intensity of capital controls (lack of offshore interest rates in a number of countries forces us to construct
a synthetic approximation of closed interest differentials using forward exchange rates). Because capital
controls are plausible sources introducing persistence in return differentials, section 2 therefore also formulates
a partial adjustment model of the price measure of financial integration. Section 3 identifies the main
determinants of capital controls. For the purpose of our empirical analysis, we link these determinants to
several relevant analytical indicators. Section 4 analyses the effect of these indicators on the intensity of
capital controls using a panel data approach. We estimate a fixed-effects model, principally attributing country
risk to an country-specific effect which of course differs among countries. The analysis is carried out for
11 EU countries over the period 1973-1993 for which relevant annual data are available. The frequency
of the data was dictated by the absence of higher frequency data on important macroeconomic determinants
of financial integration and by the fact that we want to evaluate long-term trends in financial integration.
Finally, section 5 concludes the paper.
2 Alternative measures of the degree of financial integration
This section addresses the definition and measurement of the dependent variable: the degree of financial
integration. First, we argue that the price measure is to be preferred to concentrating on the volume of capital
flows themselves. As markets become more integrated, asset prices often adjust in anticipation of capital
flows that would otherwise occur. Consequently, the volume of capital flows is less suited to measure the
degree of financial integration.3 In addition, the price measure is also to be preferred to more legal oriented
approaches of capital controls that construct dummy variables or indices of capital controls intensity. Since
financial integration is essentially a legal concept, previous research on the determinants of the degree of
financial integration (Epstein and Schor, 1992, Alesina, Grilli and Milesi-Ferretti, 1994, Gruijters, 1994,
Milesi-Ferretti, 1995 and Grilli and Milesi-Ferretti, 1995) typically constructed dummy variables or capital
control indices to measure the degree of financial integration. Alesina, Grilli and Milesi-Ferretti (1994)
use dummy variables -- taking the value 1 when capital controls are in place and 0 otherwise -- to measure
capital controls. Unfortunately, dummy variables cannot explain different degrees of intensity of capital
controls over time (Epstein and Schor, 1992, p. 143). Epstein and Schor (1992) construct an annual capital
control index compiled from the summary table at the end of the International Monetary Fund (IMF) ’s
Annual Reports on Exchange Arrangements and Exchange Restrictions. This index is composed of restrictions
on payments for capital transactions (i.e. capital controls) and the use of separate exchange rate(s) for some
or all capital transactions and/or some or all invisibles (i.e. dual exchange markets). Both types of restrictions
are given equal weight. If both restrictions are in place, the index takes the value of 2, if one restriction
is in place the index takes the value of 1, and 0 otherwise. Appendix A at the end of the paper gives an
idea of what this means for our EU countries. As follows from Appendix A, capital control indices are
more capable of explaining different degrees of intensity of capital controls than dummy variables. Epstein
and Schor (1992, p. 141), however, argue that the IMF definitions do no include some indirect measures
against capital flows which might reasonably be considered capital controls (e.g. the interest equalisation
tax). Furthermore, the IMF does not distinguish between restrictions to limit capital outflows and restrictions
to limit capital inflows (i.e. the direction of capital flows), and between restrictions on short-term and
restrictions on long-term capital flows (i.e. the maturity of capital flows). Gruijters (1994, pp. 198-213)
tries to overcome these weaknesses and constructs two capital control indices to explain the intensity of
capital import restrictions and the intensity of capital export restrictions, respectively. The indices are based
upon a historical survey of direct and indirect capital controls in 11 OECD countries. The measure implicitly
embodies numerous types of restrictions, with some being more important than others across different
countries. However, the major shortcoming endemic to all legal measures is the subjective element needed
to construct them. Ample historical evidence suggests that there have been significant discrepancies between
Besides, long time-series data of gross capital flows are not available (see Kouri and Porter, 1974).