indicate whether a merger is characterized by low or high efficiency gains in terms of both
reduced loan provision costs and lower financing costs (— high or low); the two columns
show the cases of high and low relative cost of refinancing rD.
[TABLE 2 ABOUT HERE]
As Table 2 shows, the model predicts several scenarios, depending on the value of the
parameters. The effect of mergers on expected aggregate liquidity needs is ambiguous when
the relative cost of refinancing is low, whereas it is always negative when the relative cost
of refinancing is high. Concerning competition, mergers increase loan rates when the cost
efficiency effect, as measured by the ratio cm, is small relative to the increased market power,
cc
and, vice versa, decrease loan rates when cost efficiencies dominate.
What can we say about the plausibility of the different scenarios displayed in Table 2? As
already indicated above, one may associate for example low refinancing costs with industrial
countries and high refinancing costs with developing or emerging countries. Moreover, one
may relate the magnitude of efficiency gains to the size of mergers. Even if there is an ongoing
debate in the literature on whether efficiency gains (and, in particular, scale economies)
exhaust at large or small sizes of output, the empirical consensus seems still to be that
mergers between small banks produce larger efficiency gains than mergers between large
banks.10
One plausible scenario in industrial countries (low rrD ) is therefore the occurrence of
mergers between large banks leading to higher loan rates and expected aggregate liquidity
needs, as they do not realize sufficient efficiency gains and induce greater asymmetry in the
banking system (one case in cell I). Differently, the occurrence of mergers between small
banks in industrial countries is likely to reduce both loan rates and expected aggregate
liquidity needs, as smaller mergers may realize more efficiency gains relative to the increase
in market power and make the banking system more homogenous (one case in cell II). For
developing countries (high rrD ), cells III and IV suggest that mergers would always increase
10 A substantial amount of empirical research has been spent on measuring the efficiency gains generated
by bank mergers, but results are not unanimous (see, e.g., the surveys of Carletti et al., 2002; and Rhoades,
1994 and 1998). Whereas the mainstream literature suggests that banks exhaust potential scale economies
at modest levels of size (see, e.g., Berger et al., 1987; Berger and Humphrey, 1991; and Wheelock and
Wilson, 2001), other studies (e.g., Berger and Mester, 1997; and Hughes et al., 2001) find that there are scale
economies also at large balance sheet sizes if one takes changes in risk into account.
26