Trade Liberalization and Growth: New Evidence
Romain Wacziarg and Karen Horn Welch
A new data set of on openness indicators and trade liberalization dates allows the
1995 Sachs and Warner study on the relationship between trade openness and economic
growth to be extended to the 1990s. New evidence on the time paths of economic
growth, physical capital investment, and openness around episodes of trade
policy liberalization is also presented. Analysis based on the new data set suggests that
over the 1950–98 period, countries that liberalized their trade regimes experienced
average annual growth rates that were about 1.5 percentage points higher than before
liberalization. Postliberalization investment rates rose 1.5–2.0 percentage points,
confirming past findings that liberalization fosters growth in part through its effect on
physical capital accumulation. Liberalization raised the average trade to GDP ratio by
roughly 5 percentage points, suggesting that trade policy liberalization did indeed
raise the actual level of openness of liberalizers. However, these average effects mask
large differences across countries. JEL codes: F1, F4, O4
Many developing countries have embarked on programs of external economic
liberalization in recent decades. In 1960, just 22 percent of all countries, representing
just 21 percent of the global population, had open trade policies, in the
sense defined by Sachs and Warner (1995). By 2000, some 73 percent of
countries, representing 46 percent of the world’s population, were open to
international trade (figure 1).1
Romain Wacziarg (corresponding author) is an associate professor of economics at the Stanford
Graduate School of Business; his email address is wacziarg@gsb.stanford.edu. Karen Horn Welch is
director, Domestic Public Equity, at the Stanford Management Company, in Menlo Park, California;
her email address is karen.welch@stanford.edu. The authors thank Jaime de Melo, John McMillan, Paul
Segerstrom, and Jessica Wallack; an anonymous referee; and seminar participants at Columbia
University, Harvard University, the University of Colorado at Boulder, and Stanford University for
useful comments. They thank Peter Henry and Jonas Vlachos and for sharing their data. This article
was written while Romain Wacziarg was the Edward Teller National Fellow at the Hoover Institution.
1. The main reason for the discrepancy between the share of countries that are open and the share
of the world’s population living in open countries is that as of 2000, the world’s two largest countries,
China and India, remained essentially closed. Sachs-Warner (1995) classify India as open as of 1994.
The authors revisited this issue and could not confirm their finding. In fact, in terms of both policy
indicators and trade volumes, China appears to be twice as open as India. This issue is discussed later in
the article and in an appendix to the working version of this paper (Wacziarg and Welch 2003). For an
in-depth comparison of the trade regimes of India and China, see Wacziarg (2003).
THE WORLD BANK ECONOMIC REVIEW, VOL. 22, NO. 2, pp. 187–231 doi:10.1093/wber/lhn007
Advance Access Publication June 3, 2008
# The Author 2008. Published by Oxford University Press on behalf of the International Bank
for Reconstruction and Development / THE WORLD BANK. All rights reserved. For permissions,
please e-mail: journals.permissions@oxfordjournals.org
187
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
The effect of this trend toward greater trade policy openness on per capita
income growth is the topic of a large body of research. Until recently, a
growing academic consensus had emerged that both trade policy openness and
higher ratios of trade volumes to gross domestic product (GDP) were positively
correlated with growth, even after controlling for a variety of other growth
determinants. Attempts to establish a causal link also suggested a positive
impact of trade.2
In a sweeping critical survey of this literature, Rodrı´guez
and Rodrik (2000) argue that these findings are less robust than claimed,
because of difficulties in measuring openness, the statistical sensitivity of
the specifications, the collinearity of protectionist policies with other bad
policies, and other econometric difficulties. Further research on this important
topic is called for in view of the doubts their study created about the
linkages between trade openness and growth.3
Taking over where Rodrı´guez and Rodrik (2000) left off, the article pursues
three goals. The first goal is to update the Sachs-Warner classification by
FIGURE 1. Openness to Trade, 1960–2000 Note: Openness is defined
according to the Sachs and Warner (1995) criteria. Sample includes 141
countries.
Source: Authors’ analysis based on data described in the text.
2. Particularly noteworthy are the contributions of Edwards (1992), Dollar (1992), Ben-David
(1993), Sachs and Warner (1995), Ades and Glaeser (1999), and Alesina, Spolaore, and Wacziarg
(2000). Among studies trying to establish a causal link running from openness to growth or income
levels, see Frankel and Romer (1999), who measure openness by trade volumes, and Wacziarg (2001),
who captures openness by using a composite trade policy index.
3. Harrison and Hanson (1999) also criticize the Sachs-Warner classification, in a spirit similar to
that of Rodrı´guez and Rodrik. Their criticisms are revisited in detail later in the article.
188 T H E W O R L D B A N K E CO N O M I C R E V I E W
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
presenting a comprehensive cross-country database of trade indicators (tariffs,
nontariff barriers, and other measures of trade restrictions) and policy liberalization
dates for the 1990s. The second goal is to extend the Sachs-Warner
empirical results on outward orientation and growth to the 1990s. The third,
and most important, goal is to exploit the timing of liberalization in a withincountry
setting to identify the changes in growth, investment rates, and openness
associated with discrete changes in trade policy.
The availability of almost 50 years of data makes it possible to compare the
performance of countries under liberalized and nonliberalized regimes across
time. The main empirical analysis presents estimates for the within-country
response of per capita income growth, the investment rate, and the ratio of
imports plus exports to GDP to trade liberalization, controlling for country
and time effects. New evidence is presented on the within-country path of
growth in relation to the date of major trade policy changes. Evidence from the
large sample is supplemented by a discussion of several developing countries’
experiences with trade reform.
The cross-sectional results confirm recent criticisms of the Sachs-Warner
findings by showing that these were sensitive to the openness classification used
in the 1970–89 period and do not hold for the 1990s. The vast majority of
countries in the sample used here are classified as having been open during the
1990s; a simple dichotomous indicator of openness no longer discriminates
between slow- and fast-growing countries. The findings here suggest that
researchers should exercise caution when using simple dichotomous policy
indicators such as the Sachs-Warner dummy variable. However, the dates of
trade liberalization—collected by Sachs-Warner from a comprehensive survey
of a broad country-specific case literature and updated here to the late 1990s—
can be used to estimate the within-country growth and investment effects of
trade policy liberalization. In contrast to the cross-sectional findings presented
here, the results based on within-country variation suggest that over time the
effects of increased policy openness within countries are positive, economically
large, and statistically significant.
The article examines a subsample of developing countries for which detailed
information was collected on the broader economic and political context of
trade reform. It then interprets the large sample results in the context of these
country case studies. This effort reveals two lessons. First, the extent to which
per capita income growth changed after trade reforms varied widely across
countries. While the average effect obtained in the large sample is positive,
roughly half of the countries experienced zero or even negative changes in
growth following liberalization. Second, generalizations about the factors that
may explain these differences are difficult to draw. The institutional environment
of countries, the extent of political turmoil, the scope and depth of economic
reforms, and the characteristics of concurrent macroeconomic policies all
seem to have a role to play, to varying degrees in different countries. While this
article paints a picture that is highly favorable to outward-oriented policy
Wacziarg and Horn Welch 189
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
reforms on average, it cautions against one-size-fits-all policies that disregard
local circumstances.
The article is organized as follows. Section I presents an updated data set of liberalization
dates and policy openness indicators and uses the data to replicate the
Sachs-Warner growth regressions. Section II presents within-country evidence on
trade liberalization, growth, investment, and trade volumes and discusses the
timing of these effects. Section III examines 13 country cases of trade liberalization
in order to illustrate the country-specific complexities that underlie the results
from the larger sample. The last section provides some concluding remarks.
I. TR A D E L I B E R A L I Z A T I O N I N T H E 1990S
This section updates the Sachs-Warner classification and results. It also
addresses the Rodrı´guez and Rodrik critique of their study.
The Sachs-Warner Criteria
An update of the Sachs-Warner classification is called for not only because of
the problems with their classification of open and closed countries but also
because the underlying data—on tariffs, nontariff barriers, exchange rate black
market premia, socialist economic systems, and export marketing boards—are
of independent interest. This section presents a comprehensive database of
these variables for the 1990s. It also presents the results of a painstaking check
of the Sachs-Warner classification of openness and updates their data on trade
policy openness through 2000.
Sachs-Warner constructed a dummy variable for openness based on five
individual dummy variables for specific trade-related policies. A country was
classified as closed if it displayed at least one of the following characteristics:
(1) Average tariff rates of 40 percent of more (TAR).
(2) Nontariff barriers covering 40 percent or more of trade (NTB).
(3) A black market exchange rate at least 20 percent lower than the official
exchange rate (BMP).
(4) A state monopoly on major exports (XMB).
(5) A socialist economic system (as defined by Kornai 1992) (SOC).
Tariff and nontariff barriers restrict trade directly. A black market premium
(BMP) on the exchange rate could have effects equivalent to formal trade
restrictions. If, for example, exporters have to purchase foreign inputs using
foreign currency obtained on the black market but remit their foreign exchange
receipts from exports to the government at the official exchange rate, the BMP
acts as a trade restriction. On the basis of Lerner symmetry between import
tariffs and export taxes, Sachs-Warner also included the state monopoly on
exports criterion as a trade restriction. The socialist regime dummy variable
accounts for the trade-limiting aspects of centrally planned economies.
190 T H E W O R L D B A N K E CO N O M I C R E V I E W
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
It is important to distinguish the Sachs-Warner dummy variable for openness,
which pertains to the 1970s and 1980s, from the Sachs-Warner liberalization
dates, which extend from 1950 to 1994 and were compiled independently using
a different methodology. While the Sachs-Warner dummy variable was based on
the five criteria cited above, the dates of liberalization were obtained from a
comprehensive survey of country case studies of liberalization. Where possible,
the criteria used to construct the cross-sectional dummy variable for the 1970s
and 1980s were used to establish the date of liberalization. Data limitations and
lack of consistency in the definitions of the available measures of trade restrictions
across time periods, however, prevented Sachs-Warner from using their five
criteria to establish the dates of liberalization.4
The Sachs-Warner methodology
was followed as closely as possible in the update presented here.
An Openness Dummy Variable for the 1990s
The sample is based on the 118 countries included in the Sachs-Warner data
set.5
The sample also includes the new data on 23 Eastern European countries
and former Soviet republics included in version 6 of the Penn World Tables
(Heston, Summers, and Aten 2002). The openness dummy variable
(OPEN90–99) was based on the five criteria Sachs-Warner use, in order to
maintain as much consistency as possible between their data set and the data
used here. Data limitations made it impossible to update their dummy variable
to the 1990s based on exactly the same data, however.6
The main differences between the two data sets include the following:
(1) Because of data availability problems, unweighted tariff data were used
here; Sachs-Warner used own import-weighted data. Countries that exceed
the TAR threshold in the new data set based on unweighted data could
conceivably not exceed the threshold based on weighted average data. This
is unlikely to be a big problem, however, because the use of unweighted
rather weighted tariffs does not result in countries being classified differently
in the subsample in which both measures are available.
(2) Nontariff barrier data comparable to those used by Sachs-Warner are
hard to obtain. Sachs-Warner used average nontariff barrier data for
1985–88 from the Barro-Lee data set, itself based on data from the
United Nations Conference on Trade and Development (UNCTAD).
Their data cover only 29 countries for the period 1995–98. Where
4. As Sachs-Warner write (p. 24), “Our choice of dating is surely subject to further refinement. . ..
We relied on a wide array of secondary sources, which sometimes contradicted each other.” The
appendix to their article describes how they compiled their dates of liberalization and identifies the
corresponding data sources for each country in their sample. A similar appendix for the updated dates is
available in the working paper version of this study (Wacziarg and Welch 2003).
5. Sachs-Warner characterized the openness status of only 111 of these countries.
6. The data sources are detailed in Wacziarg and Welch 2003. The full data set is available in
electronic format at www.stanford.edu/~wacziarg/papersum.html. Table 1-A displays the data used to
construct the updated openness indicator.
Wacziarg and Horn Welch 191
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
comparable data on nontariff barriers were missing, the countries were
classified based only on the other four Sachs-Warner criteria. The
limited availability of nontariff barrier data for the 1990s based on
a consistent definition required the compilation of an additional nontariff
barrier data set, which may be independently useful to researchers. In
addition to the 1995–98 average core nontariff barrier data used in the
analysis, the data set contains average core nontariff barrier data for
1989–94 and 1999 data for all nontariff barriers.7
(3) Sachs-Warner relied on an export marketing index from a World Bank
study of African countries (Husain and Faruqee 1994) as the basis for
their XMB variables and on the Kornai (1992) classification of socialist
countries as the basis for their SOC dummy variable. In the absence of
updated indices from single sources, the same methodology could not be
used with the updated data. The XMB and SOC dummy variables were
therefore obtained from a comprehensive review of country case studies.
The XMB criterion is no longer confined to African countries (as it was
in Sachs-Warner), but applies to all countries in the updated data. The
definition of an export marketing board was expanded to encompass
any form of state monopoly over major exports.8
(4) Data on the BMP from Easterly and Sewadeh (2002), the primary source
for updating these data, are missing for Belarus, Tajikistan, and
Uzbekistan, and only very limited data are available for Armenia,
Azerbaijan, Georgia, the Kyrgyz Republic, and Moldova. All are classified
as open based on the overall index drawing on limited data. Whenever
BMP data were available for former Soviet republics, the data indicate that
in 2001 all of these countries except Latvia and Lithuania were closed.
(5) Sachs-Warner deviated in some cases from their self-imposed classification
rules. Some adjustments were meant to capture the fact that some
countries had undergone changes in trade policy only mid-period, so
that a classification based on period averages could be misleading. Other
adjustments were made for others’ reasons, described in their article.
Lacking objective reasons to deviate from stated rules, the updated
classification presented here abstains from any such adjustments.
Several features of the new data are worth noting. (The underlying data used
to construct the openness status dummy variable for the period 1990–99 are
displayed in table A-1.) First, 46 countries that were classified as closed by
Sachs-Warner in the 1970–89 period are classified as open in the 1990s.
7. The difference in the definitions reflects the 1999 change in UNCTAD’s reporting. Before 1999,
UNCTAD collected data on core nontariff barriers, including quotas, licensing, prohibitions, and
administered pricing. In 1999, it began reporting all nontariff barriers, which also include technical
measures and automatic licensing.
8. Wacziarg and Welch (2003) provide additional details and country-specific sources on export
marketing boards and the political transitions from socialism.
192 T H E W O R L D B A N K E CO N O M I C R E V I E W
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
Sachs-Warner characterized nine of these countries as closed based on their
dates of liberalization. Second, 30 countries were not classified in the
Sachs-Warner study, including 23 Eastern European countries and former
Soviet republics.9
Ten of these countries remained closed in the 1990s. Third,
of the 111 countries Sachs-Warner classify, 78 were closed and 33 were
open in the 1970–89 period. In the 1990s, 32 countries were closed and 79
open. Of the 141 countries classified in the new data set, 42 were closed and
99 open during the 1990s. No country that was classified as open by
Sachs-Warner in 1970–89 was classified as closed in the updated data set.
An important and often overlooked drawback of the Sachs-Warner openness
dummy variable is that it is based on averages of BMP data over each of two
decades (1970–79 and 1980–89), averages of nontariff barriers and tariffs
(TAR) over the last years of their sample period (1985–88), and end-of-period
data for the export marketing board (XMB) and socialist (SOC) dummy variables.
In the new data set, the XMB and SOC variables are based on their
1999 values rather than beginning-of-period or decade-long data, in order to
maintain as much consistency as possible with the Sachs-Warner method-
ology.10
Similarly, the nontariff barrier data are available only for 1995–98;
decade averages of the tariff data, which are available, are therefore used. As a
result, some countries classified as closed could conceivably have become open
late in the decade, and some countries classified as open could have been
closed over most of the period. Decade dummy variables thus provide only a
rough characterization of a country’s outward orientation, especially in a
decade in which many countries actively engaged in liberalization. A better
approach is to rely more on liberalization dates, as is done below.
Trade Liberalization Dates since 1994
In principle, the liberalization date is the date after which all of the Sachs-Warner
openness criteria are continuously met (data limitations often imposed reliance on
country case studies of trade policy). The choice of liberalization dates was based
on primary-source data on annual tariffs, nontariff barriers, and BMPs. A variety
of secondary sources was also used, particularly to identify when export marketing
boards were abolished and multiparty governance systems replaced
Communist Party rule. Because of data limitations, the European Bank for
Reconstruction and Development (EBRD 1994) classification and standards of
9. The other seven countries are Cape Verde, Iceland, Lesotho, Liberia, Malta, Panama, and
Swaziland. Because of lack of data, Sachs-Warner did not classify these and four other countries
(Comoros, Fiji, Seychelles, and Suriname). The new data set did not allow for the determination of the
openness status of these four countries in the 1990s.
10. Sachs-Warner’s XMB indicators are based on data from 1991; the SOC indicators are based on
data from 1987. Using 1999 data is thus consistent with their approach, however questionable that
approach may be. Most countries that abolished export marketing boards in the 1990s did so during
the first half of the decade. However, relying on end-of-period SOC data means that some Eastern
European countries and former Soviet republics are classified as open.
Wacziarg and Horn Welch 193
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
openness were used for several transition economies, just as they are in
Sachs-Warner. Table A-2 presents the dates of trade liberalization.11
Despite the clear criteria stated above, Sachs-Warner’s dates of liberalization
could not conform to their five formal criteria for openness, because comparable
data were lacking for many time periods. Hence, there is much scope for disagreement
with the Sachs-Warner classification, especially in light of new data published
since their study. Systematic review of the Sachs-Warner dates since 1990
raised questions about the liberalization status or dates for several countries.12
Sixteen countries labeled as closed at the end of the Sachs-Warner sample period
(1994) liberalized between 1995 and 2001 (table 1).13
The dates of liberalization
cited by Sachs-Warner differ in five countries (Coˆte d’Ivoire, the Dominican
Republic, Mauritania, Niger, and Trinidad and Tobago).
Thirty-five countries remained closed as of 2001, including five that were
not classified in the Sachs-Warner study and four (Belarus, Croatia, Estonia,
and India) for which the authors disagree with Sachs-Warner’s assessment
(table 2). Of 141 countries in the sample, 18 liberalized between 1995 and
2001 and 35 remained closed as of 2001.
The Rodrı´guez and Rodrik Critique
Rodrı´guez and Rodrik (2000) find that the BMP and XMB variables played a
major role in the classification of countries as open or closed. They state that a
dummy variable for openness based on the BMP and XMB criteria alone leads
to the classification of countries as open or closed that is much closer to that
generated by OPEN (the Sachs-Warner dummy variable) than one based on the
SOC, TAR, and NTB dummy variables alone. They show that the BMP and
XMB criteria generate a dummy variable that differs from the Sachs-Warner
dummy variable in only six cases, while the TAR, NTB, and SOC criteria used
jointly generated a dummy variable that differs from the Sachs-Warner dummy
variable in 31 cases. Hence, they argue that the Sachs-Warner dummy variable
for 1970–89 largely reflected the BMP and XMB criteria. Moreover, they
argue that the XMB criterion affected only the African countries (many of
which were classified as closed based on this criterion alone) and therefore
amounted to an Africa dummy variable.14
11. The working paper version of this study (Wacziarg and Welch 2003) provides detailed country
summaries of liberalization episodes, along with an explanation of the dates chosen.
12. Wacziarg and Wallack (2004) systematically checked the Sachs-Warner liberalization dates
before 1990 in a subset of their sample, uncovering little disagreement.
13. Table 1 also presents data for Cape Verde and Panama, which were not classified in the
Sachs-Warner study.
14. Sachs-Warner based the XMB criterion entirely on the Husain and Faruqee’s (1994) study of
African countries that had been involved in a World Bank or International Monetary Fund structural
adjustment program between 1987 and 1991. Rodrı´guez and Rodrik (2000) noted that Sachs-Warner
classify all but one of the Sub-Saharan African countries as closed based on the XMB criterion, which is
not applied to any other region. This study gathered and used XMB data for countries other than
African ones.
194 T H E W O R L D B A N K E CO N O M I C R E V I E W
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
To what extent are the updated Sachs-Warner data subject to the Rodrı´guez
and Rodrik critique? BMP was the sole criterion on the basis of which 26 of
TA B L E 1. Liberalization Dates of Countries That Differ from or Were Not
Included in Sachs-Warner List
Country Date of liberalization
Cape Verde 1991
Dominican Republic 1992a
Trinidad and Tobago 1992a
Coˆte d’Ivoire 1994a
Niger 1994a
Armenia 1995
Azerbaijan 1995
Egypt, Arab Rep. of 1995
Mauritania 1995a
Mozambique 1995
Tanzania 1995
Bangladesh 1996
Ethiopia 1996
Georgia 1996
Madagascar 1996
Panama 1996
Tajikistan 1996
Venezuela, R.B. de 1996
Burkina Faso 1998
Burundi 1999
Pakistan 2001
Serbia and Montenegro 2001
Sierra Leone 2001
a
Year differs from that in Sachs and Warner (1995) (see text for explanation).
Source: Authors’ analysis based on data described in the text.
TA B L E 2. Countries that Remained Closed as of 2001
Algeria Indiaa
Russian Federation
Angola Iran, Islamic Rep. of Rwanda
Belarusa
Iraq Senegal
Central African Republic Kazakhstan Somalia
Chad Lesothob
Swazilandb
China Liberiab
Syrian Arab Republic
Congo, Dem. Rep. of Malawi Togo
Congo, Rep. of Maltab
Turkmenistan
Croatiaa
Myanmar Ukraine
Estoniaa
Nigeria Uzbekistan
Gabon Papua New Guinea Zimbabwe
Haiti
a
Disagreement with Sachs and Warner (1995) (see text for explanation).
b
Not classified in Sachs and Warner (1995).
Source: Authors’ analysis based on data described in the text.
Wacziarg and Horn Welch 195
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
42 countries were classified as closed in the 1990s; XMB was the sole criterion
on which nine countries were classified as closed. Three countries were classified
as closed based on both the BMP and XMB criteria, leaving just four
countries (Bangladesh, China, India, and Pakistan) classified as closed based on
the other three criteria. Bangladesh was classified as closed based on both the
TAR and BMP criteria. China was classified as closed based on the BMP and
SOC criteria. India was classified as closed because of its tariff and nontariff
barriers. Pakistan was classified as closed because of tariffs.
The BMP and XMB criteria generated a dummy variable that differs from
the 1990–99 updated Sachs-Warner dummy variable in only two cases, while
the TAR, NTB, and SOC criteria used jointly generate a dummy variable that
differs from the updated Sachs-Warner dummy variable in 38 cases.15
The
openness status dummy variable for 1990–99 is thus subject to the same criticisms
Rodrı´guez and Rodrik lodged against the Sachs-Warner classification for
the 1970–89 openness dummy variable.
The Rodrı´guez and Rodrik critique is valid in terms of country status based on
the OPEN90–99 dummy variable. It is less valid for the liberalization dates. As
most countries were classified as closed based on the XMB and BMP criteria, not
surprisingly, when they open up these variables change. The XMB and BMP variables
determined the year of liberalization in many countries that opened up
during the 1990s. The exceptions tend to be Eastern European countries and
former Soviet republics, which opened based on the SOC criterion (general
reforms related to liberalization). The TAR criterion was not a decisive factor in
assigning a liberalization date for any country; NTB was the determining factor
only in Panama. However, policy changes that reduced the BMP or removed
XMBs were generally accompanied by changes in the levels of other types of trade
barriers, such as tariff and nontariff barriers, that had initial values below the
Sachs-Warner thresholds of 40 percent. Hence, liberalization dates do not simply
capture changes in the BMP and XMB variables, but they also reflect broader
liberalization. Given that the dates of liberalization in the new data set were
cross-checked against a case study literature of outward-oriented reforms in
developing countries, it is likely that they reflect important shifts in trade policy.16
Updating the Sachs-Warner Results
The Sachs-Warner study attracted considerable attention in part because their
estimated effect of the cross-sectional dummy variable for openness in explaining
annual growth between 1970 and 1989 was very large (about 2 percentage
15. Among the countries in which the TAR, NTB, and SOC dummy variables and the updated
Sachs-Warner dummy variable disagree, 20 are in Africa and 10 are Eastern European countries or
former Soviet republics. These countries were classified as closed based on either the XMB criterion or
the BMP criterion, or both.
16. Wacziarg and Wallack (2004) show that the Sachs-Warner liberalization dates are good
indicators of the timing of major trade policy changes by thoroughly checking these dates against the
case study literature of trade liberalization in 25 developing countries.
196 T H E W O R L D B A N K E CO N O M I C R E V I E W
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
points). The updated data on trade policy openness make it possible to extend
the Sachs-Warner regressions through the late 1990s. As this is not the main
focus of this article, these results are reported only briefly.
As a consistency check, the Sachs-Warner regression was first replicated for
1970–89 (column 1 in table 3 replicates column 7 in Sachs-Warner’s table 11).
The only difference is that the new calculations are based on a newer release of
the Penn World Tables data (version 6 instead of version 5). The openness
dummy variable for 1970–89 enters highly significantly, with a magnitude of
1.98 percentage points of annual growth. This result is consistent with the
results in Sachs-Warner, who find a coefficient of 2.2. In contrast, the updated
Sachs-Warner dummy variable enters insignificantly in the same specification
for the 1990s (column 2 of table 3).
The cross-sectional effect of openness on growth was estimated by constructing
openness indicators based on the dates of liberalization. The openness
status for 1980, for example, takes on a value of 1 if a country had liberalized
by 1980 and a value of 0 otherwise. Subsequent growth (after 1980) can then
be regressed on this variable and other controls. Dummy variables were constructed
for each decade (1970, 1980, and 1989) in this fashion. An advantage
of this method over the period-specific dummy variables is that the periodspecific
dummy variables are based partly on information from the end of the
period (TAR, NTB, XMB, and SOC) and partly on period averages (BMP).
Constructing openness indicators based on the dates of liberalization instead
isolates only the countries that were open at the beginning of a period.
The econometric specification is identical to that in Sachs-Warner; it restricts
the time span of each regression to a single decade. The effect of the liberalization
status in the 1970s is weaker and smaller than in the 1980s but positive and
significant at the 90 percent level. The Sachs-Warner results were likely driven
by the strong effect of liberalization on growth in the 1980s (columns 3 and 4 of
table 3). This effect is positive but statistically indistinguishable from zero in the
1990s when countries are grouped according to their liberalization status as of
1989. These results suggest that the Sachs-Warner cross-sectional findings are
highly sensitive to the decade under consideration and that the updated openness
indicator can no longer effectively distinguish fast-growing from slow-growing
countries.17
I I. W I T H I N -CO U N T R Y LI B E R A L I Z A T I O N DY N A M I C S
This section argues that better use can be made of data on the dates of liberalization.
With almost 50 years of data on growth and openness, it is possible to
17. Wacziarg and Welch (2003), who conduct many more replications of the initial Sachs-Warner
cross-sectional findings, conclude that no matter how the liberalization dummy variable was defined,
the results for the 1990s show an insignificant effect of the updated dummy variable on growth. This
result is in sharp contrast with the results for the 1970–89 period.
Wacziarg and Horn Welch 197
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
TA B L E 3. Replication of Sachs-Warner Cross-sectional Regressions
(1) (2) (3) (4) (5)
Variable
Growth
1970–89
Growth
1989–98
Growth
1970–80
Growth
1980–89
Growth
1989–98
Real GDP per capita (t) 21.5929 21.150 21.292 21.397 21.261
(4.89) (1.95) (2.83) (3.84) (2.13)
Sachs-Warner openness
dummy variable(1970–89
or 1990–98 periods)
1.9845 0.136
(3.87) (0.21)
Openness status based on
liberalization dates (t)
1.387 2.574 0.521
(1.86) (4.17) (0.84)
Secondary-school enrollment
rate (t)
0.8059 4.689 0.169 1.822 4.872
(0.68) (2.43) (0.10) (1.40) (2.52)
Primary-school enrollment
rate (t)
1.4003 1.381 2.455 20.139 1.616
(1.65) (0.86) (2.01) (0.11) (0.99)
Government Consumption to
GDP ratio (t, t þ X)
20.0844 20.063 20.005 20.065 20.059
(3.02) (1.32) (0.19) (2.51) (1.26)
Number of revolutions per
year (t, t þ X)
20.4359 20.986 21.238 20.211 21.030
(0.58) (1.08) (1.12) (0.21) (1.13)
Number of assassinations
per capita per year
(t, t þ X)
0.0296 0.483 0.276 0.188 0.473
(0.13) (1.56) (0.94) (0.54) (1.54)
Deviation of the price level
of investment (t), as in
Sachs-Warner
20.1709 20.734 20.476 0.350 20.721
(0.53) (1.24) (0.99) (0.87) (1.23)
Gross domestic investment/
real GDP (t, t þ X)
0.0757 0.051 0.076 0.103 0.040
(2.64) (1.01) (2.02) (2.30) (0.76)
Extreme political repression
(from Sachs-Warner)
20.6974 0.165 20.907 20.780 0.224
(1.66) (0.28) (1.47) (1.51) (0.38)
Population density (t 2 10) 0.0006 0.0009 0.001 0.001 0.001
(0.90) (1.40) (0.60) (0.87) (1.49)
Intercept 12.2482 7.752 9.334 10.635 8.288
(4.87) (1.81) (2.84) (3.86) (1.92)
Adjusted R2
0.546 0.211 0.35 0.53 0.32
Number of observations 91 89 99 97 89
Note: Numbers in parentheses are t-statistics. The beginning date of each period (1970 in
columns 1 and 3, 1980 in column 4, and 1989 in columns 2 and 5) is denoted by t. (t, t þ X)
denotes the average computed between dates t and t þ X (X ¼ 20 in column 1 and 10 in columns
2–5). The dependent variable is defined as the real annual per capita growth rate of GDP in the
relevant period.
Source: Authors’ analysis based on data described in the text. Growth, income, and investment
data are from Heston, Summers and Aten (2002).
198 T H E W O R L D B A N K E CO N O M I C R E V I E W
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
assess the within-country effects of discrete changes in trade policy openness.18
This section compares the means of economic growth and other variables of
interest, such as physical capital investment rates and trade volumes, under liberalized
and nonliberalized regimes.
Liberalization and Growth
Fixed-effects regressions of growth on a binary liberalization indicator, defined
by the dates of liberalization, were run to assess the within-country effect of
growth on liberalization. The regressions amount to difference regressions in
growth or difference-in-difference regressions in log income:
log yit À log yitÀ1 ¼ ai þ bLIBit þ 1itð1Þ
where yit is per capita income in country i at time t and LIBit ¼ 1 if t is greater
than the year of liberalization and no reversals of the trade policy reforms have
occurred, and 0 otherwise. The sample was not restricted to countries that
underwent reforms. The residual term is modeled as 1it ¼ vi þ ht þ mit and in
all regressions, the vi terms are treated as country fixed effects and ht terms as
fixed effects.
Over the sample period 1950–98, 31.7 percent of country-year observations
occur in a liberalized regime (LIBit ¼ 1) (table 4). The conditional mean of
annual growth of per capita GDP given that a country is liberalized is 2.71
percent, while the mean is 1.18 percent in a nonliberalized regime, a difference
of 1.53 percentage points of annual growth. These simple conditional means
are based on both cross-sectional and within-country variation.
Panel (1) of table 5 displays country and time fixed-effects regressions of
growth on the liberalization indicator, in order to isolate within-country variation.
The regression for 1950–98 indicates a within-country difference in
growth between a liberalized and a nonliberalized regime of 1.42 percentage
points (column 1). This coefficient is estimated with a high level of
statistical precision (the t-statistic exceeds 5).19
The estimated within-country
18. Sachs and Warner provide some within-country evidence on liberalization and growth for a
sample of 37 reformers, presenting estimates for one fixed-effects regression of growth on dummy
variables for three time periods around liberalization episodes. They show that average growth was
depressed by 0.88 percentage points in the three years before liberalization, rose 1.09 percentage points
a year in the three years following liberalization, and rose 1.33 percentage points a year thereafter
relative to growth in the three years before liberalization. These limited results are of the same order of
magnitude as the more detailed research presented here, which investigates the robustness of these
estimates, extends them in time (the sample period spans 1950–98 rather than 1966–93) and space (the
sample includes up to 133 countries rather than 37), and presents new evidence on investment and
openness.
19. This effect was estimated allowing for first-order autocorrelation of the residuals, using the
Baltagi-Wu fixed-effects method. The coefficient on liberalization was 1.32, with a t-statistic of 4.14, in
line with the fixed-effects results reported here. The simpler fixed-effects estimates, with t-statistics
based on robust standard errors, are reported here because of concerns over the small T properties of
the Baltagi-Wu estimator, particularly when the sample is restricted to specific decades.
Wacziarg and Horn Welch 199
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
effect increases over time, reaching its maximum in the 1990s (column 2–4).
These results stand in sharp contrast to the cross-sectional results: countries
that liberalized in the 1990s experienced a larger postliberalization increase in
growth than countries that liberalized in any other decade. Indeed, the estimated
difference in growth in the 1990s is roughly 2.55 percentage points.
TA B L E 5. Fixed-Effects Regressions of Growth, Investment, and Openness on
Liberalization Status, 1950–98
Item (1) 1950–98 (2) 1950–70 (3) 1970–90 (4) 1990–98
Dependent variable: Growth
Liberalization 1.417 0.611 1.787 2.547
(5.05) (1.29) (3.11) (2.39)
Number of observations 4,936 1,728 2,312 1,116
Number of countries 133 108 112 133
Adjusted R2
0.05 0.03 0.04 0.04
Dependent variable: Investment rate
Liberalization 1.937 2.545 1.237 0.762
(9.06) (7.57) (2.91) (2.16)
Number of observations 5,078 1,844 2,321 1,140
Number of countries 136 110 117 136
Adjusted R2
0.04 0.10 0.11 0.02
Dependent variable: Openness
Liberalization 5.531 2.302 4.097 –1.803
(7.42) (1.89) (3.74) (0.83)
Number of observations 5,078 1,844 2,321 1,140
Number of countries 136 110 117 136
Adjusted R2
0.22 0.02 0.14 0.08
Note: Numbers in parentheses are robust t-statistics. Regressions are based on the specifications
in equations (1)–(3).All regressions include time and country fixed-effects (estimates not
reported).
Source: Authors’ analysis based on data described in the text.
TA B L E 4. Summary Statistics for Variables Used in Fixed-Effects Regressions
Variable
Number of
observations Mean
Standard
deviation Minimum Maximum
Liberalization 7,191 0.317 0.465 0.0 1.0
Investment rate
(percent)
5,078 15.291 9.128 23.590 52.880
Openness ratio 5,078 60.505 42.880 3.110 473.860
Growth (annual
percent)
4,936 1.784 6.153 248.732 43.754
Per capita GDP
(purchasing power
parity US$)
5,072 5,739.380 5,826.636 276.000 39,129.000
Source: Authors’ analysis based on data described in the text.
200 T H E W O R L D B A N K E CO N O M I C R E V I E W
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
Liberalization and Investment
The empirical literature on trade and growth suggests that the effects of liberalization
on economic growth are mediated largely by the rate of physical capital
investment. Several researchers, including Levine and Renelt (1992), Baldwin
and Seghezza (1996), and Wacziarg (2001), suggest that the investment rate is
an important channel linking trade and growth. This finding is based largely
on cross-country findings. Fixed-effects regressions of investment rates on the
liberalization indicator were run in order to investigate this issue in a withincountry
context:
Iit
Yit
¼ hi þ fLIBit þ vitð2Þ
where Iit is physical capital investment and Yit is GDP in country i at time t,
and vit captures country and year effects.
Panel (2) of table 5 reports the estimates of such regressions. The withincountry
evidence confirms past cross-country findings. For the period 1950–
98, countries with liberalized regimes experienced average rates of physical
capital investment that were 1.94 percentage points higher than those of
countries with nonliberalized regimes. This represents 20 percent of this variable’s
standard deviation in the pooled sample. The effect is largest in the
initial period of the sample (1950–70).
Fixed-effects regressions of growth on the investment rate were run in order
to get a rough notion of how much of the effect of trade policy openness on
growth can be attributed to the investment channel. The coefficient on investment
in the baseline 1950–98 regression was 0.15 percentage points, with a
t-statistic of 8.05.20
The effect of liberalization on investment in the corresponding
regression was 1.94 percentage points. Multiplying the two yields an
estimate of the effect of liberalization on growth through investment of roughly
0.29 percentage points, about 21 percent of the total effect of liberalization on
growth. The analysis provides suggestive evidence that investment constitutes
an important channel through which trade-centered liberalization affects
growth within countries.
Liberalization and Openness
Is trade policy liberalization followed by a break in the volume of trade, as
measured by the ratio of imports plus exports to GDP? If this is the case, it suggests
that liberalization did increase the level of openness of the economy. Determining
this effect is important, because announced reforms may be poorly implemented
or counteracted by alternative trade barriers. If trade liberalization is associated
20. The full results are presented in the working paper version of this study (Wacziarg and Welch
2003).
Wacziarg and Horn Welch 201
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
with increases in trade volumes, one could be more confident that it actually raised
the level of exposure of the reforming country to the world economy.21
This issue is examined by running the following regression:
Xit þ Mit
Yit
¼ vi þ dLIBit þ witð3Þ
where Xit denotes exports and Mit denotes imports. The results suggest that liberalization
raises openness by 5.53 percentage points of GDP for the full sample
period (Panel (3) of table 5). This effect is indistinguishable from zero in the
1990–98 time period, however, perhaps because more time is needed to observe
the effects of recent liberalizations on trade volumes. In most periods, however,
trade liberalization is associated with sustained and large increases in the effective
level of exposure of the typical reforming country to the world economy.
Timing of Effects
The simple average difference between growth in nonliberalized and liberalized
regimes may mask important timing issues. It provides no information on how
soon the effects occur or whether they cease to be felt a few years after reform.
This subsection examines the time path of growth, investment, and openness
for an average country before and after liberalization.
Average annual growth rates, investment rates, and openness ratios are displayed
in figures 2 through 4 for 20 years before and 20 years after liberalization
in a sample of 81 countries that underwent permanent liberalizations (that
is, liberalizations that were not reversed as of 2000). As several countries had
varying numbers of years of data before and after their liberalization, the
average at each point in time is based on different samples of countries.22
Several observations can be made about the figures. First, despite not controlling
for any fixed effects, the increase in growth following liberalization is
remarkably similar to that shown in table 5: growth before trade-centered
reforms averages 1.5 percent and rises to roughly 3 percent postreform
(figure 2). Second, there does not seem to be a strong time pattern: the effects
appear to be immediate and do not die out after a few years. Third, the few
years immediately preceding liberalization are low-growth years: reforms are
often preceded by downturns or crises.
The investment rate seems to take off during the 10 years following liberalization
and remain high thereafter (figure 3). The plotted effect seems larger than
that uncovered in the fixed-effects regressions. Openness follows a more or less
21. Even absent effects on actual openness, liberalization could still have effects on growth and
investment, through pro-competitive effects or technological transfers, for example.
22. The figures did not look different when the sample was restricted to countries with continuously
available data. The availability of data forced a reduction in the time span to eight years before and
after liberalizations and in the country coverage to 39 countries. These figures are available in the
working paper version of this study (Wacziarg and Welch 2003).
202 T H E W O R L D B A N K E CO N O M I C R E V I E W
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
linear upward trend, without an apparent break at the date of liberalization
(figure 4). More formal tests based on fixed effects did reveal an effect attributable
to liberalization, even after controlling for time fixed effects, however.
Dummy variables for four (nonoverlapping) periods surrounding the reforms
were defined in order to further examine the timing of the growth, investment,
and openness responses to liberalization. Fixed-effects regressions were then run
on growth, investment, and openness. The specification is as follows:
log yit À log yitÀ1 ¼ ai þ b1D1it þ b2D2it þ b3D3it þ b4D4it þ 1itð4Þ
where D1it ¼ 1 if T 2 3 t T 2 1 and zero otherwise; D2it ¼ 1 if T t
T þ 2; D3it ¼ 1 if T þ 3 t T þ 6, and D4it ¼ 1 if t . T þ 6; and T denotes
the date of liberalization. The coefficients on these dummy variables capture
the average difference in growth between these years and the period preceding
three years before liberalization (the baseline period). The corresponding specifications
for the investment rate and openness ratio were also run (table 6).23
FIGURE 2. Sample Means for Growth before and after Liberalization
Source: Authors’ analysis based on data described in the text.
23. Countries that experienced policy reversals or multiple liberalizations, for which definitions of
the dummy variables are not straightforward, had to be dropped. Dropping these variables reduced the
size of the sample for the growth regression from 133 to 118 countries.
Wacziarg and Horn Welch 203
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
The results are consistent with the observations made about figures 2–4.
Countries that liberalize often do so following periods of economic turmoil:
growth is depressed by 0.55 percentage points in the three years before liberalization
relative to the preceding years. Tornell (1998) shows that 60
percent of episodes of economic reform, including trade reform, occur in the
aftermath of a domestic political or economic crisis. Measuring growth
differences relative to “early prereform” outcomes prevents falsely attributing
to reforms growth differences that stem from depressed economic circumstances
in the years immediately preceding the reforms. In the three years
following liberalization, growth rises slightly (by 0.30 percentage points),
but the effect is statistically indistinguishable from zero. Sustained growth
differences become apparent three years after reform, with annual increases
in growth of 1.44 points in period T þ 3 to T þ 6 and of 1.0 percentage
point after that relative to the baseline period. The typical timing pattern
revealed by these regressions shows growth to be slightly depressed before
liberalization and to increase 1.0–1.5 percentage points three years after
reforms. A similar pattern applies to investment and openness. These estimates
reflect sample averages and may mask interesting country-specific
differences, as discussed below.
FIGURE 3. Sample Means for Investment before and after Liberalization
Source: Authors’ analysis based on data described in the text.
204 T H E W O R L D B A N K E CO N O M I C R E V I E W
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
FIGURE 4. Sample Means for Openness before and after Liberalization
Source: Authors’ analysis based on data described in the text.
TA B L E 6. Fixed-Effect Regressions: Timing of the Effects of Liberalization on
Growth, Investment, and Openness
Item (1) Growth (2) Investment (3) Openness
D1 20.555 21.040 21.979
(1.14) (2.88) (1.32)
D2 0.300 20.160 0.795
(0.61) (0.41) (0.63)
D3 1.438 1.197 3.606
(3.27) (2.98) (2.21)
D4 1.015 2.129 13.371
(2.30) (5.47) (9.17)
Number of observations 4,230 4,357 4,357
Number of countries 118 121 121
Adjusted R2
0.04 0.08 0.26
Note: Number in parentheses are robust-statistics. Regressions are based on the specification
in equation (4). All regressions include time and country fixed-effects (estimates not reported).
Definition of dummy variables, where T represents the date of liberalization, is as follows: D1 ¼ 1
if T 2 3 t T 2 1 and zero otherwise. D2 ¼ 1 if T t T þ 2 and zero otherwise. D3 ¼ 1 if
T þ 3 t T þ 6 and zero otherwise. D4¼ 1 if t . T þ 6 and zero otherwise.
Source: Authors’ analysis based on data described in the text.
Wacziarg and Horn Welch 205
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
Concurrent Policies
It is difficult to attribute differences in growth purely to trade liberalization.
Countries carrying out trade reforms often simultaneously adopt policies favoring
domestic deregulation, privatization, and other microeconomic reforms and
macroeconomic adjustments, making it difficult to interpret the coefficient on
liberalization in a within-country growth regression as the total effect of trade
liberalization per se.24
A more realistic interpretation of these estimates is that
they capture the impact of trade-centered reforms more broadly. In what
follows, we describe our efforts to address this important concern.
SCOPE OF REFORMS. The working paper version of this study (Wacziarg and
Welch 2003) distinguishes countries that carried out overall reforms from those
that carried out external sector reforms in relative isolation from other domestic
reforms. Wacziarg and Wallack (2004) examine 22 episodes of trade liberalization,
most of them in developing countries in the 1980s. Fourteen of these
episodes were accompanied by market-oriented domestic reforms; eight
occurred in relative isolation from major shifts in domestic policy. The distinction
between pure trade reforms and overall reforms was based largely on
whether the countries implemented a substantial program of privatization and
deregulation at the same time as trade reforms.
Isolating the sample of countries that were part of the Wacziarg and Wallack
(2004) study and examining whether the within-country effects of liberalization
on growth differed between trade reformers and overall reformers reveal
several noteworthy findings. First, even though the sample was restricted to 22
countries, the estimates were remarkably similar to those obtained for the full
sample of 133 countries. Second, the estimates of the impact of trade liberalization
in countries that carried out trade reforms in isolation were similar to the
corresponding estimates for countries that also reformed their domestic sectors,
despite the crude nature of the distinction between overall reformers and pure
trade reformers. While the interpretation of these suggestive results requires
caution, a plausible conclusion is that the effect of trade-centered reforms is in
large part attributable to an external reform component. This issue is further
addressed below in the context of individual country experiences.
OTHER EXTERNAL REFORMS. Trade reforms are sometimes associated with other
types of external reforms, such as capital market liberalization. To the extent
such reforms are adopted simultaneously, estimates may capture the impact of
these financial reforms rather than trade reforms. This argument is frequently
invoked to criticize the type of estimates presented above.
24. An analogous point is often made in a cross-country context. Rodrı´guez and Rodrik (2000) and
other observers suggest that “bad” government policies tend to go together, making it difficult to
disentangle the effects of protectionist trade policy from those of poor macroeconomic management,
poor governance, or poor institutions in general.
206 T H E W O R L D B A N K E CO N O M I C R E V I E W
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
This issue is investigated by looking at data on the timing of financial reforms.
Bekaert, Harvey, and Lunblad (2001) examine the impact of capital market liberalization
on economic growth in a panel context, using both cross-sectional
and within-country time variation. Using data from Bekaert and Harvey (2000)
on the dates of official regulatory reforms pertaining to financial markets, they
find robust positive effects of financial liberalization.25
Their dates are compared
with the dates of trade liberalization in the data set used here.
Bekaert and Harvey (2000) characterize the date of official financial liberalization
for 40 of the 106 countries in the sample that had liberalized by 2001.26
Of these, only two (Brazil and Turkey) have exactly the same year of official
financial regulatory reform and trade liberalization. Only nine countries
implemented financial sector reforms within three years before and after the
date of trade liberalization, and just 17 did so within five years before and
after. Many countries that enacted trade reforms never enacted financial liberalization,
so the numbers cited above overstate the extent of coincidence between
financial and trade liberalization dates. There is thus little evidence that trade
reform and financial market liberalization occur concurrently and that the estimates
may confound the effects of the two types of reform.
III . CO U N T R Y CA S E ST U D I E S
The econometric results presented above summarize the effect of trade liberalization
on growth and other variables for a sample of very diverse countries.
Fixed-effects regressions allow all time-invariant country characteristics to be
controlled for. The estimated coefficients on liberalization are not countryspecific,
however; they represent average responses. The reaction of individual
countries to reforms is likely to vary, especially as the depth and scope of
reforms differed across countries.
Much can be learned from the considerable heterogeneity in the response of
growth to trade reform. This section examines specific cases of reform in
countries representative of the broader sample for which enough data on
growth, investment, and openness are available before and after reforms. The
goal is to get a sense of the subtleties of reform in specific cases and to illustrate
the economic mechanisms that give rise to the average estimated effects.
The time paths of growth, investment, and openness are first examined for a
subsample of 24 developing countries for which data are available for at least
eight years before and after liberalization. A more detailed discussion then
focuses on 13 of these countries.
25. Henry (1999, 2000) uses data on economic and political reforms for a smaller set of 18
developing countries.
26. Details of the comparison between the Bekaert and Harvey (2000) dates and the dates presented
here are available on request.
Wacziarg and Horn Welch 207
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
The average difference in growth, investment rates, and openness ratios
between the pre- and postliberalization periods is shown for 24 countries
(table 7). The countries were chosen from the sample of 39 countries for which
at least eight years of data are available on either side of the date of liberalization,
restricting the sample to emerging markets, the main focus of this study.
The data reveal positive growth differences in 13 of the 24 countries and
negative differences in six of them; the remaining five countries exhibit an
effect close to zero. Postliberalization growth effects appear large in Mauritius,
Indonesia, Uruguay, Republic of Korea, Chile, Taiwan (China), and Uganda.
Among countries that experienced positive differences, the magnitude of the
growth increase ranged from 0.83 percentage points of per capita income
growth in Poland to 3.62 points in Mauritius. The range of growth decline was
of a similar magnitude.
Before and after comparisons of investment rates and openness also reveal
large variations across countries. The postliberalization surge in investment
rates was particularly strong in the Republic of Korea, Taiwan (China),
Indonesia, Jordan, and Guinea-Bissau. About half of the 24 countries exhibited
zero or negative differences in investment rates.
TA B L E 7. Mean Growth, Investment, and Openness Changes in 24 Countries
Country
Growth
difference
Investment
difference
Openness
difference
Year of
liberalization
Sample
period
Mauritius 3.62 0.34 35.90 1968 1951–98
Indonesia 3.32 9.80 25.96 1970 1961–98
Uruguay 3.08 21.01 11.22 1990 1951–98
Korea, Rep. of 3.02 18.44 43.40 1968 1954–98
Chile 2.80 21.12 26.33 1976 1952–98
Taiwan 2.29 9.91 55.77 1963 1952–98
Uganda 2.24 1.63 26.60 1988 1951–98
Ghana 1.99 23.91 9.13 1985 1956–98
Guinea 1.85 22.74 7.28 1986 1960–98
Guyana 1.80 27.49 84.49 1988 1951–98
Benin 1.74 1.64 8.72 1990 1960–98
Mali 1.19 0.86 15.68 1988 1961–98
Poland 0.83 24.30 3.35 1990 1971–98
Paraguay 0.42 2.01 49.71 1989 1952–98
Cyprus 0.34 24.05 29.13 1960 1951–96
Colombia 0.18 0.48 5.91 1986 1951–98
Tunisia 20.30 25.58 31.94 1989 1962–98
Philippines 20.40 1.03 39.54 1988 1951–98
Israel 20.96 26.10 21.42 1985 1951–98
Botswana 21.99 3.98 22.27 1979 1961–98
Mexico 22.16 24.59 17.56 1986 1951–98
Hungary 22.41 21.19 24.17 1990 1971–98
Guinea-Bissau 22.95 5.59 9.89 1987 1961–98
Jordan 24.28 5.75 40.61 1965 1955–98
Source: Authors’ analysis based on data described in the text.
208 T H E W O R L D B A N K E CO N O M I C R E V I E W
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
Closer examination of postliberalization changes in growth, investment, and
openness for a restricted sample of developing countries thus reveals considerable
heterogeneity in their experiences with reform. The following case studies
develop hypotheses that could account for these differences.
From the sample of 24 developing countries for which there are at least
eight years of data on either side of liberalization, a subsample of 13 countries
was selected to study in greater detail. A set of countries was chosen that was
small enough to allow their preexisting conditions, overall policy environment,
and macroeconomic circumstances to be examined while maintaining a geographically
diverse sample reflecting the range of country-specific growth
effects identified above. The goal was to uncover patterns that could explain
cross-country differences in individual countries’ responses to liberalization and
suggest directions for future research.
The subsample was selected to include a geographically diverse set of
countries that experienced growth effects of liberalization in roughly the same
proportions as the 24 countries discussed above. It includes 13 countries, seven
of which experienced higher mean growth rates following liberalization
(Indonesia, Republic of Korea, Chile, Taiwan (China), Uganda, Ghana, and
Poland). The growth difference was negative in four countries (Israel,
Botswana, Mexico, and Hungary). In two countries, Colombia and the
Philippines, liberalization was associated with roughly zero difference in their
mean growth rates. Table A-3 describes all countries’ concurrent reforms,
macroeconomic environment, and political context.
Examination of these case studies suggests that the packaging and timing
of reforms are important factors in explaining differences in postliberalization
growth patterns. Countries that followed through by deepening trade
reforms over time did better than countries that did not. Neither active governmental
disengagement from industrial policy nor broad-based reforms
were necessary conditions for success. Countries that counteracted shortlived
programs of external liberalization with domestic interventions and
countries that adopted tight macroeconomic policies, faced unfavorable
terms of trade shocks, or suffered from political instability did not perform
as well as other countries.
Sustained Reforms
In the majority of countries that experienced higher growth following liberalization,
trade reforms were not strictly limited to the period of liberalization;
these countries continued to deepen trade reform after liberalization. Chile, for
example, which liberalized in 1976, recovered from the Latin American debt
crisis and continued to grow during the late 1980s. During this period, it
decreased tariffs and implemented several bilateral free trade agreements. Both
Korea and Taiwan (China), which liberalized in the 1960s, continued to lower
tariffs and remove nontariff barriers, particularly during the mid-1980s and
Wacziarg and Horn Welch 209
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
1990s. Indonesia sustained the initial reforms of 1970 with reductions in
export duties in 1976 and additional trade-centered liberalization throughout
the 1980s. In Uganda, the 1988 liberalization was followed by a second wave
of external reforms in 1993–94.
Scope of Reforms
Whether trade reforms were part of a package of other domestic reforms or
occurred in relative isolation does not seem to help predict the effect on
growth. Among countries that implemented broad-based reforms, and in which
postliberalization growth increased, Chile and Poland stand out as prototypical
success stories of reform. Both implemented broad-based domestic reforms, of
which trade liberalization was only a part. In Colombia, Hungary, and
Mexico, which Wacziarg and Wallack (2004) classify as broad-based reformers,
average growth following liberalization actually fell. Political instability
is probably at the heart of Colombia’s lack of increased growth. In Hungary,
the decline may have occurred because the domestic portion of the reform
program (banking sector reforms, privatizations) was in large measure delayed
until 1995. To the extent that external and domestic reforms are complementary,
the full effects for Hungary may not be apparent in the growth data,
which extend only to 1998.
The case of Mexico is more complex. The privatization program began
before trade liberalization, in 1984, with the sale of small- and medium-sized
businesses, and continued after 1986, with the sale of larger enterprises, such
as the national telephone company, parts of the banking industry, and the
national airline. While Mexico maintained large government oligopolies that
prevented broad industrial restructuring and resource reallocation, one can
hardly argue that its entry into the General Agreement on Tariffs and Trade
(GATT) in 1986 and the concurrent reduction in external barriers occurred in
isolation from other domestic reforms.
The flip side of this coin is a country like Ghana, which, according to Wacziarg
and Wallack (2004), implemented trade reforms in relative isolation (privatization,
for instance, did not begin until the early to mid-1990s). It experienced a 2
percentage point increase in mean growth after the 1985 liberalization.
Other interesting cases are the success stories of Southeast Asia, where many
economies, including Korea and Taiwan (China), implemented policies aimed
at increasing foreign direct investment (FDI) at the same time or after external
liberalization. Indonesia, Korea, and Taiwan (China) pursued growth strategies
with widespread government involvement in the economy. In Indonesia,
government involvement increased during the 1970s, after external liberalization
began. Both Korea and Taiwan (China) adopted activist industrial policies,
with the government involved in “picking winners.” That the growth
performance of these countries was unprecedented the 1998 Asian crisis shows
that government disengagement from the economy is not a necessary condition
for successful trade reforms. What all these countries shared was an
210 T H E W O R L D B A N K E CO N O M I C R E V I E W
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
outward-oriented development model in which increasing exports was a
central pillar of the growth strategy.
One cannot point to the breadth of reform as an unambiguous criterion
explaining differences in the growth response to liberalization. The picture that
emerges is far from simple. The set of economies that experienced higher
growth following liberalization includes both those that maintained heavy government
involvement in the economy (Indonesia, Korea, and Taiwan [China])
and those that actively reduced the role of government (Chile and Poland). The
set of countries that experienced negative or zero growth differentials after liberalization
includes Colombia, Hungary, and Mexico, countries that actively
disengaged the government from domestic economic activity at the time of
trade reforms.
Counteractive Policies
Some of the 13 countries in the sample implemented policies that actively
counteracted the effects of trade reform and as a result did not experience
increases in growth rates.27
In Israel, social pacts based on broad coalitions of
labor, government, and industry set the patterns for prices, wages, and the
exchange rate in ways that mitigated the effects of trade openness on domestic
producers. In the Philippines, trade liberalization was accompanied by a large
increase in the share of state-owned enterprises in the economy, including a
doubling of the share in GDP of financial transfers from the government to
state-owned enterprises between 1987 and 1989. Such interventions, designed
partly to protect domestic producers in the face of increased import competition,
may have precluded the realization of gains from trade.
Macroeconomic Factors
Countries that did not experience growth increases after liberalization often
suffered from mitigating circumstances, associated in particular with restrictive
macroeconomic policies or terms of trade shocks. In Hungary and Mexico,
two countries in which growth fell following liberalization, trade reform was
followed by tight monetary policies involving high interest rates, which
depressed growth. In Mexico, currency overvaluation undid the effects of trade
liberalization in the late 1980s and early 1990s.
In Botswana, terms of trade considerations account for the absence of a
postliberalization growth surge. Volatility on world diamond markets increased
shortly after Botswana implemented trade reforms in 1979. The weak diamond
marked caused a recession in 1981–82 that resulted in a postliberalization
growth rate that was about 2 percentage points lower than the preliberalization
rate. Thus, terms of trade considerations are essential in accounting for the
absence of a postliberalization growth surge in Botswana.
27. Wacziarg and Wallack (2004) discuss some of these cases in greater detail.
Wacziarg and Horn Welch 211
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
Political Instability
Several countries suffered from severe forms of political instability, preventing
realization of the gains from trade liberalization. A prime example is
Colombia, where instability persisted throughout the 1990s. Other examples
include Israel and the Philippines. In contrast, economies that seem to have
experienced higher growth following reform also witnessed periods of relative
political stability. Taiwan (China) is a case in point, as are Chile, Indonesia,
and Korea, where liberalization coincided roughly with the rise to power of
authoritarian regimes, resulting in a degree of lasting political stability following
periods of political unrest.
I V. CO N C L U S I O N
This article presents an updated data set of trade policy indicators and liberalization
dates. It revisits the evidence on the cross-country effects of
Sachs-Warner’s simple dichotomous indicator of outward orientation on economic
growth, confirming the pitfalls of this indicator first identified by
Rodrı´guez and Rodrik (2000). It shows that the Sachs-Warner dichotomous
indicator effectively separates fast-growing from slow-growing countries in the
1980s and to a lesser extent in the 1970s, but fails to do so in the 1990s.
Simple dichotomous indicators of outward orientation are too crude to capture
the complexities of trade policy.
Instead, liberalization dates that capture episodes of discrete shifts in trade
policy can be useful for estimating within-country growth responses The
Sachs-Warner dates of liberalization were painstakingly checked and updated,
based on quantitative data and a thorough review of country-specific case
studies of reform. The new and robust evidence indicates that these dates of
liberalization mark breaks in growth, investment, and openness within
countries. Over the 1950–98 period, countries that liberalized their trade
regimes experienced average annual growth rates that were about 1.5 percentage
points higher than before liberalization. Postliberalization investment rates
rose 1.5–2.0 percentage points, confirming past findings that liberalization
fosters growth in part through its effect on physical capital accumulation.
Liberalization raised the average trade to GDP ratio by roughly 5 percentage
points, after controlling for year effects, suggesting that trade policy liberalization
did indeed raise the actual level of openness of liberalizers. Trade-centered
reforms thus have significant effects on economic growth within countries.
These within-country estimates represent the average effect of liberalization
on growth, investment, and openness; they mask differences in the individual
responses of countries to trade liberalization. Restricting the sample to 13
countries sheds light on the sources of these differences. Countries that experienced
positive effects tended to deepen trade reforms. But active industrial
policies, such as those implemented in Southeast Asia, did not preclude growth
212 T H E W O R L D B A N K E CO N O M I C R E V I E W
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
gains from trade liberalization, and broad-based reforms appear to be neither a
necessary nor a sufficient condition for reaping these gains. Countries that
experienced negative or no effects on growth tended to have suffered from
political instability, adopted contractionary macroeconomic policies in the
aftermath of reforms, or undertaken efforts to counteract trade reform by
shielding domestic sectors from necessary adjustments. Future research should
seek to clarify the factors accounting for heterogeneity in the growth effects of
trade reform.
A P P E N D I X
TA B L E A-1. Trade Policy Variables for Economies in Sample, 1990s
Economy
OPEN90–99
(1 ¼ open)a
Average
tariff, 1990–
99 (percent)b
Core nontariff
barrier coverage
rate (percent),
1995–98c
Average BMP,
1990–99
(percent)d
Export
Marketing Board
(1 ¼ country has
exporting board)e
Socialist
(1 ¼ country
is socialist)e
Albania 1 15.90 — 7.53 0 0
Algeria 0 23.97 — 177.91 0 0
Angola 0 — — 23.62 0 0
Argentina 1 12.54 2.1 9.30 0 0
Armenia 1 — — 0 0 0
Australia 1 7.91 — 0 0 0
Austria 1 6.91 — 0 0 0
Azerbaijan 1 — — 0 0 0
Bangladesh 0 43.70 — 83.27 0 0
Barbados 1 15.58 — 2.31 0 0
Belarus 0 12.63 — — 1 0
Belgium 1 6.91 — 0 0 0
Benin 1 28.61 1.0 1.93 0 0
Bolivia 1 10.34 — 1.49 0 0
Botswana 1 20.55 — 7.82 0 0
Brazil 1 17.32 21.6 13.76 0 0
Bulgaria 1 17.37 — 7.44 0 0
Burkina Faso 1 29.13 — 1.98 0 0
Burundi 0 7.40 — 29.55 0 0
Cameroon 1 18.43 — 1.98 0 0
Canada 1 6.81 — 0 0 0
Cape Verde 1 22.05 — 0 0 0
Central African
Republic
0 12.80 — 1.55 1 0
Chad 0 15.80 — 1.98 1 0
Chile 1 11.33 5.2 9.84 0 0
China 0 31.06 — 35.89 0 1
Colombia 1 14.30 10.3 8.87 0 0
Congo, Dem.
Rep. of
0 25.47 — 34.67 1 0
Congo, Rep. of 0 17.97 — 1.98 1 0
Costa Rica 1 10.60 6.20 5.37 0 0
Coˆte d’Ivoire 1 22.00 30.90 1.98 0 0
Croatia 0 — — 37.76 0 0
Cyprus 1 10.64 21.60 2.16 0 0
(Continued)
Wacziarg and Horn Welch 213
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
TABLE A-1. Continued
Economy
OPEN90–99
(1 ¼ open)a
Average
tariff, 1990–
99 (percent)b
Core nontariff
barrier coverage
rate (percent),
1995–98c
Average BMP,
1990–99
(percent)d
Export
Marketing Board
(1 ¼ country has
exporting board)e
Socialist
(1 ¼ country
is socialist)e
Czech Republic 1 6.08 — 0.22 0 0
Denmark 1 6.91 — 0 0 0
Dominican
Republic
1 16.70 6.20 16.31 0 0
Ecuador 1 11.29 — 9.34 0 0
Egypt, Arab
Rep. of
1 30.23 — 12.45 0 0
El Salvador 1 9.38 5.20 13.59 0 0
Estonia 0 1.12 — 25.09 0 0
Ethiopia 0 22.55 — 111.43 0 0
Finland 1 6.91 — 0 0 0
France 1 6.91 — 0 0 0
Gabon 0 19.87 — 1.98 1 0
Gambia, The 1 13.55 — 4.69 0 0
Georgia 1 — — 0 0 0
Germany 1 6.91 — 0 0 0
Ghana 1 14.93 — 2.96 0 0
Greece 1 6.91 — 1.24 0 0
Guatemala 1 10.27 — 6.03 0 0
Guinea 1 — — 3.99 0 0
Guinea-Bissau 1 — — 0 0 0
Guyana 0 13.70 — 28.23 0 0
Haiti 0 10.00 — 81.12 0 0
Honduras 1 8.90 — 9.21 0 0
Hong Kong
(China)
1 — 2.10 -0.02 0 0
Hungary 1 12.11 — 5.40 0 0
Iceland 1 3.98 — 1.24 0 0
India 0 48.65 93.80 7.45 0 0
Indonesia 1 16.27 31.30 7.10 0 0
Iran, Islamic
Rep.
0 — — 1,199.31 0 0
Iraq 0 — — 138,935.90 0 0
Ireland 1 3.98 — 2.50 0 0
Israel 1 7.80 — 2.09 0 0
Italy 1 6.91 — 0 0 0
Jamaica 1 14.68 — 15.46 0 0
Japan 1 5.98 — -0.35 0 0
Jordan 1 15.83 — 3.37 0 0
Kazakhstan 0 — — 55.34 0 0
Kenya 1 27.47 — 15.94 0 0
Korea, Rep. of 1 11.28 25.00 0.03 0 0
Kyrgyz
Republic
1 — — — 0 0
Latvia 1 5.73 — 7.29 0 0
Lesotho 1 17.40 — 3.49 0 0
Liberia 0 — — 2,306.86 0 0
Lithuania 1 4.33 — 7.45 0 0
Luxembourg 1 6.91 — 0.38 0 0
Macedonia,
FYR
1 — — 18.45 0 0
Madagascar 1 7.13 — 5.93 0 0
Malawi 0 19.80 — 28.83 0 0
Malaysia 1 11.70 19.60 1.35 0 0
(Continued)
214 T H E W O R L D B A N K E CO N O M I C R E V I E W
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
TABLE A-1. Continued
Economy
OPEN90–99
(1 ¼ open)a
Average
tariff, 1990–
99 (percent)b
Core nontariff
barrier coverage
rate (percent),
1995–98c
Average BMP,
1990–99
(percent)d
Export
Marketing Board
(1 ¼ country has
exporting board)e
Socialist
(1 ¼ country
is socialist)e
Mali 1 15.66 — 1.98 0 0
Malta 1 7.23 — 1.20 0 0
Mauritania 1 28.23 — 1.55 0 0
Mauritius 1 27.00 16.70 5.25 0 0
Mexico 1 12.53 13.40 2.24 0 0
Moldova 1 — — 0 0 0
Morocco 1 23.75 13.40 3.54 0 0
Mozambique 1 16.25 — 6.87 0 0
Myanmar 0 5.70 — 2,280.77 0 0
Nepal 0 15.28 — 24.23 0 0
Netherlands 1 6.91 — 0 0 0
New Zealand 1 6.35 — 2.50 0 0
Nicaragua 1 9.90 — 9.98 0 0
Niger 1 18.30 — 1.87 0 0
Nigeria 0 29.74 11.50 151.32 0 0
Norway 1 4.87 — 0 0 0
Pakistan 0 54.73 — 9.74 0 0
Panama 1 10.67 — 0 0 0
Papua New
Guinea
0 16.67 — 16.57 1 0
Paraguay 1 10.91 0.00 11.83 0 0
Peru 1 16.80 — 8.75 0 0
Philippines 1 19.09 — 4.36 0 0
Poland 1 12.46 — 2.42 0 0
Portugal 1 6.91 — 2.04 0 0
Romania 0 13.50 — 104.30 0 0
Russian
Federation
0 11.24 — 50,979.69 1 0
Rwanda 0 38.40 — 50.78 0 0
Senegal 0 13.05 — 1.98 1 0
Sierra Leone 0 30.25 — 61.47 0 0
Singapore 1 0.32 2.10 0.80 0 0
Slovak
Republic
1 7.35 — 5.34 0 0
Slovenia 1 10.60 — 10.06 0 0
Somalia 0 — — 246.55 0 0
South Africa 1 9.05 8.30 3.46 0 0
Spain 1 6.91 — 1.71 0 0
Sri Lanka 1 24.34 22.70 7.84 0 0
Swaziland 1 15.10 — 7.62 0 0
Sweden 1 6.91 — 0.00 0 0
Switzerland 1 1.38 — 0.00 0 0
Syrian Arab
Republic
0 16.00 — 279.97 0 0
Taiwan
(China)
1 9.85 — 0.95 0 0
Tajikistan 1 — — — 0 0
Tanzania 0 25.12 — 22.17 0 0
Thailand 1 29.54 17.50 1.80 0 0
Togo 0 15.25 — 1.98 1 0
Trinidad and
Tobago
1 14.86 — 13.22 0 0
Tunisia 1 28.25 — 3.67 0 0
Turkey 1 15.28 19.80 1.15 0 0
(Continued)
Wacziarg and Horn Welch 215
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
TABLE A-1. Continued
Economy
OPEN90–99
(1 ¼ open)a
Average
tariff, 1990–
99 (percent)b
Core nontariff
barrier coverage
rate (percent),
1995–98c
Average BMP,
1990–99
(percent)d
Export
Marketing Board
(1 ¼ country has
exporting board)e
Socialist
(1 ¼ country
is socialist)e
Turkmenistan 0 — — 42.86 1 0
Uganda 1 14.37 3.10 19.33 0 0
Ukraine 0 9.73 — 9.02 1 0
United
Kingdom
1 6.91 — 0.00 0 0
United States 1 5.96 — 0.00 0 0
Uruguay 1 14.00 0.00 9.88 0 0
Uzbekistan 0 — — Dual
exchange rate
0 0
Venezuela 1 14.31 17.70 4.13 0 0
Yemen, Re. of 1 20.00 — 8.34 0 0
Serbia and
Montenegro
0 — — 106.44 0 0
Zambia 0 18.43 1.00 62.55 0 0
Zimbabwe 0 20.43 — 132.81 0 0
— not available.
a
Based on application of Sachs and Warner (1995) criteria; see Wacziarg and Welch (2003)
for details.
b
Unweighted average tariff, 1990–99, based on data from UNCTAD (2001), World Bank
(2000), and WTO (various years).
c
Coverage rate of core nontariff barriers (quotas, licensing, prohibitions, and administered
pricing) on capital good and intermediates, based on data from Michalopoulos (1999).
d
Figures represent [(parallel exchange rate/official exchange rate) – 1] *100, based on data
from Easterly and Sewadeh (2002).
e
Based on literature reviews; see Wacziarg and Welch (2003) for details.
Source: Author compilation.
TA B L E A-2. Liberalization and Openness Dates for Countries in Sample
Year uninterrupted openness begana
Economy
Period of temporary
liberalization
(where applicable)
Sachs and
Warner (1995)
Wacziarg and
Welch (2003)
Albania 1992 1992
Algeria Closed Closed
Angola Closed Closed
Argentina 1991 1991
Armenia Closed 1995
Australia 1964 1964
Austria 1960 1960
Azerbaijan Closed 1995
Bangladesh Closed 1996
Barbados 1966 1966
Belarus 1994 Closed
(Continued)
216 T H E W O R L D B A N K E CO N O M I C R E V I E W
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
TABLE A-2. Continued
Year uninterrupted openness begana
Economy
Period of temporary
liberalization
(where applicable)
Sachs and
Warner (1995)
Wacziarg and
Welch (2003)
Belgium 1959 1959
Benin 1990 1990
Bolivia 1956–79 1985 1985
Botswana 1979 1979
Brazil 1991 1991
Bulgaria 1991 1991
Burkina Faso Closed 1998
Burundi Closed 1999
Cameroon 1993 1993
Canada 1952 1952
Cape Verde n.a. 1991
Central African Republic Closed Closed
Chad Closed Closed
Chile 1976 1976
China Closed Closed
Colombia 1986 1986
Congo, Dem. Rep. of Closed Closed
Congo, Rep. of Closed Closed
Costa Rica 1952–61 1986 1986
Coˆte d’Ivoire Closed 1994
Croatia 1993 Closed
Cyprus 1960 1960
Czech Republic 1991 1991
Denmark 1959 1959
Dominican Republic Closed 1992
Ecuador 1950–82 1991 1991
Egypt, Arab Rep. Closed 1995
El Salvador 1950–61 1989 1989
Estonia 1992 Closed
Ethiopia Closed 1996
Finland 1960 1960
France 1959 1959
Gabon Closed Closed
Gambia, The 1985 1985
Georgia Closed 1996
Germany 1959 1959
Ghana 1985 1985
Greece 1959 1959
Guatemala 1950–61 1988 1988
Guinea 1986 1986
Guinea-Bissau 1987 1987
Guyana 1988 1988
Haiti Closed Closed
Honduras 1950–61 1991 1991
(Continued)
Wacziarg and Horn Welch 217
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
TABLE A-2. Continued
Year uninterrupted openness begana
Economy
Period of temporary
liberalization
(where applicable)
Sachs and
Warner (1995)
Wacziarg and
Welch (2003)
Hong Kong (, China) Always open Always open
Hungary 1990 1990
Iceland n.a. n.a.
India 1994 Closed
Indonesia 1970 1970
Iran, Islamic Rep. of Closed Closed
Iraq Closed Closed
Ireland 1966 1966
Israel 1985 1985
Italy 1959 1959
Jamaica 1962–73 1989 1989
Japan 1964 1964
Jordan 1965 1965
Kazakhstan Closed Closed
Kenya 1963–67 1993 1993
Korea, Rep. of 1968 1968
Kyrgyz Republic 1994 1994
Latvia 1993 1993
Lesotho n.a. Closed
Liberia n.a. Closed
Lithuania 1993 1993
Luxembourg 1959 1959
Macedonia, FYR 1994 1994
Madagascar Closed 1996
Malawi Closed Closed
Malaysia 1963 1963
Mali 1988 1988
Malta n.a. Closed
Mauritania 1992 1995
Mauritius 1968 1968
Mexico 1986 1986
Moldova 1994 1994
Morocco 1956–64 1984 1984
Mozambique Closed 1995
Myanmar Closed Closed
Nepal 1991 1991
Netherlands 1959 1959
New Zealand 1986 1986
Nicaragua 1950–60 1991 1991
Niger Closed 1994
Nigeria Closed Closed
Norway Always open Always open
Pakistan Closed 2001
Panama n.a. 1996
(Continued)
218 T H E W O R L D B A N K E CO N O M I C R E V I E W
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
TABLE A-2. Continued
Year uninterrupted openness begana
Economy
Period of temporary
liberalization
(where applicable)
Sachs and
Warner (1995)
Wacziarg and
Welch (2003)
Papua New Guinea Closed Closed
Paraguay 1989 1989
Peru 1948–67 1991 1991
Philippines 1988 1988
Poland 1990 1990
Portugal Always open Always open
Romania 1992 1992
Russian Federation Closed Closed
Rwanda Closed Closed
Senegal Closed Closed
Serbia and Montenegro Closed 2001
Sierra Leone Closed 2001
Singapore 1965 1965
Slovak Republic 1991 1991
Slovenia 1991 1991
Somalia Closed Closed
South Africa 1991 1991
Spain 1959 1959
Sri Lanka 1950–56; 1977–3 1991 1991
Swaziland n.a. Closed
Sweden 1960 1960
Switzerland Always open Always open
Syrian Arab Republic 1950–65 Closed Closed
Taiwan (China) 1963 1963
Tajikistan Closed 1996
Tanzania Closed. 1995
Thailand Always open Always open
Togo Closed. Closed
Trinidad and Tobago Closed. 1992
Tunisia 1989 1989
Turkey 1950–53 1989 1989
Turkmenistan Closed Closed
Uganda 1988 1988
Ukraine Closed Closed
United Kingdom Always open Always open
United States Always open Always open
Uruguay 1990 1990
Uzbekistan Closed Closed
Venezuela, R.B. de 1950–59; 1989–93 Closed 1996
Yemen, Rep. Always open Always open
Zambia 1993 1993
Zimbabwe Closed Closed
Note: n.a. means not classified. Closed denotes countries closed as of 1994 in the
Sachs-Warner column and closed as of 2000 in the Wacziarg-Welch column.
a
Based on the Sachs and Warner (1995) criteria and broader literature review. See Wacziarg
and Welch (2003) for details.
Wacziarg and Horn Welch 219
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
TA B L E A-3 Trade Liberalization and Concurrent Events in Subsample of 13
Countries
Country (year of
liberalization)
Sample
period Concurrent events
Countries that experienced negative or zero growth after liberalization
Botswana (1979) 1961–98 Since gaining independence in 1966, Botswana has had
one of the fastest growth rates in the world (IMF 2002),
growing at an annual rate of 7.7 percent between 1965
and 1998 (Rodrik 2003). Income per capita (in
purchasing parity power–adjusted terms in 1998) was
four times the African average.
Botswana experienced a mean growth difference of 21.99
percent in the years before and after liberalization. This
differs that despite trade liberalization and an
export-oriented economy, government intervention has
been high in Botswana, where the public sector
accounts for a large share of the economy.
Botswana’s economy expanded when diamond mining
began in 1971. The recession of 1981/82 was partly a
result of a weak world diamond market. The late 1980s
were a period of new mining activity and strong
demand that supported overall economic growth.
During the early to mid-1990s, recessionary conditions
on the world diamond market led to a severe economic
slump in Botswana. Because diamond exports account
for 70 percent of Botswana’s export earnings and more
than a third of its GDP, volatile diamond prices have
had a significant impact on the country’s overall
economic growth. Despite volatility, growth remained
positive throughout most of the sample period, however
(EIU various years; IMF 2002).
Colombia (1986) 1951–98 In December 1990, Colombia was unable to repay its debt
principal payments; it was unable to refinance its debt
until April 1991. In the wake of this crisis, Colombia
pursued a variety of market-oriented reforms in
addition to further trade liberalization. Price controls
were lifted, a financial sector reform was implemented,
the exchange control system was liberalized, the
regulatory framework was modernized, and some
industries were privatized (Wacziarg and Wallack
2004).
Throughout the 1990s, substantial trade reforms were
implemented, including bilateral trade agreements with
other Latin American countries in 1993/94 (Henry
1999). The rise in civil unrest beginning in 1992 and
the political instability that persisted throughout the
1990s likely limiting postliberalization economic
growth.
(Continued)
220 T H E W O R L D B A N K E CO N O M I C R E V I E W
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
TABLE A-3 Continued
Country (year of
liberalization)
Sample
period Concurrent events
Hungary (1990) 1971–98 Hungary experienced a period of declining growth and poor
economic conditions between 1971 and 1991. In 1988–
89, the leader of the majority party changed after 30 years,
and a period of political uncertainty ensued. In 1989, the
new government implemented a stabilization program that
included higher taxes, tighter monetary policy, and the
devaluation of the currency (World Bank 1995).
In 1990/91, Hungary implemented an IMF restructuring
program. In 1995, it implemented structural reforms,
including currency devaluation, a new exchange rate
mechanism, a tight wage policy in the public sector, and
fiscal measures to enhance revenues and cut expenditures.
Hungary accelerated privatization efforts, restructuring
enterprises (including major commercial banks) and
implementing financial sector and public finance reforms
in the mid-1990s. Significant improvements were also
made in the legal and regulatory framework of the
financial sector (Wacziarg and Wallack 2004).
Economic recovery began in 1992/93. During the
mid-1990s, Hungary adhered to the IMF plan and
experienced gradual stabilization and recovery. Growth
did not return to the levels seen before liberalization,
however. Persisting high levels of debt and current
account deficits may have limited the gains from trade
liberalization. In addition, in 1993 the government
tightened monetary policy and increased interest rates,
which likely dampened the economic recovery (World
Bank 1995). While structural reforms were implemented
in 1995, the full effects may not have been evident
before the end of the study period (1998).
Israel (1985) 1951–98 The wars of 1967 and 1973 limited economic growth. In
1977, both tariff and currency barriers were relaxed; in
1979, the government approved a five-year plan to
reduce inflation and customs rates. In January 1980,
tariffs were further reduced on imports from the
European Economic Community.
Israel invaded Lebanon in June 1982; roughly a year later it
entered a deep economic crisis, characterized by
triple-digit inflation, a widening trade gap, rapidly
mounting foreign debt, and significant real exchange rate
appreciation. In July 1985, the government implemented
an emergency economic stabilization plan in order to
stop hyperinflation; it also signed a free trade agreement
with the United States (Henry 1999). Inflation dropped
significantly in late 1985 and 1986, and the IMF
announced its support of Israeli reform efforts. In 1986,
Israel fixed the exchange rate to a trade-weighted
currency basket. In January 1987, it devalued the
currency 19 percent and implemented other changes
affecting the tax system and money markets.
(Continued)
Wacziarg and Horn Welch 221
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
TABLE A-3 Continued
Country (year of
liberalization)
Sample
period Concurrent events
Despite devaluations of the currency in 1988 and 1989,
the interest rate increased, because of currency
volatility. In 1991, Israel implemented a crawling band
exchange rate system. The shekel was devalued by 6
percent in order to boost the economy, which was
suffering as a result of the Gulf War. In November
1995, a free trade area treaty affirming Israel’s special
trade status with the European Union was signed
(Henry 1999).
Despite trade reforms implemented throughout the period,
Israel’s heterodox stabilization program may have offset
the effects of trade liberalization. Social pacts based on
broad coalitions of labor, government, and industry set
the patterns for prices, wages, and the exchange rate
(Wacziarg and Wallack 2004). In addition, inflation,
currency volatility, and high interest rates in the late
1980s and early 1990s reduced Israel’s competitiveness
and the gains from trade.
Mexico (1986) 1951–98 The 1940s–1960s was a period of political and social
stability and relatively rapid economic growth in
Mexico (Tornell 2002). In the early 1970s,
expansionary fiscal and monetary policy led to
increasing levels of debt, escalating prices, and an
overvaluation of the exchange rate. By 1976, inflation
was increasing and private investment decreasing. In
August 1976, the government was forced to devalue
the peso and decrease government expenditure
(Gonzalez 1994).
The discovery of oil in 1977 stimulated the economy
between 1978 and 1982: in 1981, oil accounted for
three-fourths of Mexico’s exports. Government
spending, financed by international borrowing,
increased, however, resulting in the overvaluation of
the peso. By mid-1981, the international price of oil
had fallen; by 1982, Mexico declared itself unable
to service its debt. The government devalued the
peso by 30 percent in February 1982 and implemented
a two-tiered foreign exchange system in August 1982.
Mexico experienced a severe recession during the
Latin American debt crisis of 1982–83 (Gonzalez
1994).
(Continued)
222 T H E W O R L D B A N K E CO N O M I C R E V I E W
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
TABLE A-3 Continued
Country (year of
liberalization)
Sample
period Concurrent events
In 1984, Mexico pursued a policy of privatization and
liberalization in order to attract FDI (Henry 1999). In
1985, it implemented a program of stabilization and
structural adjustment, including trade liberalization. It
joined GATT in 1986 and significantly reduced import
restrictions and tariff barriers. A debt-rescheduling
agreement was signed in August 1985. In July 1986, an
IMF agreement was implemented, facilitating additional
debt restructuring. Further trade liberalization measures
were implemented in August 1987 (Henry 1999).
Mexico also pursued a privatization program during the
1980s, which continued into the 1990s, with the
privatization of the banking industry (Henry 1999;
Wacziarg and Wallack 2004).
An economic and fiscal crisis occurred again in 1994–95.
It was accompanied by a period of political unrest,
including the Chiapas uprising and the assassination of
several PRI figures (Henry 1999). In December 1994,
Mexico devalued the peso and implemented a floating
exchange rate regime. In 1995, it received a bailout,
which prevented it from defaulting on its debt and
granted it continued access to international capital
markets (Tornell 2002).
Despite the economic recovery and trade liberalization that
occurred in the late 1980s, Mexico never recovered to its
precrisis levels of growth. The persisting macroeconomic
instability and lack of additional structural reforms
appear to have been key factors in limiting the gains from
trade liberalization by preventing economic restructuring
and reallocation of resources. According to the IMF
(1999), further banking sector reforms and continued
economic restructuring were necessary to sustain
economic growth. The macroeconomic environment was
hindered by the volatile price of oil, uncertainty regarding
debt negotiations, and speculative attacks on the peso. As
the government decreased expenditure under the
structural adjustment program, domestic demand fell.
During the late 1980s and early 1990s, the currency
became overvalued again, effectively offsetting trade
liberalization measures. Nontrade barriers to
competition also existed, in the form of government
monopolies and oligopolies, which limited
restructuring. The high interest rate, aimed at preventing
speculative attacks and attracting foreign capital,
limited domestic demand and restructuring (Gonzalez
1994).The economy improved between 1995 and 1998
as a result of the implementation of structural reforms
and the success of the floating exchange rate in
mitigating the effect of external shocks (IMF 1999).
(Continued)
Wacziarg and Horn Welch 223
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
TABLE A-3 Continued
Country (year of
liberalization)
Sample
period Concurrent events
Philippines (1988) 1951–98 During the 1960s the Marcos regime increased trade
barriers, which remained in effect until the 1980s.
During the 1983–86 economic crisis, the inflation rate
increased significantly. The currency was devalued by
50 percent in 1984, and expansionary monetary policy
limited capital inflow and economic growth. The
Philippines secured debt rescheduling agreements
between 1985 and 1988, and the IMF approved a
stabilization plan in 1989.
In 1986 (the end of the Marcos era) the Philippines
implemented trade liberalization measures, including
the lifting of import restrictions. Despite these reforms
government investment in state enterprises roughly
doubled during the sample period, as did state
enterprises’ percentage of total economic activity,
employment, and net financial flows (Wacziarg and
Wallack 2004)
The Philippines implemented capital market liberalization,
including reform of the foreign exchange rate, in 1992.
The IMF approved the country’s economic performance
and rescheduled additional debt. Further trade reforms,
including the removal of quantitative restrictions, were
also implemented during the early 1990s (Henry 1999).
Despite further trade liberalization measures, the
Philippines has not witnessed the increased economic
growth experienced in other countries following
liberalization, possibly because of limited structural
reforms and the high level of government involvement
in state enterprises. Pritchett (2003) cites the
institutional uncertainty that arose from political
instability in the Philippines following liberalization as a
factor that may have limited investment and economic
growth.
Countries that experienced positive growth increased after liberalization
Chile (1976) 1952–98 When Salvatore Allende assumed power in 1970, he
nationalized Chile’s copper mines, banks, and other
enterprises. Government expenditure increased
dramatically: the country’s budget deficit rose from 2.7
percent of GDP to 25.0 percent between 1970 and
1973. The black market currency premium exceeded
600 percent in 1972; inflation exceeded 100 percent in
1973 (Easterly and Sewadeh 2002; Stallings and Brock
1993).
(Continued)
224 T H E W O R L D B A N K E CO N O M I C R E V I E W
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
TABLE A-3 Continued
Country (year of
liberalization)
Sample
period Concurrent events
In 1973, Augusto Pinochet overthrew President Allende in
a military coup. Between 1975 and 1982, structural
changes to liberalize the financial system were
implemented. Quantitative restrictions were eliminated
in 1973; tariffs were significantly reduced between 1973
and 1979, when they were set at a uniform rate of 10
percent. In 1979, the exchange rate was fixed to the
U.S. dollar, capital controls reduced, the tax system
simplified, and privatization pursued (Stallings and
Brock 1993).
Trade reform in the early years of the pro-market Pinochet
administration was accompanied by privatization,
elimination of the fiscal deficit, and the lifting of price
and interest rate controls. Liberalization of the labor
market also facilitated overall economic restructuring
(Wacziarg and Wallack 2004).
In 1980/81, Chile privatized its social security system and
implemented banking reforms. It experienced an
economic crisis during the Latin American debt crisis,
during which it was unable to access credit markets and
the government assumed control of troubled banks. In
1982, GDP fell 14 percent and inflation doubled.
Between 1982 and 1985, the peso was devalued, tariff
rates were raised to 35 percent, and privatization efforts
were reversed (Stallings and Brock 1993). In 1985, the
peso was gradually depreciated with a crawling peg,
tariffs were reduced to 15 percent, and privatization
resumed. During the mid- to late-1980s, Chile decreased
tariffs, rescheduled its debt, and reprivatized the
banking sector. During the 1990s, it signed free trade
agreements with Colombia and Mexico and engaged in
substantial capital market liberalization (Henry 1999).
Ghana (1985) 1956–98 Upon gaining independence in 1957, Ghana pursued a
strategy of import substitution. It implemented a series
of restrictive trade policies, including tariffs, nontariff
barriers, and exchange rate controls. It also established
a variety of state enterprises. By 1966, Ghana’s currency
was overvalued and a cycle of political instability
(including military coups) began. Rising inflation
followed by currency devaluations ensued during the
late 1960s and 1970s (Leith and Lofchie 1993).
(Continued)
Wacziarg and Horn Welch 225
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
TABLE A-3 Continued
Country (year of
liberalization)
Sample
period Concurrent events
Ghana experienced another economic crisis in 1982,
during which inflation increased and foreign exchange
reserves dropped to very low levels. In 1983, the
government launched a four-year economic recovery
program that included restructuring the country’s
physical infrastructure and economic institutions and
reducing inflation through prudent monetary, fiscal, and
trade policies. The 1985 trade liberalization program
was part of the Rawlings administration’s World Bankand
IMF-supported economic recovery program.
Multiple exchange rates implemented to promote
exports were later replaced with unified rates and
subjected to a series of devaluations. Public sector
employment (including in state enterprises) was cut and
distortions in wages reduced (Wacziarg and Wallack
2004). Ghana continued to implement trade and capital
market reforms through the late 1980s and 1990s.
Indonesia (1970) 1961–98 Indonesia suffered an economic crisis during the early
1960s, during which budget deficits rose and annual
inflation reached 640 percent. Under pressure from the
army, in March 1966 President Sukarno transferred
some power to Suharto; in March 1967 Suharto was
named president. A five-year development plan to
stabilize the economy and promote growth was
implemented that successfully stabilized the economy.
Capital market liberalization occurred in 1970. In
February 1976, the government reduced the 10 percent
export duty on a wide range of commodities. During
the 1970s, government intervention increased despite
the implementation of trade liberalization reforms. The
government increased its control of state-owned banks
and other enterprises. Oil revenue was significant during
the 1970s; economic growth weakened in the early
1980s, partly as a result of falling oil prices. However,
the impact was mitigated by the country’s rapid
adjustment and a relatively low debt burden (Temple
2003).
In June 1983, the government announced a series of bank
liberalization reforms, followed by further reforms in
1988 when credit subsidies were removed (Temple
2003). Devaluations occurred in 1983 and 1986.
(Continued)
226 T H E W O R L D B A N K E CO N O M I C R E V I E W
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
TABLE A-3 Continued
Country (year of
liberalization)
Sample
period Concurrent events
During 1984/85, Indonesia entered into bilateral trading
agreements with the Soviet Union, the United States,
and several other countries. In May 1986, the
government announced new measures to attract foreign
investment. The oil market crashed in the second
quarter of 1986. During 1986, further trade and
investment liberalization measures were implemented,
with the gradual removal of qualitative restrictions and
nontariff barriers.
The government implemented a large-scale privatization
program during the late 1980s and early 1990s.
Between 1991 and 1995, it implemented banking
reforms to strengthen the system; to stimulate lending, it
later weakened these regulations. During 1991–95,
capital market reforms aimed at improving stock
exchange were implemented (EIU various years; Henry
1999).
Korea, Rep. of
(1968)
1954–98 Political turmoil in Korea in the late 1950s forced
President Syngman Rhee’s resignation in 1960. A
military coup followed in 1961, along with continued
political unrest. Inflation increased and foreign
exchange reserves decreased significantly before Korea
stabilized and started its slow transition to democratic
rule in 1964 (Haggard, Cooper, and Moon 1993).
Korea transitioned from a policy of import substitution to
export-oriented growth during the mid-1960s. Tariffs
and nontariff barriers were reduced, and the
government created export-processing zones and
adopted other mechanisms for increasing FDI (Sakurai
1995). The currency was devalued, the tax system and
interest rates reformed, and capital markets liberalized.
In 1965, the Export Development Committee was
established; in 1966, quantitative restrictions were
eliminated. Liberalization was not universal, however;
certain sectors remained protected, and government
involvement in the economy remained pervasive.
The assassination of President Chung-Hee Park in
November 1979 unleashed a year of political and
economic crisis. In 1980, significant banking reforms were
announced; in 1981, a five-year economic plan of
structural adjustment was initiated. Economic growth was
dampened during the Asia financial crisis of 1982–84.
(Continued)
Wacziarg and Horn Welch 227
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
TABLE A-3 Continued
Country (year of
liberalization)
Sample
period Concurrent events
Capital and banking sector reforms were implemented in
1984. Further trade reforms, including reductions in
tariffs and nontariff barriers, were implemented in the
mid- to late-1980s. Banking and capital market reforms
were deepened in 1991 in an effort to attract FDI. In
1993, a five-year plan for reform and further financial
system regulation was adopted (Henry 1999).
Poland 1971–98 Poland’s economy collapsed during the 1970s. In August
1980, the Solidarity movement began, and a period of
political unrest ensued. Martial law remained in effect
through December 1982. In 1986, Poland was accepted
into the IMF and began to pursue debt restructuring. In
1989, hyperinflation impeded economic growth, causing
Poland’s debt to reach 74 percent of GDP (de Menil
2003).
In 1990, the government implemented a swift and
comprehensive set of market reforms, including trade
liberalization, in order to stabilize the economy. The
Balcerowicz Plan included removal of price controls,
reduction of government expenditure and investment,
devaluation of the exchange rate, and removal of
subsidies for energy (Wacziarg and Wallack 2004).
Trade liberalization measures included the liberalization
and elimination of exchange controls and the abolition
of state trading monopolies and nearly all quotas and
tariffs. The currency was devalued by more than 50
percent in January 1990 and then gradually depreciated
based on a crawling peg until 1995 (de Menil 2003).
Despite a deep recession in 1991, Poland persisted in its
liberalization program, implementing a new IMF plan
in 1993 that included tax reform and continued
privatization. During the mid-1990s, Poland continued
to implement reforms, including currency reform,
privatizations, and policies, to promote FDI. It applied
for EU membership in 1994 and became a member of
the Organisation for Economic Co-operation and
Development in 1996.
According to de Menil (2003), productivity gains appear
to have been the primary factors in Poland’s growth
during the 1990s. He believes that comprehensive
structural reforms facilitated economic transformation,
the reallocation of resources, and the rapid adoption of
Western principles of management and standards of
efficiency.
(Continued)
228 T H E W O R L D B A N K E CO N O M I C R E V I E W
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
TABLE A-3 Continued
Country (year of
liberalization)
Sample
period Concurrent events
Taiwan (China)
(1963)
1952–98 Most of the economies in this subsample implemented
trade liberalization in the wake of economic and often
political crises. In contrast, Taiwan had a stable
economic environment and relatively low tariff rates at
the time of liberalization. Trade liberalization in the
early 1960s involved further tariff reductions as well as
incentives, such as the creation of export-promotion
zones, to attract FDI (Sakurai 1995).
Between 1985 and 1987, Taiwan further reduced tariffs
and nontariff barriers. In 1985, it implemented polices
to promote FDI and liberalize the foreign exchange
market. In 1987, it tightened capital controls. In 1988,
it implemented capital market reform measures along
with additional trade reform measures. In 1989–92, it
implemented banking reforms and privatization
measures (Henry 1999).
Uganda (1988) 1951–98 Between independence (in 1962) and 1980, Uganda
experienced economic devastation, as a result of
mismanagement and war. Capital was destroyed, and
manufacturing operated at extremely low capacity. In
1981, Uganda implemented an IMF reform program
that included floating the currency, removing price
controls, and imposing fiscal austerity. The reform
program was initially successful, but success was not
sustained and the IMF withdrew its support in 1984, a
year that marked the beginning of a period of economic
collapse and civil war.
In 1985, Uganda implemented policies to promote FDI
and liberalize the foreign exchange market. A new
economic recovery program was launched in 1987.
Political unrest led to a tightening of the capital market,
however. In 1988, further trade and capital market
liberalization measures resumed, followed by banking
reforms and privatization between 1989 and 1992. In
1993–94, further trade and capital market
liberalization measures were implemented, including the
liberalization of the interest rate. A variety of currency
regimes was implemented between 1988 and 1992; the
currency was pegged to the U.S. dollar and a composite
of other currencies before a flexible exchange rate
system was adopted in 1996 (Amvouna 1998).
Source: Author compilation.
Wacziarg and Horn Welch 229
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
REFERENCES
Ades, Alberto F., and Edward L. Glaeser. 1999. “Evidence on Growth, Increasing Returns and the
Extent of the Market.” Quarterly Journal of Economics 114(3):1025–45.
Alesina, Alberto, Enrico Spolaore, and Romain Wacziarg. 2000. “Economic Integration and Political
Disintegration.” American Economic Review 90(5):1276–96.
Amvouna, Anatolie Marie. 1998. “Determinants of Trade and Growth Performance in Africa: A
Cross-Country Analysis of Fixed versus Floating Exchange Rate Regimes.” African Economic Policy
Discussion Paper 16. USAID Equity and Growth through Economic Research (EAGER) Project. U.S.
Agency for International Development, Washington, D.C.
Baldwin, Richard, and Elena Seghezza. 1996. Testing for Trade-Induced, Investment-Led Growth.
NBER Working Paper 5416. Cambridge, Mass.: National Bureau of Economic Research.
Bekaert, Geert, and Campbell R Harvey. 2000. “Foreign Speculators and Emerging Equity Markets.”
Journal of Finance 55(2):565–614.
Bekaert, Geert, Campbell R. Harvey, and Christian Lundblad. 2001. Does Financial Liberalization Spur
Growth? NBER Working Paper 8245. Cambridge, Mass.: National Bureau of Economic Research.
Ben-David, Dan. 1993. “Equalizing Exchange: Trade Liberalization and Income Convergence.”
Quarterly Journal of Economics 108(3):653–79.
de Menil, Georges. 2003. “History, Policy, and Performance in Two Transition Economies: Poland and
Romania.” In Dani Rodrik, ed., In Search of Prosperity: Analytic Narratives on Economic Growth.
Princeton, N.J.: Princeton University Press.
Dollar, David. 1992. “Outward-Oriented Developing Economies Really Do Grow More Rapidly:
Evidence from 95 LDCs 1976–1985.” Economic Development and Cultural Change 40(3):523–44.
Easterly, William, and Mirvat Sewadeh. 2002. Global Development Network Growth Database. World
Bank, Washington, D.C.
EBRD (European Bank for Reconstruction and Development). 1994. Transition Report: Economic
Transition in Eastern Europe and the Former Soviet Union. London: European Bank for
Reconstruction and Development.
Edwards, Sebastian. 1992. “Trade Orientation, Distortions and Growth in Developing Countries.”
Journal of Development Economics 39(1):31–57.
EIU (Economist Intelligence Unit). Various years. Country Profiles and Quarterly Economic Reports.
London.
Frankel, Jeffrey A., and David Romer. 1999. “Does Trade Cause Growth?” American Economic
Review 89(3):379–99.
Gonza´lez, Diana Alarco´n. 1994. Changes in the Distribution of Income in Me´xico and Trade
Liberalization. Tijuana, Me´xico: El Colegio de la Frontera Norte Press.
Haggard, Stephen, Richard N. Cooper, and Chung-in Moon. 1993. “Policy Reform in Korea in
Political and Economic Interactions.” In Robert H. Bates, and Anne O. Krueger, eds., Economic
Policy Reform: Evidence from Eight Countries. Cambridge, Mass.: Blackwell.
Harrison, Ann, and Gordon Hanson. 1999. “Who Gains from Trade Reform? Some Remaining
Puzzles.” Journal of Development Economics 59(1):125–54.
Henry, Peter Blair. 1999. “Chronological Listing of Major Policy Events in Emerging Markets.”
Working paper. Stanford, Calif.: Stanford University, Graduate School of Business. (https://
faculty-gsb.stanford.edu/henry/Homepage/PDF/Reform%20List.pdf)
———. 2000. “Stock Market Liberalization, Economic Reform, and Emerging Market Equity Prices.”
Journal of Finance 55(2):529–64.
Heston, Alan, Robert Summers, and Bettina Aten. 2002. Penn World Table Version 6.0. Center for
International Comparisons (CICUP), University of Pennsylvania, Philadelphia.
Husain, Ishrat, and Rashid Faruqee. 1994. Adjustment in Africa: Lessons from Country Case Studies.
Washington, D.C.: World Bank.
230 T H E W O R L D B A N K E CO N O M I C R E V I E W
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017
IMF (International Monetary Fund). 1999. Press Release 99/28. 7 July.
———. 2002. “IMF Concludes 2002 Article IV Consultation with Botswana.” Public Information
Notice, 1 November. Washington, D.C.
Kornai, Janos. 2002. The Socialist System: The Political Economy of Communism. Oxford: Clarendon
Press.
Leith, J. Clark, and Michael F. Lofchie. 1993. “The Political Economy of Structural Adjustment in
Ghana.” In Robert H. Bates, and Anne O. Krueger, eds., Political and Economic Interactions in
Economic Policy Reform: Evidence from Eight Countries. Cambridge, Mass.: Blackwell.
Levine, Ross, and David Renelt. 1992. “A Sensitivity Analysis of Cross-Country Growth Regressions.”
American Economic Review 84(4):942–63.
Michalopoulos, Constantine. 1999. “Trade Policy and Market Access Issues for Developing Countries.”
Policy Research Working Paper 2214. World Bank, Washington, D.C.
Pritchett, Lant. 2003. “A Toy Collection, a Socialist Star, and a Democratic Dud: Growth Theory,
Vietnam, and the Philippines.” In Dani Rodrik, ed., In Search of Prosperity: Analytic Narratives on
Economic Growth. Princeton, N.J.: Princeton University Press.
Rodrı´guez, Francisco, and Dani Rodrik. 2000. “Trade Policy and Economic Growth: A Skeptic’s Guide
to the Cross-National Evidence.” In Ben Bernanke, and Kenneth Rogoff, eds., NBER
Macroeconomics Annual 2000. Cambridge, Mass.: MIT Press.
Rodrik, Dani. 2003. In Search of Prosperity: Analytic Narratives on Economic Growth. Princeton, N.J.:
Princeton University Press.
Sachs, Jeffrey D., and Andrew Warner. 1995. “Economic Reform and the Process of Global
Integration.” Brookings Papers on Economic Activity 1:1–118.
Sakurai, Makoto. 1995. “East Asian Agricultural Trade and Protectionism.” In Toshihiko Kawagoe,
and Sueo Sekiguchi, eds., East Asian Economies: Transformation and Challenges. ASEAN Economic
Research Unit, Institute of Southeast Asian Studies, Singapore.
Stallings, Barbara, and Philip Brock. 1993. “The Political Economy of Economic Adjustment: Chile.” In
Robert H. Bates, and Anne O. Krueger, eds., Political and Economic Interactions in Economic
Policy Reform: Evidence from Eight Countries. Cambridge, Mass.: Blackwell.
Temple, Jonathan. 2003. “Growing into Trouble: Indonesia after 1966.” In Dani Rodrik, ed., In Search
of Prosperity: Analytic Narratives on Economic Growth. Princeton, N.J.: Princeton University Press.
Tornell, Aaron. 1998. Reform from Within. NBER Working Paper 6497. Cambridge, Mass.: National
Bureau of Economic Research.
———. 2002. “Economic Crises and Reform in Mexico.” In Stephen Haber, ed., Crony Capitalism and
Economic Growth in Latin America: Theory and Evidence. Stanford, Calif.: Hoover Institution Press.
UNCTAD (United Nations Conference on Trade and Development). 2001. Trade Analysis and
Information System (Trains). CD-ROM. Version 8.0. Geneva: United Nations Conference on Trade
and Development.
Wacziarg, Romain. 2001. “Measuring the Dynamic Gains from Trade.” World Bank Economic Review
15(3):393–429.
———. 2003. “India in the World Trading System.” Working paper. Stanford, Calif: Graduate School
of Business, Stanford University.
Wacziarg, Romain, and Jessica Seddon Wallack. 2004. “Trade Liberalization and Intersectoral Labor
Movements.” Journal of International Economics 64(2):411–39.
Wacziarg, Romain, and Karen Horn Welch. 2003. Trade Liberalization and Growth: New Evidence.
NBER Working Paper 10152. Cambridge, Mass.: National Bureau of Economic Research.
World Bank. 1995. Hungary: Structure Reforms for Sustainable Growth. World Bank Country Study.
Washington, D.C.
———. 2000. International Trade and Development: Data on Import Tariff and NTBs. Washington,
D.C. (www1.worldbank.org/wbiep/trade/TR_data.html).
WTO (World Trade Organization). 1996–2002. Trade Policy Reviews. Various countries. Geneva.
Wacziarg and Horn Welch 231
Downloaded from https://academic.oup.com/wber/article-abstract/22/2/187/1736899
by University Library UUtrecht user
on 31 December 2017