Historically Black Colleges Pay Higher Costs in Bond Market, Research Finds

April 11, 2016

Professor Bill Mayew finds evidence that race is driving up trader fees for HBCUs

When universities issue bonds to raise money, there's cost involved. Schools contract with underwriters to sell bonds to investors, much like the seller of a home does with a real estate agent. It is the job of bond traders to find a willing buyer and they get paid a search cost based on how much time and effort it takes.

When bond traders told Professor Bill Mayew of Duke University's Fuqua School of Business how much more difficult it was to place bonds from historically black colleges and universities (HBCUs), his curiosity was piqued. The contention was that racial animus among wealthy white investors made it harder to sell bonds and drove up search costs. Mayew and his colleagues — Chris Parsons of the University of California at San Diego, Paul Gao at the University of Notre Dame and Casey Dougal of Drexel University — combed through more than 20 years of public data to see if the story had any merit. They found evidence that race does appear to impose higher search costs on underwriters and bond traders, which ends up putting more financial burden on HBCUs. The findings were recently presented at the National Bureau of Economic Research Corporate Finance meeting and Mayew discusses them in this Fuqua Q&A.

Q: How did you go about studying this effect?

We collected data for 4,145 tax-exempt higher education municipal bond issues between 1988 and 2010, worth a total of $150 billion. The bonds were issued by 965 individual colleges, 102 of which were HBCUs. We looked at differences between HBCUs and other schools in the publicly disclosed underwriter spread - how much underwriters charged schools to take a bond and put it in the hands of investors. We found HBCU issuance costs are about 20 percent higher than non-HBCUs, because it is more difficult for underwriters to find a buyer. To issue a $30 million bond costs an HBCU about $290,000, versus about $242,000 for a non-HBCU. This $48,000 difference is a significant financial burden.

Q: How can you attribute this effect to racial prejudice?

We had to rule out alternative explanations and rule in racism. To reap the largest benefit from the state tax breaks that accompany municipal bonds, investors tend to buy bonds that are issued in their own states. So if you're a school in Louisiana, your investor audience is wealthy people in Louisiana. This allows us to better identify race effects. If race is making it more difficult to sell HBCU bonds, then we should see larger effects in states where there has historically been higher racial animus.

We scored states using a composite of measures established in past academic research examining racism: racial resentment and opposition to affirmative action as measured in the Cooperative Congressional Election Study survey, racially-charged Google searches and frequency of racist tweets after President Obama was elected.

In the three states that ranked highest on those measures — Louisiana, Alabama and Mississippi — we found HBCUs were paying underwriters three times more to place their bonds relative to HBCUs in other states.

Q: Could it be that investors are simply concerned that under-resourced HBCUs are more likely to default?

The municipal bond market has two institutional features that allow us to rule out this possibility. First, credit rating agencies provide a direct measure of credit quality. We see similar HBCU effects even among the issuances that were AAA rated, and the probability of default is virtually zero among AAA rated bonds. Second, many bonds are insured and in such cases the default risk becomes that of the insurer, not the school. If we solely look among the subsample of bonds that are insured, we also see virtually identical HBCU effects. 
Q: Could HBCUs be paying more because they don't have the in-house resources to prepare the bond packages, and thus have to pay underwriters more up front to issue them?

We control for the size and experience of the underwriter syndicate in our analysis, but the best evidence to rule this out comes from secondary market trading data. After bonds are initially issued, the underwriter is out of the equation, but racial animus is not. It should be just as hard for a bond trader to find a buyer in the secondary market when a seller wishes to trade. We find transaction costs in the secondary market mimic those in the initial market — they are roughly 20 percent higher for HBCUs. Specifically, round-trip transaction costs for an average trade are 170 basis points for non-HBCUs versus 204 basis points for HBCUs. Also, in the secondary market, while the bond trader searches for a buyer, the bond must sit in dealer inventory. There's no reason for traders to delay offering HBCU bonds to potential buyers, because holding those bonds in inventory puts the traders at risk. We find HBCU bonds sit in dealer inventory 20 percent longer than non-HBCUs, and so a very consistent pattern of evidence emerges consistent with search costs.

Q: Is it possible that HBCUs are simply less well known among investors in a state, and investors don't want to buy bonds from schools with which they are unfamiliar?

We control in our analysis for various school attributes, such as enrollment, alumni giving and whether the school is public or private. However, the most compelling evidence is the cross-state analysis. If our effects are due to familiarity, it would have to be the case that HBCUs are three times less familiar in Louisiana, Alabama and Mississippi than in other states. It's not clear why that would be the case.

Q: What can be done to mitigate this effect?

Finding a solution is difficult because no one's breaking the law. To make this go away, you have to make the search costs go down. One way to do that is to make the state tax benefit transferable, so schools can better attract buyers from other states. But that would require coordination and compromise among political leaders, which is difficult. Another approach would be to lower the price point for investors to enter this market. Right now the smallest denomination for a bond is typically $5,000. If you made it smaller, that could widen the pool of buyers. However, the smaller the investment the smaller the tax benefit.

Contact Info

Gregory Phillips
Duke University
The Fuqua School of Business
100 Fuqua Drive
P.O. Box 90120
Durham, NC 27708-0125

Tel: 1.919.660.7801
Email: gregory.phillips@duke.edu

Erin Medlyn
Duke University
The Fuqua School of Business
100 Fuqua Drive
P.O. Box 90120
Durham, NC 27708-0125

Tel: 1.919.660.8090
Email: erin.medlyn@duke.edu