Journal of Development Economics, July 2009, vol.89, iss.2, pp.210-222. Available From:Link to Source
This paper evaluates the relationship between financial liberalisation, growth and volatility at the industry and aggregate country levels. A wealth of theoretical work hypothesises that liberalisation has a positive impact on growth but a magnifying effect on volatility. However, to date no empirical work has examined the impacts on both growth and volatility within a single empirical framework. Using an industry-level panel data set of 56 countries, the authors explore the impact of financial liberalisation on growth and volatility, the channels through which liberalisation leads to these impacts, and whether these impacts are temporary or permanent.
The authors first construct a difference-in-differences model using the non-liberalising countries in a given time period as the control group for each country that liberalised at this time, and then using a propensity score matching (PSM) algorithm to identify a specific non-liberalising country as the control group for each liberalising country. Next, they compare these results to the results of a difference-in-differences sector-level model , comparing the impact of liberalisation on sectors within a given country based on their level of dependence on external finance. This model has the advantage of controlling for both observed and unobserved country-level characteristics, whereas PSM can control only for observed country-level characteristics. Finally, the authors present the results of both models using both de jure (in law) and de facto (in practice) levels of financial liberalisation. The former is of primary concern for the study’s policy implications, whereas corroborating the results with de facto measures of liberalisation increases the robustness of the results.
To examine the channels through which liberalisation affects growth and volatility, the authors use a standard growth accounting framework. According to this framework, the impacts of liberalisation may work through increased firm entry or through growth in total production, including capital accumulation, increased employment or total factor productivity. Finally, to examine the permanence of these impacts, the authors divide the 12 years after liberalisation into four 3-year periods and calculate the growth and volatility over those periods using a PSM model.
The three questions posed by this paper concern how financial liberalisation affects growth and volatility, through which channels it does so and whether the change is temporary or lasting. The difference-in-differences analysis shows that de jure financial liberalisation increased both the rate of output growth and the level of market volatility significantly, by standard deviations of 0.17–0.4 and 0.13–0.21, respectively. The alternative model based on sector characteristics reveals that those industries with relatively more reliance on external finance had faster growth rates and more volatility. The de facto measure of liberalisation produced the same results as the de jure measure.
The impacts appear to have worked primarily through capital accumulation and increased employment (positive and significant across all models), with weak results indicating an increase in firm entry (significant only in the PSM model), and no indication from any model of an increase in total factor productivity. The analysis of long-term growth and volatility reveals that both impacts were transitory. Growth attributable to liberalisation rose immediately, accelerated, and then fell to a level statistically insignificant from zero by the end of the 12-year period.
Volatility likewise rose immediately, but fell to the point of being indistinguishable from pre-liberalisation levels in subsequent periods. There was also a persistent negative impact on the average product mark-up, suggesting long-term pro-competitive impacts of liberalisation. Though the growth impacts were temporary, the temporary increase in growth led to a long-run-level increase in wealth. This result, paired with the negligible increase in long-run volatility, leads the authors to conclude that a long-run welfare benefit is gained from financial liberalisation.