Other control variables consider banks’ costs, which may influence deposit rates. The simple idea is
that banks consider the whole structure of costs when they fix deposit rates. In this paper we consider two
aggregates: the ratio between costs and total assets and the average staff costs per employee.13
Descriptive statistics are reported in Table 2, together with the correlation coefficients of the
independent variables. As expected we get a negative correlation between the concentration indicators and
the number of banks and branches in the provinces.14
4. Econometric results
Table 3 reports the estimates for the dependent variable of the interest rate on total deposits. We
consider the years 1990-99. The regressions contain yearly dummy variables to control the effects on
deposit rates of changes in the monetary policy stance and the business cycle. Dummies are significant,
indicating the importance of the evolution of rates over time.
The Herfindahl index does not have a statistically significant influence on deposit rates. A negative
effect is produced, however, by the concentration ratio R3. It seems that only the market shares of the
most important banks in each province are relevant for the level of interest rates.
Average staff costs per employee (COSPER) have a negative effect on the remuneration of deposits:
the higher the value of this variable, the lower the deposit rate.15 The negative sign of COSPER (Table 3) is
also found in a regression in which it is used as the only variable together with HER (Table 4).
The number of banks per province (NBANKS) has a positive effect on deposit rates (Table 3): a
larger number of banks may correspond to a higher degree of competition or diffusion of financial services,
leading to a higher return for depositors.
13 Neuberger and Zimmerman (1990) find a negative influence of the average wage on deposits’
remuneration; Berger e Hannan (1988) find a coefficient with a positive sign. Other variables which could
influence the deposits’ remuneration are the interbank position and banks’ securities issues; we do not deal
with these variables now.
14 Appendix 2 describes the variables used in the regression.
15 This is the same result of Neuberger and Zimmerman (1990)’s paper. The ratio costs/total assets
(COSRAT) has a positive coefficient (tab. 3). We offer two interpretations of this result. First, banks with
high costs are not able to reduce them in the short period; therefore, they increase their funding to enlarge the
scale of operations. Second, inefficient banks (i.e. with higher costs) are subject to a greater failure risk;
therefore, they must pay a premium to attract deposits. In a regression where COSRAT is used as regressor
with HER, however, the former variable is not statistically significant (table 4).