Capital controls and their effects on private investment
This dissertation estimates the effects of capital controls on investment, using two alternative investment models: an empirical investment model motivated by the neoclassical flexible accelerator model and a structuralist investment model. Two new dummy variable capital control indices are constructed and are used along with covered interest rate differentials as proxies for capital controls. Based on annual panel data for 12 countries (11 industrialized countries plus Malaysia) from 1986 to 1997, both models show that restrictions on short-term capital outflows (inflows) have negative (positive) effects on domestic investment. This robust finding is obtained using both covered interest rate differentials and dummy variables as proxies for controls on short-term capital flows, although the results with dummy variables are more significant in the structuralist model. The effects of controls on short-term capital flows in general, regardless of direction, were tested using absolute values of covered interest rate differentials and with squared covered interest differentials. The results of these tests were coefficients with varying signs that were often insignificant and very sensitive to model specification. The panel data investment models also yield ambiguous results about the effects of controls on inward or outward foreign direct investment on private investment. Dummy variables for direct investment inflow (outflow) controls have negative (positive) effects in the empirical model and positive (negative) effects in the structuralist model, possibly due to differences in the other variables included in the two models. The individual time-series version of the model using quarterly data for eight industrialized countries gives ambiguous results about the effects of short-term capital controls on private investment, using covered interest rate differentials. These results suggest that the most of the findings about the effects of short-term capital controls in the panel data regressions are due to international cross-sectional effects. Overall, the results imply that the effects of capital controls on investment depend on the economic structure and overall economic policy regime in each country. The empirical tests in this dissertation were limited by the availability of the necessary data and future research with larger data sets for more countries can help to clarify all of these issues.