The impact of financial liberalization on bank efficiency: evidence from Latin America and Asia

Abstract

This article investigates the impact of financial liberalization on bank efficiency, using data for a sample of over 4000 bank-year observations from ten emerging economies for the period 1991–2000. We use Data Envelopment Analysis (DEA) to calculate bank efficiency at the individual bank level. Bank efficiency measures are then aggregated at the country level to investigate the relationship between financial liberalization and bank efficiency, using a panel least square fixed-effects model. Overall, we find strong support for the positive impact of financial liberalization programmes on bank efficiency.

Notes

¹ During the 1970s and 1980s, countries have also experimented with financial liberalization policies. Especially during the 1970s, countries such as Argentina, Chile and Uruguay implemented financial liberalization. However, the wave of financial liberalization policies was most apparent from the early 1990s.

² Note that financial liberalization may also have quantity effects, i.e. it increases the amount of resources that are intermediated between savers and investors. By introducing market principles and competition in financial markets interest rates on deposits be raised, leading to higher saving and investment rates. This is not the focus of this article, however.

³ One of our referees pointed out that there may be a quality dimension of loans that is unmeasured in this study, and more generally in the literature on bank efficiency and financial liberalization. The quality dimension refers to the fact that before liberalization, many loans may require low screening and monitoring due to government involvement and may thus be low-quality loans, whereas after liberalization loan decisions are taken by banks that put in screening and monitoring efforts, potentially leading to higher quality loans. Thus, a simple ratio of quality–unadjusted loans to inputs may be a poor measure of efficiency. Although we agree with the referee on this point, we would also like to stress that it is very difficult to come up with better measures of efficiency, which take into account such quality adjustments. Based on the available data, we were not able to produce better measures, which is why we have used a crude measure like loans to inputs, a measure that has been used also in several other studies in this literature.

⁴ The data in Table 1 show that the financial liberalization index does not reverse for any of the countries in any of the years.

⁵ Laeven (2003) provides financial liberalization data for 13 countries. In the analysis we have left out Chile, Malaysia and Taiwan, because BankScope provides data for only a few banks from these three countries during the period 1991–1996.

⁶ Except for Pakistan in 2000; in this case OTE and PTE are equal.

⁷ t-Value is 1.37.

⁸ The average PTE of Latin American banks over the period 1991–2000 is 0.829; for Asian banks this is 0.813.

⁹ Fries and Taci (2005) also find a negative coefficient for DD. In their study the coefficient is not statistically significant, however.