securities industry to find that the industry was in monopoly equilibrium in the 1980s and
was in monopolistic competition equilibrium in the 1990s.
We assess the degree of competition by direct estimation of the first order condition of
profit maximization. This is essentially the method that is formalized by Bresnahan (1982)
and Lau (1982), and has been applied to many studies.2 For example, Shaffer (1989, 1993),
Shaffer and DiSalvo (1994), and Zardkoohi and Fraser (1998) applied this method to the
banking industries in the U.S. and Canada. The method uses time-series data, and
therefore only elucidates a long-run average degree of competition. In contrast, we use
panel data, which enable us to investigate short-run changes in the degree of competition.3
The next section reviews the history of the liberalization in the Japanese banking
sector. In section 3, we present a model for an estimation of the degree of competition.
In section 4, we explain our data and present the empirical results. Section 5 is devoted to
checking the robustness of the results. In section 6, we investigate what elements had
critically affected the degree of competition. Section 7 provides the conclusion.
2. Liberalization of the Japanese Banking Sector
The liberalization of the Japanese banking sector was caused, at least partially, by the
enormous issuance of government bonds from the early 1970s. Initially all the newly
that the results may suffer from some problems.
2 See Bresnahan (1989, 1997) for the development of this method.
3 Angelini and Cetoreli (1999) analyzed Italian banks with panel data for 1983-97 to estimate the
Lerner index.