and forestry cooperatives, are much smaller, so they probably compete within more limited
areas. Therefore, the analysis of these institutions should require a somewhat different
methodology than that applied in this paper, which should be interesting as a future study.
Regional banks, together with city banks, constitute ‘ordinary banks’. Legally,
regional banks are not distinguished from city banks, but they form their own business
organizations.11 They differ from city banks in that they are much smaller and basically
operate in restricted areas. However, Kano and Tsutsui (2003) found that the loan markets
of regional banks are not segmented by prefecture. In 1996, their share of the loan market
in Japan was 33.1%.12
City banks have nation-wide branch networks and operate diversified businesses,
including international bank business, while regional banks tend to be specialized in
traditional deposit-loan business. In 1996, city banks’ share of the loan market in Japan
was 49.6%. Thus, by analyzing regional and city banks, we cover 82.7% of the total
amount of loans of Japanese banks. We estimate the model with separate samples because
city and regional banks have different production functions and their markets are somewhat
segregated.
We used panel data of city banks and regional banks for the estimation. All the data
were extracted from Nikkei NEEDS Company (Bank) Data File, except for those of IIPt ,
11 Regional banks consist of regional banks and second regional banks. The latter were converted
from mutual banks during 1989-92, and had previously been specialized financial institutions for
small- and medium-sized firms.
12 The total loan outstanding in Japan is defined as the sum of city banks, long-term credit banks,
9