accordance with the Monti-Klein model, in which the bank is price-maker in the deposit and loan markets.
Conigliani and Lanciotti (1979) found a negative correlation between concentration and deposit rates.
More recently, in the new situation of liberalized banking markets, Italian studies have pursued some
subjects which recall the American literature. Pittaluga (1994) criticized the structure-conduct-performance
approach, because in Italy non-price competition became more important in recent years. More precisely,
urban banks applied price competition in the deposit market; while rural banks mostly used forms of non-
price competition, such as enlarging their branch network. Corbisiero and Pesaresi (1999) studied
competition in the provinces in which large mergers had been concluded. These merges caused an initial
increase in concentration followed by a subsequent decrease. This result stems from the redistribution of
market shares from the merged banks to new entrants which adopt more aggressive strategies. The merged
bank increases its market share only if it applies higher rates on deposits. This paper, like others in USA,
emphasizes the role of the new competitors in influencing the incumbents’ behavior.
Looking at macro data, deposit rates seem more upwardly rigid than money market rates in Italy as
well. Figure 1 shows that in the years 1995-98 rates on current accounts and certificates of deposit with a
maturity of six months slowly followed the increases in six-month Treasury bill rates; on the contrary,
deposit rates were quick to follow reductions in Treasury bills’ yields (the only exception concerns some
months in 1995). This phenomenon is stronger for current accounts, while CD rates appear more reactive
to market conditions7.
In this paper, however, attention is focused on micro data, whose analysis appears to be of interest
even on the basis of the few Italian studies on the same subject.
3. The data and the hypotheses to test
The data on interest rates have been taken from the Central Credit Register. A sample of banks
produce quarterly data on deposit and loan interest rates. Deposit rates consider deposits of residents in
lire greater than 20 million lire (about 10,000 €).
In this paper, the data refer to the 95 Italian provinces and consider around 45 banks.8 The
dependent variables in the regressions are the interest rates on current accounts, savings accounts,
7 On bank deposits in Italy see Cesarini, Conti, Di Battista (1994) and Di Battista (1996).
8 The sample contains 41 banks in 1990, 46 in 1991, 47 in 1992, 47 in 1993, 47 in 1994, 47 in 1995, 46 in 1996,
45 in 1997, 44 in 1998, 43 in 1999. Mergers are mainly responsible for these annual changes.