Are minority-owned businesses underserved by financial markets? Evidence from the private-equity industry

Abstract

Our study addresses a longstanding question—whether discrimination exists in financial markets. Although empirical evidence demonstrating disparate treatment of minorities is vast, studies have inadequately explained why minority customers seeking financing are targeted for discriminatory treatment. We develop a theoretical framework explaining why profit-maximizing capital suppliers may choose to offer minority clients worse terms than those provided to comparable white customers. Our framework stresses search costs and reservation prices. We then test this by comparing the relative profitability of investing private equity in minority- and white-owned small firms, an approach advocated by Gary Becker. Using three empirical tests, we consistently find the financial returns derived from investing in minority firms exceed those of white-firm investments. Conducting Becker’s test, in this instance, indicates that disparate treatment of minority clients does not result in loss of profitable investing opportunities for private equity funds but, instead, higher profits.

Appendix: The NAIC Database.

Our due diligence indicated that serious problems compromised the integrity of conventional data sources, like the Venture Economics database, commonly used to analyze the private-equity industry. Researchers familiar with such sources told us of low survey response rates, high item-nonresponse rates, and glaring internal inconsistencies in the data. Creators of the NAIC database sought to generate high response rates from the surveyed funds, along with clean, internally consistent responses, i.e., better data than conventional sources (Bates and Bradford 2002). NAIC member funds strongly endorsed this approach. Institutional investors viewed investing in minority-oriented funds as social investing, resulting in few and small investments by these sources. If the social investing stereotype was to be broken, solid evidence of attractive returns had to be forthcoming, and members (in 2000) felt such returns existed.

The NAIC member funds were first surveyed in 2001. The survey forms completed by each of the surveyed funds contained sections describing all annual cash flows for each individual realized and unrealized portfolio company investment, traits of the individual funds, and traits of the individual GPs. Survey response rates were high; item non-response rates were low. Of the steps undertaken to generate high quality data, particularly important were cross checks designed to ensure consistency and accuracy. The funds were surveyed in 2001, 2004, and 2007. The 2001 survey revealed only 118 portfolio company investments had been fully realized by yearend 2000; many were in the process of maturing. Cash flows for each of these were collected, along with cash flows for those not fully realized.

In the 2004 survey, an important redundancy was built in. Funds responding were asked, in 2004, to list the cash flows of all investments fully realized by yearend 2003, creating 118 completely redundant responses (those realized in 2000). A comparison of the cash flows for each of these 118 investments was conducted to spot inconsistencies between data reported in 2001 and what was reported in 2004, all of which should have been identical. Cash flows reported for unrealized investments in 2000 were similarly compared to 2004 survey responses describing these same investments. Other redundancy checks were conducted. Regarding 2001 and 2004 consistency of cash flows into and out of individual portfolio companies, nearly 10% of the responses flunked. In a few cases, realized portfolio company investments reported in 2001 disappeared entirely in 2004, and each was a clear loser. Funds never forgot their winners. When the lost portfolio firms were reported to the surveyed funds, GPs quickly found them. Second, cash flows for the years 1989 through 2000 reported in 2001 sometimes differed from the same cash flows reported in 2004. Discrepancies were typically minor and embarrassed funds quickly provided corrections. Third, certain investments reported as fully realized in 2001 were not, in fact, fully realized, evidenced by inflows reported in later years. These were corrected.

While most funds reported fully consistent results in 2001 and 2004, one firm operating several funds was seriously off base, and corrections were not immediately forthcoming. Dr. Bradford visited its headquarters and audited their books. The fund GPs attributed the problem to personnel turnover and cooperated fully. The net result of correcting these and other errors was that the net returns pre-correction were higher than they were post-correction. Stated differently, some of the surveyed firms overstated their returns on realized investments, suggesting that cross checks and corrections were very worthwhile.

The end product was cleaned, cross-checked, corrected data. We view this NAIC database as a major source of strength of our analysis. The 2004 process of data cleaning initiated responses from many surveyed funds that raised subsequent data quality. Asked for clarifications, most—particularly the larger funds—simply sent complete copies of their internal spreadsheets describing all cash flows for their individual portfolio companies. When yearend 2006 data were added in 2007, over half of the portfolio company cash flows were recorded directly from such internal spreadsheets. A final issue concerned funds active in 2003 that had failed and shut down by yearend 2006. Among six such funds, four responded fully in 2007. We view the relatively clean, internally consistent data typifying the NAIC database as a major advance, compared to traditional sources like Venture Economics data.