Based on an unbalanced panel of all Bavarian cooperative banks for the year
s of 1989-97, which includes information on 283 mergers, we analyze motives
and cost effects of small-scale mergers in German banking. Estimating a fr
ontier cost function with a time-variable stochastic efficiency term, we sh
ow that positive scale and scope effects from a merger arise only if the me
rged unit closes part of the former branch network. When we compare actual
mergers to a simulation of hypothetical mergers, size effects of observed m
ergers turn out to be slightly more favorable than for all possible mergers
. Banks taken over by others are less efficient than the average bank in th
e same size class, but exhibit, on average, the same efficiency as the acqu
iring firms. For the post-merger phase, our empirical results provide no ev
idence for efficiency gains from merging, but point instead to a leveling o
ff of differences among the merging units.