20
an increase in SIGGOV leads to a significant decline of the generalized country-specific effect (B and E).
So, we agree with the prediction of Alesina, Grilli and Milesi-Ferretti (1994) that capital export controls
are more likely imposed in countries with less independent central banks and in countries with significant
government changes.
The generalized county-specific effects in regression E indicate that Belgium and Italy have potentially
greater risks for investors as follows from the negative country risk premia with respect to regression E.
Judging the negative sign of the generalized country-specific effect, capital export restrictions are more
likely. Nevertheless, the political risks are insignificant. From regression B it follows that also Finland and
Sweden probably have very low generalized country-specific effects. Investments in Finland and Sweden
apparently involve relatively high (but insignificant) political risks.
Finally, we should mention some critical measurement issues. First, the relationship between the intensity
of capital controls and its explanatory variables is often subject to uncertainty. The finding of no significant
association with the intensity of capital controls may simply reflect the crudeness of the measured dependent
and independent variables. Second, controls on capital outflows may also reduce capital inflows, as foreign
investors worry about their ability to transfer income outside the country. Or a country imposes both capital
import as well as capital export restrictions. These aspects of capital controls are difficult to grasp. Third,
multicollinearity problems may exist in regressions A and D. For example, the inflation variable is of course
closely related to the seigniorage revenue and the inflation tax variable. The same may hold for the inflation
rate and the expansion of broad money. Therefore, regressions B and E apply general-to-specific modelling
to find the main (significant) determinants. Lastly, future research may want to consider covered interest
parity for long-term bonds (calculated with the help of currency swaps) as it is unclear if the results go
through in financial assets with maturities of say more than 1 year.
5 Conclusions
The process of financial integration has received considerable attention in recent years. The paper has provided
renewed evidence on the fundamental determinants of the intensity of capital controls. Using closed interest
differentials to measure the intensity of capital controls and applying a panel data approach, we have identified
the fundamental determinants of the intensity of capital controls in the EU. The main empirical results can
be summarized as follows. The most important implication of increased financial integration is that it forces
a greater degree of interest rate parity across countries, and that it reduces the scope for independent monetary
(the interest rate instrument) and fiscal policy in the EU. Although, we emphasize different factors contributing
to deviations from closed interest rate parity, broad common ground exist for the intensity of capital export
controls to depend positively on realized inflation (INF) and government deficits (DEF) (see regressions
B and D). Particularly, this adverse effect of inconsistent national monetary and fiscal policies has been
of crucial importance in explaining the intensity of capital controls after allowing for the persistence in
capital control intensity (see regression E). Furthermore, regarding almost all other arguments considered
in this study, the basic purpose of controls on capital flows has been to provide a certain degree of autonomy
from external economic circumstances. We find credit to the domestic economy (CREDIT) to be significantly
positively correlated with the intensity of capital export restrictions (see regression B). Furthermore, also
low productivity in the business sector (PROD) and low availability of sophisticated deposit instruments
(M2M1) belong to the main determinants of the intensity of capital export restrictions (see regression B).
These explanatory variables were not taken into account by previous studies. Similar to Alesina, Grilli and
Milesi-Ferretti (1994), we find capital export restrictions to be less likely in countries with current account
surpluses (CA) (see regression B).
Remaining differentials from closed interest parity are the consequence of country risk premia imposed