آزادسازی مالی و رشد اقتصادی: A متا تجزیه و تحلیل

This study provides a systematic analysis of the empirical literature on the relationship between financial liberalization and economic growth by conducting a meta-analysis, based on 441 t-statistics reported in 60 empirical studies. We focus on explaining the heterogeneity of results in our sample in terms of study-, data- and method-specific characteristics. Although our findings indicate that, on average, there is a positive effect of financial liberalization on growth, the significance of this effect is only weak. Moreover, we find that most of the variables that may help explain the heterogeneity of results are insignificant. There are two exceptions. First, studies carried out based on data from the 1970s on average find a statistically less significant relationship between financial liberalization policies and growth (i.e. they report lower t-statistics) as compared to studies using data from the 1980s. Second, studies controlling for the level of development of the financial system report lower t-statistics for the relationship between liberalization and growth.

During the past two decades, many countries have reformed their domestic financial markets. These reforms were triggered by both domestic and international developments. Domestically, many government policies that focused on controlling financial markets were increasingly criticized, for it was felt that these policies impeded the efficient functioning and development of financial institutions. The idea that stagnating economic growth and economic crises are related to financially repressive policies has gained ground since the early 1970s (McKinnon, 1973; Shaw, 1973).1 Moreover, the globalization of financial markets has also put pressure on governments to reconsider financial market controls. The profoundness of these reforms raises questions regarding the potential consequences of financial liberalization on economic growth. Reforms of financial markets include several specific policies which generally aim at higher economic growth. Several authors suggest that the liberalization of financial markets raises the efficiency with which these markets can transform saving into investment, ultimately improving the growth performance of a country. At the same time, however, financial liberalization policies have been criticized for their potential role in triggering financial and economic crises. The question, therefore, is whether or not these policies lead to higher economic growth. Several survey studies have looked into this debate from an empirical point of view.2 The general picture that emerges from these surveys is that the evidence remains inconclusive. However, most of these studies narrowly focus on the effects of capital account liberalization. Moreover, they do not attempt to systematically investigate the outcomes of empirical studies of the financial liberalization-growth nexus. In light of this gap, this study conducts a meta-analysis of the relationship between financial liberalization (rather than capital account liberalization only) and economic growth to provide a more systematic review of the available evidence. The analysis is based on a sample of 60 empirical studies. This is a much larger sample than has been used in any of the previous review studies. Meta-analysis is a methodology that provides a statistical approach to reviewing and summarising the literature (Stanley, 2001). This methodology allows us to draw a comprehensive picture of the impact of financial liberalization on growth. By using meta-analysis, each study is taken as one single observation containing information on the nature of the relationship between financial liberalization and economic growth. Recently, a growing number of meta-analyses have been published in economics on issues such as the relationships between aid and growth (Doucouliagos and Paldam, 2008, 2009), central bank independence and inflation (Klomp and de Haan, 2010), investment and uncertainty (Koetse et al., 2009), economic freedom and growth (Doucouliagos and Ulubaşoğlu, 2006), democracy and growth (Doucouliagos and Ulubaşoğlu, 2008), income inequality and growth (De Dominicis et al., 2008) and fiscal policies and growth (Nijkamp and Poot, 2004). To our knowledge, this is the first study using meta-analysis as a tool to investigate the financial liberalization-growth nexus. In our analysis, we focus on exploring the heterogeneity of findings in 60 different studies. First, we investigate whether the choice of the financial liberalization measure has an impact on the results reported in different studies. Second, we analyse the potential impact of study design. In particular, we focus on the impact of differences between studies regarding country samples, time periods, and estimation methods. Third, we explicitly focus on indirect effects of financial liberalization on economic growth. In a separate section, we also investigate whether studies suffer from a potential publication bias, i.e. whether published results provide a biased distribution of effect measures (here: t-statistics), because there may be a tendency not to publish results that show insignificant results. The remainder of this study is organized as follows. Section 2 provides a review of the debate on financial liberalization and its potential effects on economic growth. In Section 3 we discuss how studies have dealt with the measurement of financial liberalization. Section 4 provides an overview of the data collection procedure and discusses the descriptive statistics. Section 5 discusses the methodology and the results of the meta-analysis in detail. The study ends with a conclusion, in which we also discuss the limitations of this review and suggestions for further research.

Since the early 1970s, the relationship between financial liberalization and economic growth has been subject of debate, both in policy as well as academic circles. Academic research investigating the nature of the relationship between financial liberalization and economic growth remains inconclusive. In this study we aimed at providing a systematic analysis of the empirical literature by conducting a meta-analysis of the relationship between financial liberalization and economic growth based on 60 different empirical studies and 441 t-statistics. As far as we know, this is the first study using meta-analysis as a tool to investigate the financial liberalization-growth nexus. In the analysis, we concentrated on explaining the heterogeneity of results regarding the relationship between financial liberalization and economic growth reported in the studies in our sample. In particular, we focused on a long list of study-, data- and method-specific characteristics. The meta-regression analysis provided the following main results. First of all, we showed that, on average, there is a positive albeit weak effect of financial liberalization on growth. This result suggests that financial liberalization is not a panacea for achieving strong economic growth. Liberalization policies may perhaps be more successful if combined with other reform measures and/or institutional changes, such as fiscal and/or monetary policies, and/or institutional changes focussing on regulation of financial markets (Kose et al., 2010). Empirical research on the growth effects of such combinations of policies is, however, hardly available. It is true that some studies use measures such as inflation rates and government expenditure as control variables. Although these variables could, at least theoretically, be linked to monetary and/or fiscal policy changes, changes in these variables may also be the result of many other confounding factors. We have therefore refrained from further digging into the issue of the combined impact of financial liberalization and other government policies on economic growth. Instead, we suggest that this may be a very fruitful direction for future research. Next, for most of the variables that may help explain the heterogeneity of results regarding the relationship between financial liberalization and economic growth we do not find significant results. There are two exceptions. First, in many of our specifications data from the 1970s generate more negative financial liberalization coefficients. Second, our results show that studies that take into account a measure of the level of financial development report lower t-statistics for the relationship between liberalization and growth. This outcome may show that for countries with less developed financial systems, financial liberalization has more value in terms of stimulating economic growth. These results remain valid after having carried out a large number of robustness checks, including a test for publication selection bias. The results regarding the importance of the financial system are especially interesting and we therefore suggest future research should focus on analysing the exact nature of the relationship between financial liberalization, financial depth and economic growth. We end by making a few qualifying remarks. First, the quality of our meta-analysis can only be as good as the quality of the empirical analysis of the underlying studies. The studies in our dataset are carried out in the country growth regression tradition. This literature has received quite some criticism recently, especially with respect to potential endogeneity problems. However, solving these problems in a convincing and satisfying way has proven to be very difficult in this strand of literature. Second, the studies in our sample are based on data from a limited set of observations (i.e. all countries in the world) and in many cases the data come from similar sources such as IMF and World Bank databases. We acknowledge the potential limitations of the underlying literature and data they use and recommend that our results be interpreted with some caution. Notwithstanding these qualifying remarks, we do believe that the results of our meta-analysis are valuable as they come from the most comprehensive and systematic overview of the literature on the relationship between financial liberalization and growth available to date.