18
The government debt ratio (DEBT) enters insignificantly positive in the regression. Resulting, higher debt
ratios are positively related to capital import controls -- contrary to our prior expectation. Apparently, foreign
investors demand compensation in the form of higher domestic returns to compensate for the increased
probability of loan default. Current account surpluses (CA) are associated with capital import restrictions.
With relatively high domestic credit expansion (CREDIT) more capital export restrictions are in place. Little
evidence was found for justifying an important role of broad money (M2) and the unemployment rate (UN)
in explaining the intensity of capital controls. Monetary policies that accommodate high rates of domestic
credit expansion should lead to lower domestic interest rates relative to offshore interest rates and a negative
coefficient for M2 arises. However, if such accommodating monetary policies are perceived to eventually
lead to higher future inflation or current account deficits a positive coefficient for M2 arises. The openness
of the economy (OPEN) and the productivity of the business sector (PROD) are effective control variables.
The productivity in the business sector provides an important (significant) indication of the direction of
capital flows and, hence, for the objective of capital controls. Low productivity of the business sector may
be an influential argument for restricting capital exports. Apparently, even with capital export restrictions,
capital find their way to the most productive investment opportunities. The openness of the economy is
positively related to the presence of capital export restrictions. New evidence is found to support the
hypothesis of the increasing relevance of financial market structure for the intensity of capital controls.
With less-developed financial markets (M2M1) relatively higher capital outflows are expected. Accordingly,
capital export restrictions are more feasible with low ratios of M2 over M1.
Regression B finds the variables INF, DEF, CA, M2M1, CREDIT and PROD to be the most important
explanations of capital control intensity. Regression B indicates the existence of significant country risk
premia (see country-specific effects ζi in Table 2).34 The countries are listed in ascending order of estimated
country risk premia (ζi). Consequently, it may be important to include the lagged dependent variable as
an explanatory variable in the regression.
In regression C we explain closed interest differentials by the persistence in the intensity of capital controls
as measured by coefficient β1 in equation (19).
N t∙β.∏ . ∙c^ (19)
Coefficient β1 reflects the direct effect of capital controls in place. The country-specific effects capture
all other determinants of closed interest differentials. The within-group estimator for fʒɪ is 0.548 with standard
error of 0.055 and a coefficient of determination adjusted for degrees of freedom (R2) of 0.33. Again, the
countries are listed in ascending order of estimated country risk premia (ζi). This regression may be seen
as the reference point for regressions D and E identifying the fundamental determinants of the financial
integration in the EU.
We now arrive at our principal models of the fundamental determinants of capital controls. The estimation
results of the partial adjustment model in equation (20) are summarized in regressions D and E of Table
2.
(i-ia")i., = G÷ β1(i-iaro)i,t-ι ÷ β2xi,,÷ei3 (20)
In regression D all determinants are included. Regression E is equivalent to D but now only reports the
most significant variables applying general-to-specific modelling including only the significant variables.
Regressions D and E tests for the relative importance of the persistence in capital controls (i.e. lack of
34 The significance may also be caused by the common constant (β0) to be significantly different from zero.