Bank competition, relationship banking and integration of financial systems

This dissertation first provides a theory of competition in bank lending; it shows that many extreme implications of Bertrand competition in bank lending disappear when banks engage in independent information gathering about the borrower before lending. It determines that a bank's lending strategy depends on the number of competing banks such that, as the number of banks increases: (a) banks become more conservative in offering a loan; (b) interest rate paid by the borrower with higher (lower) credit quality tends to decrease (increase). It demonstrates that restricted entry into banking gives banks positive profits, which decline with the number of banks. Any change in the cost of funds to banks is amplified in the interest rates on bank loans and affects borrowers with lower credit quality disproportionately. Predictions of the theory are consistent with empirical evidence. Further economic implications are provided.;Based on this theory, the dissertation then provides a theory of relationship banking to study how bank reputation can be an effective enforcement mechanism of bank commitment. It demonstrates that cases, in which increased bank competition or developed bond markets prevent relationship banking, are not general, contrary to a widely held conjecture; competition or developed bond markets may also facilitate relationship banking. It also shows that relationship banking may allow borrowers to raise more funds by issuing bonds than in arm's length banking, contrary to another widely held conjecture. Further testable implications are provided.;The dissertation also studies whether history of financial institutions is enough to lead path dependencies. It demonstrates that, even if relationship banking is selected in a country because of regulatory restrictions on bond markets, it need not disappear after deregulation or financial integration. This possibility of institutional path dependence that does not depend on borrower-specific factors indicates that different financial systems need not converge to a single system.