13
246), financial market structure refers to "[...] the mix of finance between equity and debt and to the source
of funds -- through financial markets or through financial intermediaries." The term "financial structure"
may be defined fairly broadly to include such features as the degree of development of money and financial
markets; the degree of competition within the banking system, and between banks and other intermediaries,
the existence of constraints on capital movements; the ownership structure of the financial intermediaries
(Cottarelli and Kourelis, 1994, p. 590). Among the facts that interfere with the relationships between financial
integration and financial market structure are the persistent difference in bank-based and market-based systems,
the uneven development and structure of financial balances and the unequal responsiveness of interest rates
to monetary policy (Bank for International Settlements, 1994, pp. 136-140). In many of the national financial
markets bank deposit and loan rates were subject to regulation and other distortions (see Marston, 1995,
Chapter 2). Policies that seek to maintain nonmarket interest rates on loans or deposits are undermined
by international arbitrage unless backed up by controls. Taxes on financial intermediation can cause financial
activity to move offshore unless backed up by controls.
Unfortunately, it is rather difficult to find relevant indicators of financial market structure. Due to data
availability we had to employ rather crude measures of financial market structure.27 We conjecture that
the intensity of capital controls is negatively related to the development of the financial system. We
hypothesise that less-developed financial markets are characterized by relatively low intermediation by
financial institutions to the domestic economy (CREDIT); that is, domestic credit expansion is relatively
low, and relatively low ratios of broad money (M2) over narrow money (M1) (M2M1); that is, the
development of more sophisticated deposit instruments is relatively low.28 Both variableds were not included
in previous research. On the other hand, CREDIT may also indicate inconsistent monetary policies and
the presence of capital export restrictions. So, we will let the data decide which of the two arguments holds.
4 The fundamental determinants of financial integration in the European Union: Empirical results
and interpretation
This section determines empirically the fundamental determinants of financial integration in the EU. Because
we would like to infer conclusions about a country-specific effect which cannot be observed directly, we
resort to the use of panel data. With the use of panel data the partial adjustment model specified in equation
(6) of section 2 takes the following form:29
(i - (Eu)°\,t ( β0 + βι (i - *aro)i,f.1 ÷ β2xi∙.i ÷ γi∙ (ei,t (10)
where β0=ψδ1 is the common intercept term, β1=(1-ψ) is the partial adjustment coefficient, β2=ψδ2 is
coefficient for the determinants of the intensity of capital controls, (i-iEuro)i,t is the closed nominal interest
rate differential of country i in period t, (i-iEuro)i,t-1 is the closed nominal interest rate differential of country
i in period t-1, xi,t are the determinants of the intensity of capital controls excluding EXR, ES, LEFT and
SIGGOV of country i in period t, γi is the country-specific effect and εi,t is the error term for country i
27 To our knowledge, only the OECD has comparable data available to examine sectoral financial balances (OECD, Financial
Statistics, Part II, Financial Accounts of OECD countries). See also Mooslechner (1994) and Lemmen and Eijffinger (1995b).
28 Of course, there are more relevant indicators such as the market share of the five largest banks, or the number of bank branches
per 100,000 inhabitants etc. (see e.g. King and Levine, 1993 and Cottarelli and Kourelis, 1994). Since data are difficult to obtain,
we leave this for further research.
29 Actually, we should write β3γi, but this term can only be estimated in a composite form. We are not able to disentangle the
constant coefficient β3 and the country-specific effect γi. So, without loss of generality β3 can be normalized at one.