Assessing bank competition for consumer loans

Bank loans generate more than half of all U.S. bank revenues and differ between business and consumer loans in both size and borrower sophistication.  Consumers are viewed as less informed in financial matters and so are the focus of most state and federal legislation and regulatory concern. This led U.S. Congress to establish the Consumer Financial Protection Bureau after passing the 2010 Dodd-Frank Act. This paper assesses bank competition for consumer loans.

Bank competition for consumer loans is assessed using three standard competition measures---the HHI, an approximate Lerner Index, and H-Statistic---along with a new one based on frontier analysis.  It turns out that these measures are only weakly economically related to one another for U.S. banks. The issue then is, which measure(s) may best reflect the performance and conduct aspects that bank competition for consumer loans is expected to influence and control?  We suggest an answer to this question, explain why the four measures differ in assessing competition, and show how they identify similar or different characteristics of the most and least competitive banks making consumer loans.

In our paper, the four competition measures are defined and estimated for 2,644 U.S. banks using quarterly data over 2008-2010, a period of financial stress for both banks and many consumers.  We illustrate the degree to which each competition measure generally, and at their frequency tails, are related to one another. We analyze how the average loan price, profitability, and industry asset share vary when moving from most to least competitive banks for each competition measure. We then seek to determine a "best" measure of consumer loan competition based on the association of each measure with consumer loan price conduct and profit performance that most would associate with potentially competitive versus uncompetitive behavior.

Using our Frontier Competition Efficiency measure which we suggest is superior to the other competition measures, the sets of banks deemed to be most or least competitive both contain very large banks, with the least competitive banks---as expected---being considerably more profitable as well as having the highest prices for consumer loans.  Both sets of banks are concentrated in smaller U.S. states where, as a group, the decile of most competitive banks operate 40% of the branches.  Average per capita income differs only by 3% between these two sets of states.  Overall, competitiveness is heterogeneous across bank size classes and appears homogeneous across depositor income levels.