specific written permission for such copying must be obtained from the Per-missions Department, Brookings Institution Press, 1775 Massachusetts Avenue, N.W., Washington, DC 20036,fax 202/797-6195; e-mail [email protected].


4/257

Editor’s Summary vii

EDUARDO LORA and MAURICIO OLIVERA

The Electoral Consequences of the Washington Consensus 1Comments by Sebastián Galiani and Ernesto Dal Bó 46

ANTONI ESTEVADEORDAL and KATI SUOMINEN

Rules of Origin in Preferential Trading Arrangements: Is All Well with the Spaghetti Bowl in the Americas? 63Comments by Pablo Sanguinetti and Alberto Trejos 93

GIOVANNI MAJNONI and ANDREW POWELL

Reforming Bank Capital Requirements: Implications of Basel II for Latin American Countries 105Comments by Patricia Correa and Philip Brock 141

JUAN CARLOS ECHEVERRY, ANA MARÍA IBÁÑEZ,

ANDRÉS MOYA, and LUIS CARLOS HILLÓN

The Economics of TransMilenio,a Mass Transit System for Bogotá 151Comments by Mauricio Cárdenas and Andrés Gómez-Lobo 189

DAVID S. KAPLAN, GABRIEL MARTÍNEZ GONZÁLEZ,

and RAYMOND ROBERTSON

What Happens to Wages after Displacement? 197 Comments by Naércio Menezes-Filho and Omar Arias 235

2005

Volume 5 Number 2 economíaSpring



5/257

LATIN AMERICAN AND CARIBBEAN ECONOMIC ASSOCIATION

The Latin American and Caribbean Economic Association (LACEA), or

Asociación de Economía de América Latina y el Caribe, is an internationalassociation of economists with common research interests in Latin

America. It was formed in 1992 to facilitate the exchange of ideas among

economists and policymakers.

Membership in LACEA is open to all individuals or institutions

professionally concerned with the study of Latin American and Caribbean

economies. For membership information, please visit the LACEA website

at www.lacea.org and click on “join LACEA.”

O F F I C E R S

President

Mariano Tommasi, Universidad de San Andrés, Argentina

Vice President

Andrés Velasco, Harvard University and Universidad de Chile

Past Presidents

Sebastián Edwards, University of California–Los AngelesGuillermo Calvo, Inter-American Development Bank and University of Maryland

Nora Lustig, Universidad de las Américas, Puebla

Albert Fishlow, Columbia University

Secretary

Ariel Fiszbein, World Bank

Treasurer

Sergio Schmukler, World Bank

E X E C U T I V E C O M M I T T E E

Nancy Birdsall, Center for Global Development

François Bourguignon, World Bank

Raquel Fernández, New York University

Francisco H. G. Ferreira, Pontifícia Universidade Católica, Rio de Janeiro, and

World Bank

Nora Lustig, Universidad de las Américas, PueblaJosé A. Ocampo, United Nations

Guillermo Perry, World Bank

Carola Pessino, Universidad Torcuato Di Tella, Argentina


6/257

Carmen Reinhart, University of Maryland

Andrés Rodríquez-Clare, Inter-American Development Bank

Jaime Saavedra, World Bank

Cristina T. Terra, Fundação Getúlio Vargas

ECONOMIA

Editor

Andrés Velasco, Harvard University and Universidad de Chile

Editorial Associate

Jennifer Hoover

Managing Editor

Magdalena Balcells

Editorial Board

Rafael Di Tella, Harvard University

Eduardo Engel, Yale University

Francisco H. G. Ferreira, Pontifícia Universidade Católica, Rio de Janeiro, and

World Bank

Carmen Pagés, World Bank

Roberto Rigobón, Massachusetts Institute of Technology

Andrés Rodríguez-Clare, Inter-American Development Bank

Roberto Steiner, International Monetary Fund

Miguel Urquiola, Columbia University

Sebastián Edwards (ex officio), University of California–Los Angeles

Mariano Tommasi (ex officio), Universidad de San Andrés, Argentina

ECONOMIA PANEL FOR VOLUME 5

Mauricio Cárdenas, Fedesarrollo-Colombia

Gerardo Esquivel Hernández, Colegio de México

Ronald Fischer, Universidad de Chile

Ilan Goldfajn, Pontifícia Universidade Católica, Rio de Janeiro

Eduardo Levy Yeyati, Universidad Torcuato Di Tella, Argentina

María Soledad Martínez Pería, World Bank

Francisco Rodríguez, Instituto de Estudios Superiores de Administración,

VenezuelaNouriel Roubini, New York University

Jaime Saavedra, World Bank

Ernesto Schagrodsky, Universidad Torcuato Di Tella, Argentina


7/257

AUTHORS AND DISCUSSANTS

Omar Arias, World Bank

Philip Brock, University of WashingtonMauricio Cárdenas, Fedesarrollo-Colombia

Patricia Correa, Super Intendencia de Bancos de Colombia

Ernesto Dal Bó, University of California–Berkeley

Juan Carlos Echeverry, Universidad de los Andes

Antoni Estevadeordal, Inter-American Development Bank

Sebastián Galiani, Universidad de San Andrés

Andrés Gómez-Lobo, Universidad de Chile

Luis Carlos Hillón, Universidad de los Andes

Ana María Ibáñez, Universidad de los Andes

David S. Kaplan, Instituto Tecnológico Autónomo de México

Eduardo Lora, Inter-American Development Bank

Giovanni Majnoni, World Bank

Gabriel Martínez González, Inter-American Conference on Social Security

Naércio Menezes-Filho, University of São Paulo

Andrés Moya, Universidad de los Andes

Mauricio Olivera, Inter-American Development Bank

Andrew Powell, Universidad Torcuato Di Tella

Raymond Robertson, Macalester College

Pablo Sanguinetti, Universidad Torcuato Di Tella

Kati Suominen, Inter-American Development Bank

Alberto Trejos, Instituto Centroamericano de Administracion de Empresas INCAE


8/257

Editor’s Summary

Nearly two decades after a wave of policy changes swept through Latin

America, economic reforms continue to be the focus of much dis-

cussion. Critics claim that the promarket reforms have failed todeliver economic growth, and that the time has come to try something else.

Advocates claim that the reforms were never given a fair chance—too lit-

tle was done, often too late. Complete the reform process, they claim, and

growth will come.

Both sides do agree on one point: Latin America seems to be suffering

from reform fatigue, and another wave of reforms is unlikely to happen any

time soon. Certainly not in countries led by left-leaning populists, such as

Argentina’s Néstor Kirchner or Venezuela’s Hugo Chávez. The reformmomentum has even stalled in countries led by promarket conservatives—

Mexico under Vicente Fox and Colombia under Alvaro Uribe are two

examples. If such reforms are now unpopular in many quarters, did the

politicians who initially adopted them bear a political cost? Was the Wash-

ington Consensus electorally bad for friends of Washington? That is the

question studied by Eduardo Lora and Mauricio Olivera in the lead article

of this, the tenth issue of Economía.

Lora and Olivera analyze the outcome of sixty-six presidential elec-tions and eighty-one parliamentary elections in seventeen Latin American

countries from 1985 to 2002. Their general conclusion is striking: reform-

ing parties and politicians were rewarded electorally only when reforms

involved macroeconomic stabilization and a sharp reduction in inflation;

otherwise, their reforming zeal cost them dearly at the polls. Economic out-

comes do matter for electoral outcomes. Lora and Olivera find that the

incumbent’s party is rewarded in presidential elections for reductions in the

inflation rate and in legislative elections for increases in the growth rate.

Changes in unemployment and income distribution, however, do not appear

to influence voters’ behavior.

vii

A N D R É S V E L A S C O


9/257

What is even more surprising is that, at the polls, policies matter irre-

spective of their results—that is, their effects on growth or inflation. Elec-

torates seem not to like reform policies of the kind applied in Latin

America in the 1990s. In a regression with electoral outcomes on the right

hand side, reform indexes have a negative and significant effect, even when

the authors control for changes in inflation and growth. The point esti-

mate of the effect of policies on electoral results implies that the incum-

bent’s party typically lost 15 percent of its vote in presidential elections on

account of the average amount of promarket reforms introduced during

its term. More aggressive reformers (say, those reforming one standard

deviation above the mean) sacrificed 27 percent of their vote on account of promarket reforms. Statistically, this seems to be a very robust result for

presidential elections.1

Moreover, lying about one’s true intentions does not seem to be a good

electoral strategy. Several Latin American politicians—including Fujimori

in Peru, Menem in Argentina, and more recently Gutiérrez in Ecuador—

first ran as opponents of the Washington Consensus, then followed ortho-

dox policies. The paper shows that a candidate that said one thing on tax

policy and then did another was, on average, punished more severely atthe polls. Campaign promises do not seem to matter for the effect of other

policies on voting behavior.

These results raise two kinds of questions. For academics, the issue is

why inputs (policies) matter and not just outputs. Is it ideology, pure and

simple? Or is it that because outcomes represent an extremely noisy signal

of politicians’ competence, the choice of policies conveys some informa-

tion that voters find useful? For policymakers, the question is political:

what has to change in Latin America before ambitious reforms become fea-sible again? Are all large-scale reforms out of the question, or only those

that carry the Washington Consensus label? Both sets of questions remain

very much open.

The unpopularity of the reforms does not mean, however, that policy is

frozen everywhere. Trade is one area in which reform has not stopped dead

viii E C O N O M I A , Spring 2005

1. The total effect of reforms on electoral outcomes is the sum of two effects: a direct

effect that runs from policies to votes and an indirect effect that runs from policies to eco-

nomic outcomes to votes. The first is typically large and negative, while the second is pos-itive insofar as the reforms lowered inflation and stimulated growth. The figures given

correspond to the total effect—that is, after the positive indirect effects have been taken into

account. The direct negative effects are much larger. See the paper for details.


10/257

on its tracks. Liberalization agreements, of both the bilateral and regional

kind, continue to be signed, though at a less frenzied pace than a decade

ago. Nearly fifty deals have been forged in the Americas since 1990s. But

this veritable “spaghetti bowl” of overlapping and sometimes contradictory

agreements has costs as well as benefits. One cost, studied by Antoni Este-

vadeordal and Kati Suominen in the second paper of this issue, results from

the rules of origin applied.

At heart, the matter is simple: if Brazil gives Paraguay preferential

access to its market, Brazilian policymakers want to make sure that the new

imports entering Brazil are, in fact, made in Paraguay and not in a third

country attempting to benefit from Paraguay’s preferential status. But whatsounds simple in theory becomes devilishly complicated in practice. What,

precisely, is a Paraguayan good? Goods often have imported inputs, and in

few or no items is 100 percent of value added likely to originate in

Paraguay. Where, then, should a country draw the line?

Rules of origin attempt to settle the issue, but in doing so they face many

pitfalls. If the set of rules is too stringent, then the bulk of Paraguayan

goods may be left out of the Brazilian market. Indeed, such rules can be

used as protectionist devices that effectively undercut trade preferences andcontradict the avowed liberalizing intent of free trade agreements. Another

problem is that rules of origin are almost inevitably complex (the paper

identifies several kinds, each with its own subcategories). Applying them

can be very costly, especially for the poorer economies in the region.

Estevadeordal and Suominen offer three main conclusions. First, putting

stringent rules of origin into an agreement makes it more politically feasi-

ble, since the rules can be used as a tool to pay off protectionist interests.

Second, there is evidence that restrictive rules of origin undercut the liber-alizing potential of free trade agreements. NAFTA is a particularly egre-

gious example of this, with many Mexican goods subjected to rules that

verge on ludicrous. It is unfortunate, therefore—and this is the third con-

clusion of the paper—that the NAFTA model of rules of origin is increas-

ingly being used in other agreements in the region. This does not bode

well for free trade in the Americas.

Not all is lost, however. NAFTA-type rules are at least precise, and they

leave less room for arbitrary application than do other types of rules of

origin used in earlier agreements. Moreover, the growing homogeneity of

rules that follow the NAFTA model simplifies the life of customs officials

and lowers transaction costs. Last, and most important, the NAFTA rules of

Andrés Velasco ix


11/257

origin have what trade experts call lenient facilitation devices. In English,

this means that the rules themselves include ways to reduce their restric-

tiveness. A key aspect is diagonal cumulation, which allows countries tied

by the same set of origin rules to use products that originate in any part of

the common rules-of-origin zone as if they originated in the exporting

country. Therefore, argue the authors, the rules of origin in a future Free

Trade Area of the Americas—if one ever materializes—should not be all

that restrictive. One can only hope they are right.

Financial regulation is another area in which policy is changing, as a

result of both internal needs and international changes in standards. The

1988 Basel Accord on bank capital—the so-called Basel I agreement—isnow in place throughout the region, and discussion has shifted to whether

and how Latin American countries should apply Basel II. It is widely

accepted that bank capital ought to be regulated, but how to do so remains

open to debate. The simple approach of Basel I divides assets into very

broad risk categories and establishes an 8 percent minimum capital require-

ment for risky assets. The potential for arbitraging one’s way around this

simple rule has grown, however, as risk management becomes more

sophisticated. In response, Basel II goes well beyond simple quantitativerequirements, proposing two basic approaches: the standardized approach,

which uses external credit rating agencies together with a table that maps

those ratings directly into capital requirements; and the internal ratings-

based (IRB) approach, in which the banks themselves estimate their cus-

tomers’ default probability (without relying on external rating agencies) and

then use a particular formula to determine capital requirements as a func-

tion of the estimated default probability.

The third paper in this issue, by Giovanni Majnoni and Andrew Powell,focuses on a key aspect of Basel II application. Many emerging markets

do not have many (or any) external rating agencies, so the standardized

approach may not be applicable. The internal ratings-based approach, in

turn, is complex, and its application and supervision may stretch limited

supervisory resources. Majnoni and Powell suggest a simplification of the

IRB approach that could be used as a transition arrangement. In their

centralized ratings-based (CRB) approach, banks would rate their clients,

but the regulator would determine the rating scale and the way in which

the banks’ ratings map into default probabilities. Using a centralized scale

would facilitate comparison across banks and greatly ease the monitor-

ing of banks’ ratings. Those requirements would also be easier to monitor,

x E C O N O M I A , Spring 2005


12/257

since the regulator would determine how banks’ ratings feed into capital

requirements.

The hard part of the approach is deciding what kinds of standards the

regulator should apply, since what works in rich countries may not work

in emerging economies. Basel II’s IRB approach suggests a formula for cal-

culating a bank’s capital requirement as a function of three basic variables:

default probability, exposure at default, and loss given default. A regulator

might then ask a bank to hold provisions and capital to cover a specified

percentage of the distribution of losses to ensure the continued solvency

of the bank except in highly extraordinary circumstances. The calibration of

the Basel II IRB formula uses a value at risk of 99.9 percent with a hori-zon of one year—that is, a bank is only expected to use up its capital in

one year with a probability of 0.1 percent, or once every 1,000 years.

Majnoni and Powell employ a bootstrapping technique to calculate loss

distribution functions for Argentina, Brazil, and Mexico, using data on loan

performance from public credit registries. They then use these functions to

estimate the size of expected and unexpected losses of an average-sized

bank with a loan portfolio randomly drawn from the universe of loans

within the financial system. Their results show that these three countrieshave significantly higher default probabilities than Group of Ten (G10)

countries. As a result, both current practice under Basel I and the sug-

gested standards under Basel II may be inadequate. To achieve a 99 percent

level of protection, capital requirements would need to be close to 15 per-

cent, which is significantly higher than the 8 percent level recommended

in Basel I. Even higher levels would be required to achieve 99.9 percent

protection, as intended in Basel II. They also find that Basel II’s IRB

approach would result in levels of 90–95 percent protection rather than the99.9 percent goal. This is not surprising, since the IRB was calibrated for

the safer economies of G10 countries.

If bank regulation needs modernizing in Latin America, public trans-

port does too. The spectacle of streets packed with old buses spewing black

smoke is all too common in many cities of the region, from Mexico City

to Quito and from São Paulo to Santiago. Poor public transport induces

more private cars to enter the streets, worsening congestion and pollution.

If you think that this is a textbook case of the state not doing the job of

providing public services, think again. Bus systems are private in many

cities in Latin America, and that does not seem to solve the problem. As

Juan Carlos Echeverry, Ana María Ibáñez, Andrés Moya, and Luis Carlos

Andrés Velasco xi


13/257

Hillón document in their paper in this volume, the market for urban buses

is ripe with market failures: unclear definition of property rights on the

curbside and on the road; cartelization that results in fares set above the

competitive equilibrium levels; misalignment between the incentives of bus

drivers and owners, in a typical principal-agent conflict; and congestion and

pollution externalities. In many developing countries, these market fail-

ures are exacerbated by weak regulation and enforcement. The result often

is too many buses each carrying too few passengers in unsafe conditions,

clogging the streets and soiling the air as they move (or fail to move) along.

One city in Latin America to have tackled the problem head-on is

Bogotá, Colombia. Its so-called TransMilenio system is now being imitatedin Quito and Santiago, among others, as well as several cities in Colom-

bia. Echeverry, Ibáñez, Moya, and Hillón explain the logic behind the new

system and analyze is effects. The key elements of the new system are as

follows: (i) a hybrid public-private system, with concession contracts for

private service providers; (ii) competition “for the road” (rather than “on

the road”) in a tendering process, with fare-setting based on long-term

investment recovery; (iii) remuneration based on kilometers traveled rather

than passengers transported, so as to prevent drivers from fighting over pas-sengers on the street; (iv) separation between the transportation service and

the fare collection process; and (v) exclusive road and curb-side service in

metro-like stations.

Congestion, pollution, traffic accidents, travel times, and waiting times

all fell dramatically along the corridors where TransMilenio was first put to

work. The system was initially hailed as the solution of Bogotá’s serious

transport problems. Not all results were unambiguously positive, however,

as the paper makes clear. Increased ridership resulted in jammed busesand rising waiting times. Moreover, the full system covering the entire

city is not expected to be operating until 2015. This gradual transition did

not help: older buses were displaced to secondary streets, where traffic

and pollution increased.

A cost-benefit analysis of the system as is, with approximately 25 per-

cent of the routes in operation, reveals welfare gains for users of the new

routes, but an overall negative effect stemming primarily from increases

in travel time for passengers using the traditional transport system. Since

congestion costs are highly nonlinear, the welfare losses from heightened

congestion in unserved corridors more than offset the benefits from Trans-

Milenio, even though those benefits are sizeable. The authors conclude by

xii E C O N O M I A , Spring 2005


14/257

arguing that the adoption of a new public transport system must be cou-

pled with improved regulation of all other public transport providers, so as

to avoid the problem that arose in Bogotá.

What happens to workers’ wages and employment prospects once they

are displaced from their current jobs—for instance, by trade reform? If they

are likely to be re-hired quickly at comparable wages, then no policy

response is called for; but if some wage losses are large and lasting, then

targeted help for displaced workers may be called for. David Kaplan,

Gabriel Martínez, and Raymond Robertson study the issue for the case of

Mexico, using an administrative data set that allows them to follow indi-

vidual workers over a period of thirty-two quarters in four regions that varysignificantly in labor market conditions. They focus on the differences in

institutions, inequality, and labor market conditions that may explain dif-

ferences in wage behavior after displacement.

One striking result is the heterogeneity of worker experiences, which

range from large wage losses to many instances of gains after displacement.

This is consistent with earlier results for other countries, but it cannot be

attributed to differences in institutions (rates of unionization) or inequal-

ity, which are quite similar across Mexico. Rather, Kaplan, Martínez, andRobertson argue that labor market conditions, which vary quite a bit across

time and regions within Mexico, explain the heterogeneity of experiences.

In good times and in the most economically active regions, postdisplace-

ment wages are generally higher than they were in the previous jobs. How-

ever, workers who are fired during times of high unemployment and in

less economically active regions experience lasting effects on wages. If any

public assistance is to be disbursed, Kaplan, Martínez, and Robertson

argue, it should go to these workers.All papers but one included in this issue were presented at the panel

meeting held in San José, Costa Rica, in October 2004. The local hosts, and

particularly Juan Rafael Vargas, provided much help. As usual, associate

editors of Economía, members of the 2004 panel, and outside discussants

and referees have done an outstanding job. Thanks is due to them all.

Andrés Velasco xiii


15/257


16/257

The Electoral Consequences of the

Washington Consensus

No country in Latin America escaped the dictums of the Washington

Consensus. From Brazil under left-leaning Fernando Henrique Car-

doso to Mexico under ultra-orthodox economist Ernesto Zedillo

and Peru under Alberto Fujimori’s yoke, macroeconomic imbalances were

brought under control, barriers to international trade were lifted, and state-

owned enterprises were privatized. Whether this one-size-fits-all prescrip-

tion was imposed from outside or adopted at will by the governments elected

on the promise of improving the lot of their peoples may be a matter of debate. But all sides seem to agree on one point: the results did not meet

the expectations created both by outsiders and by those in power.

Up to the mid-1980s only two countries in Latin America had adopted

a package of policies similar to what became to be known as the Wash-

ington Consensus at the turn of the decade. Those two were undemocratic

Chile and impoverished Bolivia, by then among the most politically and

economically unstable countries, if not in the world, then certainly in Latin

America. Extreme cases, extreme policies: that was a common interpreta-tion of the two experiences. Less common was the expectation that those

policies were about to be adopted by virtually every Latin American coun-

try in the next few years, both those in which democracy had been the rule

1

E D U A R D O L O R A

M A U R I C I O O L I V E R A

Lora and Olivera are with the Inter-American Development Bank.

We are grateful for the valuable research assistance of Carlos Andrés Gómez. We also

thank Benito Arrunada, Mauricio Cárdenas, Stephen Kay, Ugo Panizza, Andres Rodríguez-

Clare, Mariano Tommasi, Jessica Wallack, and seminar participants at ISNIE-University

Pompeu Fabra, LACEA-PEG, Econnet-IADB, and the Economia panel meeting for com-ments and suggestions. We are especially grateful to Rafael Di Tella, Sebastian Galiani, and

Ernesto Dal Bó for their detailed and very useful comments and suggestions. We would like

also to thank Sebastian Saiegh for allowing us to use his data on political coalitions.


17/257

for decades, like Colombia, Costa Rica, and Venezuela, and those where

the third wave of democratization was just arriving, such as Argentina,

Brazil, and Uruguay.

The years of high expectations, both about democratization and about

Washington Consensus–type policies, are over. Latin Americans are still

convinced democrats, but enthusiasm has waned. Three out of every four

Latin Americans see democracy as the best form of government—or rather,

as the least bad, since 68 percent think that democracy is not function-

ing well in their countries. Latin Americans are even more sceptical about

the benefits of promarket economic policies. Only one out of four Latin

Americans considers privatization to have been beneficial for his or her coun-

try and barely 16 percent think that the market economy is doing a good job.1

Malaise is getting the upper hand in a number of places. Electricity and

water privatizations were blocked in Arequipa (Peru) and Cochabamba

(Bolivia), following violent clashes between vociferous opponents and the

police. An ambitious project to attract foreign direct investment to

Bolivia’s gas sector was derailed by the Indian communities. While these

events may be dismissed as isolated expressions of popular feeling, a new

crop of presidents from Néstor Kirchner in Argentina to Lucio Gutiérrez

in Ecuador and Tabaré Vásquez in Uruguay has won clear majorities inpopular elections after campaigning against the excesses of market-

oriented policies.

In an attempt to establish whether this malaise is justified or not, econ-

omists have devoted substantial effort to assessing the economic and social

consequences of the Washington Consensus policies. The dominant view

seems to be that they have had positive effects on economic growth and

income levels, though there is intense debate over the size of those effects,

over whether they are transient or permanent, and over the importance of

each of the components of the Washington Consensus. The dominant viewalso holds that the effects have been muted by lack of regulatory and insti-

tutional support for the liberalization efforts, though the specific forms of

regulation and institutions necessary for that purpose are far from clear.

Even more intense is the debate over the social and distributional effects of

fiscal stabilization and promarket reforms, which are the two main pillars of

the Washington Consensus.2

2 E C O N O M I A , Spring 2005

1. Opinion data come from the 2003 issue of Latinobarómetro, a public opinion surveyconducted by the Corporación Latinobarómetro, Santiago, Chile.

2. These debates are surveyed in Lora and Panizza (2002); Kuczynski and Williamson

(2003); and Lora, Panizza and Quispe-Agnoli (2004).


18/257

However, the future of these policies will depend not so much on their

efficacy but on whether they receive the support of the electorate. On this,

the state of knowledge is much more scant and fragmentary, as will be seen

below. This paper attempts to help fill that vacuum by evaluating through

econometric methods the electoral consequences of the Washington Con-

sensus. Although our approach is backward looking, it sheds considerable

light on the future. Our study shows that the electorate cares not only about

the outcomes of the policies (maybe about only some outcomes and not

others), but also about the policies themselves, irrespective of whether they

produce good or bad (observable) outcomes. In addition, the electorate

seems to care about whether the policies adopted by a government are

in line with the ideology of the incumbent’s party and with preelectoral

promises. Furthermore, in presidential regimes voters cast separate votes

for the executive and the legislature, and outcomes and policies affect each

vote differently. The presidential vote is more volatile and more suscepti-

ble to economic outcomes and policies, but votes for the legislature are not

completely immune: policies in which the legislature clearly plays a role,

such as privatizations, tend to have electoral consequences. These results

provide a nuanced landscape for the future of Washington Consensus poli-

cies, where neither bold backslashes nor aggressive promarket reformsshould be expected in the future. Not only is the time of high expectations

over; perhaps the time for deep reforms is also past.

In the next section of the paper we present a short survey of the literature

assessing the electoral consequences of the Washington Consensus policies

and derive our empirical hypotheses. On that basis, we then discuss the the-

oretical and econometric approaches that support the empirical analysis.

In subsequent sections we describe the data, present the econometric find-

ings, and discuss our conclusions.A note on terminology is in order before proceeding. “Neo-liberal,”

“market-oriented,” “orthodox,” and a variety of other labels have been

attached to the set of economic policies in vogue since the early nineties in

Latin America and elsewhere. We use these terms interchangeably, but not

loosely: for the sake of clarity and brevity, this paper deals with the ten poli-

cies summarized in the classic article by Williamson that made the term

“Washington Consensus” famous.3 We assume that all those labels refer to

that same set of policies (as detailed below in the section titled “Data”).

Eduardo Lora and Mauricio Olivera 3

3. Williamson (1990a).


19/257

Review of the Empirical Literature and Some Testable Hypotheses

The most straightforward view of the response of the electorate to eco-nomic policies is based on the “economic voting” argument: people base

their electoral decisions on cost-benefit calculations. If the policies bring

net benefits to them, they cast their votes to support the government, or the

party, administering those policies; if the policies bring losses to them,

they lend their support to the candidate, or the party, opposing them. Eco-

nomic voting is usually assumed to be retrospective: voters observe past

performance and assume that past trends will persist into the future if the

government or the party remains in power. If those trends are deemed

acceptable, given a set of standards or expectations voters decide to reelect

the incumbent, or his party if the option of reelection does not exist. There-

fore, in retrospective economic voting policies play no direct role, since

voters decide entirely on the base of past outcomes.4

Considerable evidence from advanced industrial democracies supports the

view that past economic performance influences people’s voting decisions

and their support for governments.5 An important empirical finding from this

literature is that voters base their decisions on aggregate (or “sociotropic”)

economic outcomes such as growth, inflation, and unemployment, ratherthan on individual (or “pocketbook”) outcomes. Most of the empirical liter-

ature on developed countries comes from single-country analyses, based

either on time-series electoral outcomes or public opinion polls. The eco-

nomic voting hypothesis is more robust for public opinion polls than for

actual electoral outcomes.6 Empirical studies of electoral behavior in the

United States using state-level data lend support to the simple economic

voting hypothesis, in the sense that voters are able to evaluate their state’s

economic performance relative to that of the national economy. Further-more, they (irrationally) reward state governors for economic fluctuations

that are unrelated to gubernatorial actions, which implies that they have lim-

ited ability to filter aggregate economic information.7 The ability of voters


4. Stokes (2001b, pp. 1–18) provides a concise review and discussion of the theoretical

underpinnings of retrospective economic voting.

5. Based on the seminal work by Downs (1957); among the initial papers on economic

voting in the United States are Kramer (1971); Meltzer and Vellrath (1975); and Arcelus and

Meltzer (1975).6. Lewis-Beck (1988) is a salient example of the early empirical literature based on opin-

ion polls in European countries. For a review of this literature, see Stokes (2001b, pp. 2–8).

7. Wolfers (2002).


20/257

to gather and update information is a central issue in the theoretical and

empirical literature on economic voting.8 Although some evidence points

out to the presence of prospective as well as retrospective behavior, uncer-

tainty about the workings of the economy and the relatively high cost of

gathering and processing the information necessary to forecast outcomes

is consistent with the importance of retrospective voting in the empirical

findings.9

Empirical support for the economic voting hypothesis in Latin America

has been uncovered by Karen Remmer, Michael Coppedge, Kenneth Roberts

and Erik Wibbels, and Susan Stokes.10 A concise summary of these findings

is presented in table 1. Based on data for twenty-one competitive elections

between 1982 and 1990, Remmer has found that conditions of economic

crisis undermine support for incumbents and provoke high levels of elec-

toral volatility.11 The magnitude of the electoral change is found to be asso-

ciated with the depth of the crisis during the campaign period, with variations

in exchange rates, GDP, and inflation highly correlated with various

indicators of electoral outcomes. Her results also suggest that the effect

of economic conditions on electoral instability are mediated by the struc-

ture of the party system (insulating two-party systems from the volatility

experienced by more fragmented systems). However, as Stokes pointsout, these results are anomalous given the predictions of normal economic

voting, as she “finds that incumbent parties suffered larger losses at the

polls when inflation went down (significant) and when GDP rose (not

significant).”12

In a subsequent paper, Remmer presents new estimates on the influence

of inflation and growth on the incumbent vote in presidential elections.13

Her new database covers forty-nine elections for seven countries between

1983 and 1999. Her results indicate that after controlling for the advantageof incumbency as well as major differences in the structure of party sys-

tems, electoral outcomes are strongly influenced, in the direction expected,

by macroeconomic performance in the year before the election. That is, infla-

tion is found to be negatively correlated with electoral support, whereas


8. For a review of this debate, see Duch and Stevenson (2004); and Keech (1995).

9. On prospective behavior, see, for instance, Lewis-Beck (1988).

10. Remmer (1991, 2003); Coppedge (2001); Roberts and Wibbels (1999); Stokes (2001b).11. Remmer (1991).

12. Stokes (2001b, p. 27).

13. Remmer (2003).


21/257

growth is positively correlated with it. Furthermore, inflation is significant

in all the regressions presented, while growth is more significant for the

elections held in the 1990s than for those in the 1980s, indicating that the

sensitivity of the electorate to economic performance has increased rather

than waned over time.Coppedge’s empirical work focuses on the impact of changes in inflation

on legislative vote shares. His dependent variable consists of 132 changes in

legislative vote shares for major parties in eleven countries from 1978 to

1995. His only indicator of economic performance is the change in (the

log of) inflation from the last year of the previous government to the last

year of the current government. By interacting this variable with appropri-

ate dummies, Coppedge finds that changes (whether increases or decreases)

in inflation affect electoral support for the incumbents’ parties in theexpected way, while only increases in inflation improve the vote share of

the opposition parties. However, these results apply only to parties “with a

fluid base,” that is, parties that do not count on a strong party identification.

When there is such identification, voters are reluctant to question their

party identification on the basis of macroeconomic outcomes.

Roberts and Wibbels consider economic voting as a possible explana-

tion of electoral volatility in Latin America. Their database includes fifty-

eight congressional elections and forty-three presidential elections in sixteen

Latin American countries during the 1980s and 1990s. Their results show

that economic performance has an effect on electoral stability. Economic

growth stabilizes partisan support in legislative elections, whereas sharp


T A B L E 1 . Summary of Empirical Findings on Economic Voting in Latin America

Election type

(number of Estimation

Study Dependent variable countries) Period method Main results

Remmer (1991)

Remmer (2003)

Roberts and

Wibbels (1999)

Coppedge (2001)

Stokes (2001b)

Source: Authors’ calculations.a. Not significant.

Electoral volatility

Vote shares

Electoral volatility

Vote shares

Probability of a

security-oriented

candidate beingelected

Inflation −; GDPgrowth +a

Inflation −; GDPgrowth +

Inflation −a; GDPgrowth +

Inflation

Inflation −; GDPgrowth +

Presidential (12)

Presidential (8)

Legislative and

presidential (16)

Legislative (11)

Presidential (15)

1982–90

1983–99

1980–97

1978–95

1982–95

Pooled OLS

Pooled OLS

Pooled OLS

Pooled OLS

Probit


22/257

changes in the rate of inflation from one administration to the next, whether

positive or negative, produce the opposite effect. Short-term inflation influ-

ences support for incumbent presidents, but growth changes have only a

weak effect on the vote for incumbents, “which suggests that voters are

more inclined to hold them directly accountable for monetary stability

than economic growth.” Although electoral volatility is influenced by eco-

nomic performance, it is also related to the institutional characteristics of

political regimes and party systems, and to the structure and organization

of class cleavages.14

In her study of “neoliberalism by surprise,” Stokes uses data from

twenty-three elections in the 1980s and 1990s in order to assess how the

electorate judges incumbents who, having campaigned for stability-oriented

or protectionist policies, once in office switch to market-oriented ones. She

finds that for both, “switchers” and “non-switchers,” economic growth

and inflation affect their vote share in the expected ways. Furthermore,

voters are more sensitive to economic outcomes in the case of “switchers,”

although this result is not statistically significant (more on these results

below).15

These empirical studies taken together lend support to the retrospective

economic voting argument in both presidential and legislative elections.They make clear that voting decisions are also influenced by political, insti-

tutional, and structural factors and that some of these factors may influence

the severity with which voters judge economic outcomes. Therefore, based

on these studies, two testable propositions are derived:

1. Electoral support for the incumbent’s party is higher, the better the

aggregate economic outcomes during his or her administration.

2. The sensitivity of electoral support to economic outcomes depends on

the institutional characteristics of the political regime and the party system.As mentioned, in normal economic voting only past outcomes influence

people’s views. However, as in all six of the Stokes case studies on mar-

ket reforms in new democracies, people sometimes react to economic

deterioration by supporting the government more strongly; and conversely,

they sometimes respond to economic improvements with pessimism and

opposition.16 Normal economic voting is not the only pattern, especially in

the process of deep economic reform. If there are good reasons to believe


14. Roberts and Wibbels (1999), quote from p. 584.

15. Stokes (2001a).

16. Stokes (2001a).


23/257

that past circumstances are not good indicators of the future, information

other than past economic outcomes may influence people’s electoral deci-

sions. For instance, voters may recognize that past circumstances were

affected by factors beyond the government’s control and exonerate the

incumbent from the responsibility for past declines in their welfare. Vot-

ers may then forecast their future welfare as a function of government pol-

icy, rather than as an extrapolation of the past. This sounds simpler than

it is, of course, because future government policies are unknown and

because the relationship between policies and outcomes is diffuse. Peo-

ple’s expectations of future policies may be formed on the basis of the

policies adopted or announced by the incumbent or on the basis of his

party’s ideology. These policy expectations may then be translated into

expected outcomes through a set of beliefs and hypotheses about their pos-

sible consequences.

It is often implicitly assumed that people’s (average) beliefs conform

to the actual functioning of the real world. If that is so, assessing the

effects of economic policies would help explain voters’ electoral decisions.

Economists have devoted considerable effort to evaluating the impact of

Washington Consensus policies on economic growth, income distribu-

tion, employment levels, and a host of other variables.17

However, therehas been no comparable effort to examine whether these results are con-

sistent with how the electorate responds to those policies. The only study

on the subject, by Carlos Gervasoni, has found positive correlations

between several indicators of heterodox (that is, anti-neoliberal) policies

and losses in the vote shares of the parties of the incumbents who adopted

those policies.18 The variable with the largest and most significant effect is

money supply growth. Import protection indicators are also significant,

whereas fiscal deficit and the share of the state in GDP are not significant.These results suggest that Washington Consensus policies do not entail

electoral costs and may even produce electoral benefits, probably because

they bring positive economic effects. It is suggestive that the most signif-

icant policy variable is the money supply, because it is well known that

inflation is, ultimately, a monetary phenomenon, and as mentioned, empiri-

cal evidence suggests that inflation is a key economic outcome influencing

electoral decisions.


17. For surveys of the literature, see Inter-American Development Bank (2003, chap. 5);

Kuczynski and Williamson (2003); and Lora and Panizza (2002).

18. See Gervasoni (1997), citing a 1995 study.


24/257

However, it is a great leap of faith to assume that people’s beliefs con-

form to the actual consequences of policies. In mapping policies on out-

comes, ideology and leaders’ opinions may be more important for most

people than their limited understanding of how policies work their influence

through the social and economic structures to affect production, employ-

ment, or income distribution. Evidence on how those factors influence

electoral responses to economic policies is very scant. However, in-depth

case studies on Argentina and Venezuela by Javier Corrales clearly show

that the reaction of the electorate to the adoption of neoliberal economic

policies in the 1990s was mediated by the party structure and other insti-

tutional factors.19 The cohesion and tactics of the Partido Peronista help

explain the electorate’s support of the neoliberal reforms in Argentina in

the early 1990s, as well as their demise a decade later. Venezuela’s Acción

Democrática lacked that cohesion, and its reforms were soon rejected by

the electorate.

If voters care about policies and not only about past outcomes, the pol-

icy announcements of presidential candidates will be a key source of infor-

mation. However, campaign promises are often poor predictors of actual

policy: according to Stokes, of the thirty-three Latin American govern-

ments that adopted promarket reforms between 1982 and 1995, only abouthalf (seventeen) hinted during their campaigns that such reforms were

going to be implemented.20 This raises several empirical issues. First, do

policy announcements in fact influence electoral decisions? Empirical evi-

dence from the United States and other advanced industrialized economies

shows that they do: people seem to base their opinions in part on campaign

announcements, and voters punish ambiguous campaigns.21 Of course,

some promises may resonate more than others, depending on, among other

things, economic circumstances. For thirty-eight Latin American electionsin the 1980s and 1990s, Stokes finds that stability-oriented candidates (as

opposed to market-oriented ones) stand a better chance of being elected,

the lower the rates of GDP growth and inflation.22

A second empirical issue is whether deviating from campaign promises

carries electoral costs for the incumbent. Although deviations may in

principle be costly, they may produce a positive payoff if they signal the


19. Corrales (2002).20. Stokes (2001a).

21. For a brief review of this topic, see Stokes (2001a, pp. 4–5).

22. Stokes (2001a, pp. 93–97).


25/257

incumbent’s commitment to achieving highly desirable economic out-

comes at the expense of more immediate partisan support.23 According to

Stokes, deviating from campaign promises does carry electoral costs,

although only weakly.24 However, since her estimates control for eco-

nomic outcomes, this result implies that policy switches may still have a

positive electoral payoff if the new policies bring substantial economic

improvement. Neoliberalism by surprise may still be a good political

strategy.25

A common theme in the literature on economic voting is the conditional

nature of voters’ responses to economic outcomes and policies. As men-

tioned, the severity of their judgment depends on their attachment to the

party in power, the structure of the party system, and other institutional

considerations. It also depends, although weakly, on whether the policies

adopted by the incumbent are in line with his campaign pronouncements.

An additional variation on this theme holds that the electorate is better pre-

pared to support untested policies, even if they may cause short-term duress

or if they run counter to established beliefs, when economic conditions have

deteriorated.26 However, once conditions improve or simply stabilize, tol-

erance subsides and support for further reforms wanes. Therefore, while

uncertainty is welcome at the outset of the reform process, certainty is thekey factor for its consolidation. Based on case studies of Peru and Argentina,

Kurt Weyland offers persuasive evidence that the public was supportive to

the reform process while there was a perception of acute economic crisis.27

Even though the reformers were reelected, support for their economic pro-

grams was already diminishing. Corrales endorses this view in his analysis

of the reform process in Argentina and Venezuela, although he acknowl-

edges that in the latter case support for reform was never very strong.28

Therefore, the literature on economic voting suggests that policies, notonly outcomes, may influence electoral decisions. As with outcomes, voters’

position with respect to policies may be mediated by a host of factors,

including ideological considerations, policy pronouncements during the


23. For a theoretical approach, see Cukierman and Tommasi (1998).

24. Stokes (2001a, p. 95).

25. Cukierman and Tommasi (1998); Navia and Velasco (2003).

26. This behavioral hypothesis is based on seminal work by Thaler and others (1997),

Kahneman and Tversky (1979), and Tversky and Kahneman (1991), who find that people aremore prone, even eager, to assume risks after experiencing losses.

27. Weyland (2002).

28. Corrales (2002).


26/257

electoral campaign, and the state of the economy at the time of elections.

This gives rise to the following additional testable propositions:

3. Electoral support for the incumbent’s party depends on the economic

policies adopted. Policies may carry electoral costs even when they deliver

good economic outcomes.

4. The electorate’s tolerance of unpopular policies depends on the ide-

ology of the incumbent’s party, his or her campaign statements, and the

initial state of the economy.

Empirical Approach

None of the empirical literature just reviewed offers a full-fledged theo-

retical model of electoral behavior, and we have no intention of providing

one. However, the series of hypotheses arising from that literature can be

organized in a simple framework such that the persistence of the vote for

the incumbent’s party is a function of a vector of economic outcomes and

a vector of policies (both relative to their past values):

where V t and V

t −1are the share of the vote for the incumbent’s party at the

end and the beginning, respectively, of its term in office; X t

and X t −1

are the economic outcomes at the time of each election; and Pt and P

t −1

are the policies at those two moments. A is the set of other parameters that

may influence the stability of the vote for the party in office, and ut is an

error term. β and γ are our parameters of interest. In this simple frame-

work, hypothesis 1 states that β is positive for economic outcomes that aredesirable, such as growth, or negative for undesirable ones, such as infla-tion or unemployment (and assumes that γ is zero, since it ignores theinfluence of policies). Hypothesis 2 postulates that β is a function of somefeatures of the political system, such as party fragmentation or the ideo-

logical polarization of the party system. The stronger these features, the

higher the electorate’s response to the economic outcomes. Hypothesis 3,

which postulates that the electorate cares about the choice of policies,

implies that γ is not zero but probably negative if the policies are marketoriented. Finally, hypothesis 4 states that some aspects of the political

and economic context when the incumbent’s party was initially elected

may affect the way the electorate judges the adoption of policies. This

V

V A

X

X

P

P u

t

t

t

t

t

t

t

− − −= ∗

∗

∗1 1 1

β γ

,



27/257

hypothesis can be incorporated in our framework by assuming that γ is afunction of those factors. More specifically, γ will be smaller (in absolutevalue) when the policies adopted were those announced by the incumbent

during his election campaign, when they are in line with his party’s ideol-

ogy, or when the economy started from a situation of crisis. Although our

framework is general enough to test further hypotheses, due to sample size

limitations and for the sake of parsimony and tractability, we restrict its

application to the hypotheses identified in the literature review.

Our economic voting framework is relevant both for presidential and

for legislative elections. An important feature of presidential systems is the

separation of powers between the legislative and the executive, aimed at

imposing checks and balances in order to discipline parties and make themaccountable.29 Since checks and balances force the two powers to agree on

policies, voters should be expected to pass judgment on the performance

of the incumbent’s party in both branches on the basis of economic out-

comes and policy decisions. Of course, we should expect that the influence

of each policy on presidential vis-à-vis legislative elections will depend on

whether such policy is controlled exclusively by the executive or not. While

legislatures have very little influence on monetary, exchange rate, and tariff

policies in most Latin American countries, they do have a strong (even over-riding) influence on tax policies, privatization decisions, and the regulation

of financial, capital, and labor markets. As Brian Crisp and Gregg Johnson

show, contrary to widespread belief, Latin American legislatures make use

of their powers to influence the timing and depth of promarket reforms.30

And according to Roberts and Wibbels, the electorate holds each branch of

power more accountable for some outcomes than for others.31 When assess-

ing the role of the legislature in policy decisions in Latin America, it is impor-

tant to keep in mind that the incumbent’s party (or the coalition of parties

backing the incumbent) usually holds the majority in that body (see below).

To estimate the relevant parameters, the previous expression can be

written in logs as

where d log(V t ) corresponds to the change in (the log of) the share of votes

for the incumbent party between t, the time when its performance is eval-

d V F d X d pt t t t log log log log ,( ) = + ( ) + ∗ ( ) + ∗ ( ) +α ψ β γ ε


29. Persson, Roland, and Tabellini (1997).

30. Crisp and Johnson (2003).

31. Roberts and Wibbels (1999).


28/257

uated, and t − 1, when it was elected for office; d log( X t ) and d log(P

t ) are

the changes in (log measures of the) outcomes and policies, respectively;

εt is equivalent to log(u

t ); and

α + ψ log(F ) is equal to log( A), with

αas a

constant parameter and F as a set of political control variables.

We estimate separate models for presidential and legislative elections

with panel data for seventeen countries starting from the mid-1980s

described below. Potential problems of heteroscedasticity and endogene-

ity need to be addressed in this type of specification. The former may arise

from country or party heterogeneity and is dealt with by the use of White

robust standard errors. The endogeneity problem stems from potential omitted

variables, since differentiating countries solely by the economic and policy-

related variables included in sets X and P may not capture all the sources of heterogeneity.32 This is partly dealt with by the inclusion as controls of a set

of political variables (represented by F ). However, other country-related

factors might bias the estimations if they are correlated with the explanatory

variables. To take care of this problem, we run all the regressions with coun-

try fixed effects (although, admittedly, our sample size is too small to get

precise estimation of these effects).33 The fixed effects estimator is

where C is the set of country dummies.

Data and Sources

Table 2 presents the structure of our database, and table 3 shows correlations

between the more relevant variables. The database includes a total of sixty-

six presidential elections and eighty-one legislative elections in seventeen

d V F d X d p C t t t t t log log log log ,( ) = + ( ) + ∗ ( ) + ∗ ( ) + +α ψ β γ λ ε


32. We assume that the two other sources of endogeneity—reverse causality and measure-

ment error—are not latent in our model. Reverse causality is not a concern, since voters

evaluate the incumbent’s behavior after policies and outcomes are known. Measurement error

problems may be present, depending on the actual process of expectations formation. However,

ample empirical evidence provides support for the hypotheses of retrospective voting, which

for our framework implies that expectations are formed on the basis of past outcomes only.

33. All the regressions were also run without fixed effects: while virtually all the conclu-

sions are the same, in these regressions, some of the explanatory variables (especially those

measuring promarket policies) show higher levels of significance. We have also run the re-gressions including a common time trend, or including five-year period fixed effects, with-

out any important divergence from the results presented below. Results are available upon

request from the authors.


29/257

T A B L E 2

.

S t r u c t u r e o f t h e D a t a S e t

N u m b e r o f e l e c t i o n s ( p e

r i o d )

N u m b e r o f p a r t i e s t h a t h e l d . . .

F r a g m e n t a t i o n a

C o u n t r y

P r e s i d e n t i a l

L e g

i s l a t i v e

P r e s i d e n c y

L a r g e s

t s h a r e i n t h e l e g i s l a t u r e

M e a n

M i n i m u m

M a x i m u m

P o l a r i z a t i o n i n d e x

A r g e n t i n a

3 ( 1 9 8 9 –

9 9 )

8 ( 1 9

8 5 –

9 9 )

2

3

2 . 7

7

2 . 3

0

3 . 0

6

0 . 2

3

B o l i v i a

4 ( 1 9 8 5 –

9 7 )

4 ( 1 9

8 5 –

9 7 )

3

3

4 . 0

6

3 . 4

2

5 . 0

8

0 . 5

2

B r a z i l

3 ( 1 9 8 9 –

9 8 )

4 ( 1 9

8 6 –

9 8 )

2

2

6 . 6

0

2 . 7

6

8 . 2

7

0 . 2

5

C h i l e b

3 ( 1 9 8 9 –

9 9 )

4 ( 1 9

8 9 –

2 0 0 1 )

2

2

4 . 9

0

4 . 8

4

4 . 9

9

0 . 1

6

C o l o m b i a

5 ( 1 9 8 6 –

0 2 )

5 ( 1 9

8 6 –

9 8 )

2

1

2 . 6

6

2 . 2

1

3 . 0

9

0 . 1

6

C o s t a R i c a

5 ( 1 9 8 6 –

0 2 )

5 ( 1 9

8 6 –

2 0 0 2 )

2

2

2 . 3

1

2 . 2

1

2 . 5

6

0 . 4

2

D o m i n i c a n R e

p u b l i c

5 ( 1 9 8 6 –

2 0 0 0 )

4 ( 1 9

8 6 –

2 0 0 0 )

3

3

2 . 4

8

2 . 1

8

2 . 8

8

0 . 5

5

E c u a d o r

4 ( 1 9 8 8 –

9 8 )

7 ( 1 9

8 6 –

9 8 )

4

2

6 . 0

5

4 . 2

9

7 . 5

6

0 . 3

6

E l S a l v a d o r

4 ( 1 9 8 4 –

9 9 )

6 ( 1 9

8 5 –

2 0 0 0 )

2

2

2 . 6

8

2 . 4

1

3 . 0

6

0 . 3

9

G u a t e m a l a

4 ( 1 9 8 5 –

9 9 )

5 ( 1 9

8 5 –

9 7 )

4

4

3 . 3

1

2 . 3

5

4 . 4

4

0 . 2

4

H o n d u r a s

5 ( 1 9 8 5 –

2 0 0 1 )

5 ( 1 9

8 5 –

2 0 0 1 )

2

2

2 . 1

8

2 . 0

0

2 . 5

8

0 . 4

2

M e x i c o

3 ( 1 9 8 8 –

2 0 0 0 )

6 ( 1 9

8 5 –

2 0 0 0 )

2

1

2 . 3

8

1 . 8

5

2 . 8

2

0 . 3

2

N i c a r a g u a

3 ( 1 9 9 0 –

2 0 0 1 )

3 ( 1 9

9 0 –

2 0 0 1 )

1

2

2 . 0

5

2 . 0

5

2 . 0

5

0 . 5

8

P e r u

4 ( 1 9 8 5 –

2 0 0 0 )

4 ( 1 9

8 5 –

2 0 0 0 )

3

4

3 . 8

0

2 . 5

0

5 . 8

3

0 . 5

1

P a r a g u a y

4 ( 1 9 8 9 –

2 0 0 3 )

4 ( 1 9

8 9 –

2 0 0 3 )

1

1

2 . 2

1

1 . 8

8

2 . 5

4

0 . 4

0

U r u g u a y

3 ( 1 9 8 4 –

9 9 )

3 ( 1 9

8 4 –

9 9 )

2

2

3 . 1

9

2 . 9

2

3 . 3

2

0 . 4

2

V e n e z u e l a

4 ( 1 9 8 8 –

2 0 0 0 )

4 ( 1 9

8 8 –

2 0 0 0 )

4

1

3 . 9

2

2 . 3

4

5 . 7

9

0 . 3

0

T o t a l o r a v e r a

g e

6 6

8 1

2 . 4

2 . 2

3 . 3

9

2 . 6

2

4 . 1

1

0 . 3

7

S o u r c e : P a y n e a n d o t h e r s ( 2 0 0 2 ) , c o m p l e m e n t e d w i t h t h e

P o l i t i c a l D a t a b a s e o f t h e A m e r i c a s ( O r g a n i z a t i o

n o f A m e r i c a n S t a t e s a n d G e o r g e t o w n U n i v e r s

i t y ) .

a .

E f f e c t i v e n u m b e r o f p a r t i e s i n t h e l e g i s l a t u r e .

b .

I n C h i l e , t h e e f f e c t i v e n u m b e r o f p a r t i e s d i f f e r s f r o m t

h e

n u m b e r o f c o a l i t i o n s ( C o n c e r t a c i ó n a n d A l i a n z a p o r C h i l e ) , w h i c h a r e c l o s e t o 2 i n e f f e c t i v e t e r m s a n d o f w h i c h o n l y C o n c e r t a c i ó n h a s h e l d t h

e p r e s i d e n c y .


30/257

T A B L E 3

.

C o r r e l a t i o n s

I n fl a t i o n ( l o s s

I n s t i t u t i o n a l

V o t e s

G r o w t h

o f p u r c h a s i n g

U n e m p l o y m e n t

G i n i i n d e x

M a c r o i n d e x

S t r u c t u r a

l i n d e x

i n d e x

( s h a r e )

F r a g m e n t a t i o n

P o l a r i z a t i o n

P r o m i s e s

I d e o l o g y

( l o g , c h a n g e )

p o w e r , c h a n g e )

( c h a n g e )

( c h a n g e )

( l o g , c h a n g e )

( l o g , c h a n g e )

( l o g , c h a n g e )

P r e s i d e n t i a l e l e c t i o n s

V o t e s ( s h a r e )

1 . 0

0

F r a g m e n t a t i o

n

− 0 . 3

2

1 . 0

0

P o l a r i z a t i o n

− 0 . 2

0

− 0 . 0

9

1 . 0

0

P r o m i s e s

− 0 . 1

6

− 0 . 0

5

− 0 . 3

2

1 . 0

0

I d e o l o g y

− 0 . 1

6

0 . 0

3

− 0 . 4

5

0 . 3

7

1 . 0

0

G r o w t h ( l o g , c h a n g e )

0 . 2

1

− 0 . 0

2

− 0 . 0

6

0 . 2

5

0 . 1

4

1 . 0

0

I n fl a t i o n ( l o s s

o f

p u r c h a s i n g

p o w e r , c h a n g e )

− 0 . 2

4

− 0 . 5

5

0 . 3

3

0 . 0

2

− 0 . 1

6

−

0 . 4

1

1 . 0

0

U n e m p l o y m e

n t ( c h a n g e )

− 0 . 0

2

0 . 1

3

− 0 . 4

6

− 0 . 1

6

0 . 2

8

−

0 . 2

7

− 0 . 4

8

1 . 0

0

G i n i i n d e x ( c h

a n g e )

− 0 . 2

2

− 0 . 1

6

− 0 . 1

9

0 . 2

1

0 . 0

9

0 . 2

3

0 . 1

Documents

Andres Velasco Economia