second group consists of 44 countries, including industrial as well as emerging
economies. The sample covers the period 1990-2003 14. Trade integration and per-
capita income convergence are estimated for each country i vis-à-vis the European
Union-15 (EU-15) average, which is therefore taken to be the reference partner. That
is, in both equations, j is represented by the EU-15 average. This makes it possible to
assess the effect of financial openness on the process of economic integration of
country i with the EU-15. In fact, the main findings are qualitatively unchanged if the
United States or the richest among EU-15 economies are used as reference partners.
To account for reverse causality; that is for the possibility that financial openness is
determined by trade volumes and per-capita income growth, equations (3) and (6) will
be estimated by 2SLS, using lagged and initial values of endogenous variables as
instruments. The estimator is further corrected to account for the fact that the annual
panel is unbalanced.15
To operationalise equation (3), y is measured by a country's real per-capita GDP, n is
proxied by the fertility rate, h is proxied by the enrolment rate in tertiary schooling, k
is proxied by the real investment share of GDP, q is defined as country's ratio of M2
minus narrow money to narrow money. In equation (6), instead, T is measured by a
country's exports to and imports from the EU-15 (in logs of millions USD), Y is given
by real aggregate GDP and D is the log of distance (in kilometres) between a country
and Frankfurt-am-Main. A complete list of variables, definitions and sources is given
in the Appendix. Moreover, the next section will discuss the sensitivity of
econometric results to changes in variables definition and construction.
14 The panel is however unbalanced as for some countries the first available observation comes later
than 1990. The group of emerging economies includes: Albania, Armenia, Azerbaijan, Belarus,
Bulgaria, Croatia, Czech Republic, Estonia, Hungary, Kazakhstan, Kyrgyz Republic, Latvia, Lithuania,
Moldova, Poland, Romania, Russian federation, Slovak Republic, Slovenia, Ukraine. The group f 44
economies includes all of the above emerging economies plus: Austria, Belgium, Canada, Cyprus,
Denmark, Finland, France, Germany, Greece, Iceland, Ireland, Israel, Italy, Luxembourg, Malta,
Netherlands, Norway, Portugal, Spain, Sweden, Switzerland, Turkey, United Kingdom, United States.
15 The unbalanced panel estimator follows Verbeek and Nijman (1996). An alternative to the 2SLS
instrumental variable estimator would be a 3SLS system estimator (see Wooldridge, 2002). In this case,
equations (3) and (6) are estimated as a system together with an equation where financial openness is
the dependent variable and trade and per-capita income enter as explanatory variables. In fact, a set of
estimates from the 3SLS procedure are available from the authors upon request. The qualitative thrust
of results does not change relative to the single-equation 2SLS presented in the next section. We prefer
reporting the 2SLS and not the 3SLS because the focus of this analysis is more on the estimation of
reduced-form equations than on structural models.
10