the riskiness of banks’ activities abroad due to lacking information (informa-
tion constraint). In the two-country setup, the incentive constraint means
that the country A supervisor does not take excess costs of a national bank
breakdown into account which accrue in country B. Hence, the excess cost
factor bA in the regulator’s objective function reflects only repercussions on
the national banking system and the home economy. bA is lower than the
excess cost factor in the case of a supranational regulator. With rising in-
tegration of the two markets, the national excess cost factor increasingly
deviates from the supranational one. According to corollary 3, the probabil-
ity of failure would be too high as the bank is not forced to fully internalize
the negative effects of bankruptcy. The same holds for country B. Thus, the
overall probability of failure is inefficiently high compared to the suprana-
tional optimum and is the higher, the more integrated the interbank market
is.
Besides, the information deficit means that country A supervisor does
not know qB and, thus, is unable to assess the bank’s realized quality qA.
However, incorporating imperfect monitoring is tedious and basically pre-
vents supervisors from fully eliminating bank rents. As all qualitative results
continue to hold, I abstain from explicitely modelling this variant and in-
stead refer the interested reader to Lewis and Sappington (1989) who model
imperfect monitoring in a related context.
In summary, a supervisory regime where national supervisors lack infor-
mation on banks’ activities abroad and where they do not take international
spillovers into account is suboptimal: The induced probability of bank failure
is inefficiently high. Hence, a mechanism is to be implemented which aligns
the national supervisors’ incentives such that bA fully reflects the costs of
bank failure independent of where these costs accrue and such that supervi-
sors are forced to appropriately exchange information. Such a mechanism is
described in the next section.
However, before moving on to the next section, I analyze the effect of
improved information exchange. Using (2) and assuming a mechanism which
implements efficient information exchange among supervisor in country A
and B, the implicit definition of equilibrium probability of failure (13) be-
comes
G((1 k*)L + γ (L) |(1 - f )(qo,A + eA) + fqB )
re 1
bA
L
where eA is the optimally induced effort of bank A.
As can be seen above, the quality of banking system B can partly sub-
stitute for costly effort of banking system A. The same quality level can,
14