Large Banks and Efficient Banks: how Do they Influence Credit Supply and Default Risk?

Abstract

This study examines two questions relating to the banking market structure. First, does the banking market structure influence banks’ decisions to originate new single-family home mortgages? Second, does the banking market concentration affect mortgage default risks? Using a two-stage approach with the inputs from two data sources on the banking market in the US and mortgages in non-agency securitization pools for the period from 1999 to 2008, we find that banks operating in the markets with a low entry barrier (efficient banks) increase credit supply, while banks possessing market power restrict credit supply to the mortgage markets. Banks with market power originate loans that have lower default risk compared to loans originated by banks in the competitive markets. Efficient banks use mortgage technology indiscriminately to increase credit supply even at the expense of lowering credit quality (increasing default risks). We show that the effects of banking market structure are not correlated with legislation risks and population size in the markets.

Appendix: Measures of contestability and concentration of banking markets

1.1 Panzar and Rosse (1982, 1987) index (PR-H)

The PR-H statistics proposed by Panzar and Rosse (1982, 1987) is estimated by first regressing the logarithm of total interest income, π, against annual expense on funds, F, personal expense, PE, and physical capital expense that include furniture, fixture, equipment and auto, PCE, as follows:

$$ \ln \pi ={\beta}_0+{\beta}_1\ln F+{\beta}_2\ln PE+{\beta}_3\ln PCE+\varepsilon $$

(6)

The regression (7) is separately estimated (560 regressions) for the 50 US states and 6 other territories / islands (American Samoa (AS), Federated States of Micronesia (FM), Guam (GU), Puerto Rico (PR), Rhode Island (RI) and Virgin Islands (VI)), each year for the sample periods from 1999 to 2008, to derive at the (56 × 10) matrix of regression coefficients, [β₁, β₂, β₃]. The state-year PR-H statistics are then computed as the sum of the three input factor elasticity:

$$ PR-H={\beta}_1+{\beta}_2+{\beta}_3 $$

(7)

A bank with a high PR-H statistic is known to operate efficiently in producing loan services at the lowest marginal costs.

1.2 Herfindahl-Hirschman index (HHI)

Herfindahl-Hirschman Index (HHI) is a widely used measurement for market concentration. In computing HHI, we first determine the market share of each bank in a market sorted by state and by year:

$$ {\uprho}_{\mathrm{i}}=\frac{{\mathrm{E}}_{\mathrm{i}}\ }{\sum \limits_{\mathrm{i}\in \mathrm{j}}{\mathrm{E}}_{\mathrm{i}}} $$

(8)

where ρ_i is the market share of a sample bank in a state j. E_i is the total equity value of the bank. $\sum \limits_{\mathrm{i}\in \mathrm{j}}{\mathrm{E}}_{\mathrm{i}}$ is the cumulative equity value of all bank i in a state j.

The squared market share of each banks in state j is added up to give the HHI value, which ranges from close to zero to a large number:

$$ \mathrm{HHI}={\uprho}_1^2+{\uprho}_2^2+{\uprho}_3^2+{\uprho}_4^2+\dots +{\uprho \mathrm{S}}_{\mathrm{n}}^2 $$

(9)

where n is the total number of banks in the market segment j, and the total number of banks could be limited by a ceiling of 50 banks. A small HHI value indicates that a market is less concentrated; and a large HHI value indicates that a market is highly concentrated. If there is only one bank in the market, the value would be one. When the value approaches to zero, the market is perfectly competitive.