Abstract
We show that there is no positive relationship between trade liberalization and economic growth during the Great Liberalization of the 1990s. We replicate Estevadeordal and Taylor (Rev Econom Statist 95:1669, 2013) and find that their estimated positive relationship between trade liberalization and economic growth is not robust. Their results are driven by outliers, internally inconsistent tariff data, inappropriate selection of the treatment year, and their specific data on economic growth.
This is a preview of subscription content, log in via an institution to check access.
Source: Author’s calculations based on data from Estevadeordal and Taylor (2013)
Source: Author’s calculations based on data from Estevadeordal and Taylor (2013)
Similar content being viewed by others
Data and computer code availability
The datasets of this paper (1. code and programs, 2. data, 3. detailed readme files) are collected in the electronic supplementary material of this article.
Notes
According to the latest empirical literature review on the topic by Irwin (2019), the replicated study is the only one to directly use tariff rates to test the effects of trade policy on economic growth, which in turn highlights its substantial importance.
The database can be consulted online at https://www.fraserinstitute.org/studies/economic-freedom/.
The data on disaggregated tariffs were collected by the authors.
The authors pursue IV regressions to respond to the possible counterargument that institutions might be the underlying force driving policy changes. However, they present regression results for changes in variables from period 1 to period 2, where institutional variables are not significant. They ultimately pursue the IV regressions since these have the “benefit of addressing problems of measurement error" (p. 1683). If the measurement problems are related, as they argue, to the omission of nontariff measures, it is not clear how the IV regressions help to address them. Nevertheless, we also carried out the battery of robustness checks proposed in the main section of this paper to the IV framework. The results presented in Table 8 in Appendix reveal that the significance of the positive relationship between tariff rates and economic growth only holds for some specifications, thus confirming the main message of this paper. It is actually important to note that the IV regression with the first instrument, the one with the highest country coverage, delivers insignificant results as soon as we shift growth data and use PPP from Penn World Table (PWT) 9.1 or constant national prices from World Development Indicators (WDI).
The authors also use this difference regression framework controlling for other relevant variables of the growth literature, Table 5 in their paper. Table 9 in “Appendix” shows the results of the robustness exercises proposed in this paper controlling for these additional variables used in the original paper. The results are qualitatively the same as those presented in the main text below.
The same situation applies for Trinidad and Tobago (TTO). However, tariff data for period 1 are not inconsistent in the sense that the full average is a possible weighted average of disaggregated figures. However, the data are inconsistent in the sense that all disaggregated measures decrease from period 1 to period 2, while the full average increases. Although addressing this case would strengthen our results, we ignore it as the inconsistency is different from the one we address.
If we also extend the assumption made here to address the inconsistencies in the disaggregated tariff data for other countries with consistency problems (i.e., Ghana, New Zealand, and Turkey), the results in terms of significance and magnitude are practically the same as those presented in Table 4. These results are available upon request but not presented here.
We tried to find GDP per capita in PPP dollars in the newest WDI database, consulted in 2020. Unfortunately, this variable is only available from 1990 onward, so we were unable to find the growth data used by the authors.
The specific data used, as there are many in PWT 9.1, are the real output-side GDP at chained PPP terms.
We also ran regressions with all robustness checks included at the same time. The results, presented in Table 12, confirm the lack of robustness of the results in Estevadeordal and Taylor (2013). Moreover, there is one case in fact where the sign shifts entirely and becomes positive, and the result is significant.
All the robustness exercises here presented—excluding some of those when all the robustness checks are ran at the same time–deliver point estimates that belong to the confidence interval of the regression by Estevadeordal and Taylor (2013). Therefore, what we are really showing here is that a robust calculation of the confidence interval cannot exclude zero, so that the impact of trade policy on growth may be not different to zero.
As mentioned in Sect. 2, this implies identifying more carefully the countries that liberalized, the year of liberalization for each of them, and identifying good control groups, so that countries pertaining to the control group are those never treated (or treated much later), and countries treated a long time ago (so that they are not affected by the treatment at the moment of analysis).
References
Barro RJ, Lee JW (2013) A new data set of educational attainment in the world, 1950–2010. J Develop Econ 104:184–198
Belsley DA, Kuh E, Welsch RE (1980) Regression diagnostics: Identifying influential data and sources of collinearity. John Wiley & Sons, Hoboken
Chui M, Levine P, Mansoob Murshed S, Pearlman J (2002) North-South models of growth and trade. J Econ Surv 16(2):123–165
DeJong DN, Ripoll M (2006) Tariffs and growth: an empirical exploration of contingent relationships. Rev Econ Stat 88(4):625–640
Dollar D (1992) Outward-oriented developing economies really do grow more rapidly: evidence from 95 LDCs, 1976–1985. Econ Develop Cultural Change 40(3):523–544
Estevadeordal A, Taylor AM (2013) Is the Washington Consensus Dead? Growth, Openness, and the Great Liberalization, 1970s–2000s. Rev Econ Stat 95(5):1669–1690
Feenstra RC, Inklaar R, Timmer MP (2015) The next generation of the Penn World Table. Am Econ Rev 105(10):3150–82
Frankel JA, Romer DH (1999) Does Trade Cause Growth? Am Econ Rev 89(3):379–399
Irwin DA (2019) Does Trade Reform Promote Economic Growth? A Review of Recent Evidence. Working Paper 25297, National Bureau of Economic Research
Simon Johnson P, Larson W, Papageorgiou C, Subramanian A (2013) Is newer better? penn world table revisions and their impact on growth estimates. J Monet Econ 60(2):255–274
Lora E (2012) Structural reform in Latin America: What has been reformed and how it can be quantified (updated Version). Working Paper 346, Inter-American Development Bank
Matsuyama K (1992) Agricultural productivity, comparative advantage and economic growth. J Econ Theory 58:317–334
Mazumdar J (1996) Do Static Gains from Trade Lead to Medium-Run Growth? J Political Econ 104(6):1328–1337
Rodríguez F, Rodrik D (2000) Trade policy and economic growth: a Skeptic’s guide to the cross-national evidence. NBER Macroecon Ann 15:261–325
Sachs JD, Warner A (1995) Economic Reform and the Process of Global Integration. Brook Papers Econ Activity 1:1–118
Skott P, Larudee M (1998) Uneven development and the liberalisation of trade and capital flows: the case of Mexico. Cambridge J Econ 22(3):277–295
WTO (n.d.) The Uruguay Round. https://rb.gy/zlmba5
Young A (1991) Learning by doing and the dynamic effects of international trade. Q J Econ 106(2):369–405
Acknowledgements
I am grateful to Dani Rodrik, Michael Ash, Robert Pollin, Arslan Razmi, and Anne Mcgrew for helpful comments and suggestions. I also thank participants in the Analytical Political Economy Workshop at the University of Massachusetts Amherst. The usual disclaimer applies.
Funding
No funding was received for the elaboration of this manuscript.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflicts of interests
The author has no relevant financial or nonfinancial interests to disclose.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Below is the link to the electronic supplementary material.
Appendix
Appendix
See Tables 8, 9, 10, 11, 11 and 12.
Table 8 shows regression results using the robustness checks proposed in this paper to the IV regression carried out by Estevadeordal and Taylor (2013). The framework uses the tariff rate in capital and intermediate goods, and not the full average tariff rate. The results are mixed, but again prove that the positive relationship between trade liberalization and economic growth cannot be ascertained. When dropping the observations of China and Côte d’Ivoire, the results hold. The same happens for the exercise proposed to address the inconsistency in disaggregated tariff data for the cases of Paraguay and Côte d’Ivoire. The results stop holding when the treatment year is moved to 1995, when using PWT 9.1 data, and when using constant prices national data for growth from the WDI. Using the second instrument, as the sample changes substantially, from 44 observations to 31, the only result where significance disappears is the one where the treatment year is 1995.
Table 9 shows that the battery of robustness checks proposed in the paper leads again to lack of robustness in the results by Estevadeordal and Taylor (2013) when controlling for the additional variables considered by Estevadeordal and Taylor (2013), relevant to explain growth performance.
Table 10 shows the results of the original regressions compared to those when China and Côte d’Ivoire are excluded, following the visual inspection in Figure 1. The results reveal that, by excluding these two countries, the significance of the tariff rate on capital goods and intermediate inputs in relation to the change in economic growth disappears, and the magnitude of the coefficient is reduced by almost half.
Table 11 shows the results when changing the treatment year to years 1991, 1992, 1993, 1994, and 1995, in the preferred specification of the authors, thus complementing Table 5. The results show that the magnitude of the coefficient accompanying the trade policy variable is reduced as the treatment year is changing, and the significance is only preserved for the original case and for the alternative case of the treatment year in 1991.
Table 12 shows, finally, the results when all robustness checks are included at the same time. In one specification, the result totally reverts, so that trade liberalization affects growth negatively, and the coefficient is significant, thus confirming the lack of robustness in the results by Estevadeordal and Taylor (2013).
Rights and permissions
About this article
Cite this article
Hoyos, M. Did the trade liberalization of the 1990s really boost economic growth? a critical replication of Estevadeordal and Taylor (2013). Empir Econ 63, 525–548 (2022). https://doi.org/10.1007/s00181-021-02139-8
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00181-021-02139-8