Race Discrimination in Higher Education Bond Markets

Race Discrimination in Higher Education Bond Markets

October 13, 2019 1 By Ronny Jaskolski


This is Duke University. Today we’re going to talk about
race discrimination in higher education bond markets. Race discrimination is
an ongoing hot button issue in our society,
and usually when we hear the term
race discrimination, we think about one race– whites– treating a
different race– blacks– in an inappropriate
manner, or in a way that suppresses or makes one
race inferior to another. We commonly think about
this in labor markets. You hear a lot of
stories about how a white employer
has a preference to not hire a black worker– maybe not hire them,
maybe pay them less. In economics, we would call
that taste-based discrimination, where you hire someone based
on your personal preferences, for whether you want
to be around them. That’s what we think about
when we say the word race discrimination. But in real life, when you
start pushing, and you say, are you really discriminating
against this individual? Some would say, no, no. I’m not a racist. It’s not that I don’t like
the color of their skin, it’s that the
color of their skin tells me something about them. So back to the labor example. The white employer
might say no, it’s not because they’re black, per se. It’s because, on average, blacks
come from poorer neighborhoods, which means they may not have
health insurance, which means they might miss work more. And so that’s the reason that
I don’t want to employ them, not because they’re
black per se. In economics, we
have a name for that. That’s called statistical
discrimination, where the color of
your skin actually tells you something
statistically, or on average, about someone. When you’re studying
race discrimination, distinguishing between those two
issues is critically important. And it’s very difficult
to do in many settings. We’re going to try to
crack the code on this, or try to make some progress
in understanding race discrimination by looking at the
higher education bond market. So there’s two pieces
there– higher education and bond market. So why would we focus
on higher education– colleges and universities? The reason for that is we
can delineate them into, in some sense, black
schools and white schools. So the Higher
Education Act of 1965 designates certain
schools as what we call “Historically Black
Colleges and Universities.” And obviously, if you
get that designation, we can clearly and
objectively delineate schools between black and white. Another novelty,
or another reason that it’s important
this is a novel setting, is because that same Higher
Education Act of 1965 acknowledged that,
in historical terms, the federal and
state government had discriminated against these very
schools in terms of funding. And so we have a nice historical
setting where discrimination has been found to exist,
and what we’re going to do is see if that discrimination
translates over into the municipal bond market. So you might say,
well, why bond market? And why are you doing that? Well, it turns out that
bond market investors tend to be wealthy individuals,
and the wealth distribution in the United States
is well-known to be tilted in favor of whites. What that implies is we’re
setting up a design, or a test, that essentially
does the following. If I have a white investor who’s
going to invest in some bonds– some coming from a white
school, some coming from a black school– which one would the individual
be more likely to invest in? That notion is what we’re
going to study in this paper to figure out whether investors
have a preference, if you will, for a certain type
of school’s bonds. It’s going to turn
out the bond market’s going to be very, very
advantageous for us to rule out statistical
discrimination for reasons that you’ll see in a moment. So in order to assess
race discrimination in a bond market, what
we have to do first is have a quick primer
on how bonds work. So imagine a school
wants to build a dorm. Let’s suppose that they’re
going to issue a $30 million bond to build this dorm. So think of a bond
as a piece of paper that just entitles the
holder of that piece of paper to a stream of cash payments. So for example,
if the $30 million bond pays a 5%
annual coupon, and it has a maturity of 20
years, what that means is that piece of
paper in your hand entitles you to a $1.5 million
payment 20 times once a year for 20 years. In real life, one
person probably won’t go buy that
entire $30 million bond. Many, many, many
individuals will buy pieces of that $30
million, and finding those individual investors
to hold that investment is not so easy for schools. So they want to build a dorm. They’re not really good
at finding investors to buy their bonds. In comes the underwriter
to fix that problem. So there’s a service that
underwriters provide, like big investment banks. They will come in and buy
the entire $30 million block from the school. And they say, don’t worry. I’ll go find buyers. I’ll buy the whole
block from you, and I will turn around
and sell it to buyers. I’ll find them on my own,
and all you have to do is pay me a fee. So the schools will
essentially pay the underwriter to go find individual
investors, so the schools don’t have to worry about it. That cost, the amount that the
underwriter charges the school, is called a gross spread. So our story in this paper,
what we’re going to investigate is whether gross spreads in
some sense, are higher for HBCUs than non-HBCUs. And the idea here is as follows. If it’s harder to find a buyer
because the average buyer is a wealthy white
individual, you might have to search
much harder to find a willing buyer for an HBCU
relative to a non-HBCU. If that’s true,
the underwriter has to work harder and in
turn charge the HBCU more to place the bonds in
the hands of investors. So I know what you’re thinking. You’re thinking, this is
not race discrimination. It’s probably statistical
discrimination. Why do you say that? HBCUs are well known to be
of poorer financial health than non-HBCUs, and
so you’re saying it’s not that investors
don’t like HBCUs because they’re black. It’s because the HBCUs
are less likely to make the payments on
the bond, and that has nothing to do with racism. That would be an argument about
statistical discrimination. And I agree. That’s a potential problem. It may well be that
people just don’t want to hold bonds that
aren’t going to pay them back. And it will be much harder
to find such a buyer. There’s two novelties
in the bond market that’s going to help
rule this out for us. Those are credit ratings. So credit rating agencies will
come in and rate the bonds, and they will say, here’s
your credit rating. So if your AAA rated, that means
that the likelihood that you’re not going to pay the bond back,
make the bond payments on time and in full, is very, very low. Well, schools have
these bond ratings, and so we could compare
to schools, for example, that have AAA bond ratings. So it’s very unlikely that
the ability to not pay you back as an investor is
due to the type of school because we can actually
see a measure that tells us whether you’re likely
to pay back or not. The other novelty is
called bond insurance. And so some schools
say, you know, I know you’re worried
about the fact that I might not pay you back. So they’ll go pay an insurance
company that will make the bond payment in the event
that the school cannot. And so these two notions
about not getting paid back, while plausible on the
surface, are ruled out, or we’re able to rule
them out quite forcefully in our research setting by
exploiting the municipal bond market. So let me give you
an example of what our paper is all about in
terms of making an inference. So let’s just
compare two schools– Alabama State. That’s an HBCU founded in
1867 in Montgomery, Alabama. And let’s take another
university, the University of North Alabama. That’s a non-HBCU founded in
1830 in Florence, Alabama. So we have two schools
that are in the same state. They’re both public
schools, about the same size enrollments, and what
we’re going to do here is essentially
compare a bond that comes from Alabama
State and a bond that comes from North Alabama. They’ll have similar sized
bond issuances, similar credit ratings, similar insurance
companies insuring the bond payments, and so
we can pretty much rule out all these other effects
related to the ability to pay. In our prediction,
of course, here would be gross spread for
Alabama State will be higher than the University
of North Alabama. If we find a difference,
it’s hard to say it’s due to statistical
discrimination pertaining to the inability of one
school to pay over another. So here are our results
for our paper overall. We get a 23-year sample period– 1988 to 2010. It’s about $150
billion worth of bonds that were issued by
colleges and universities. What we find is, on average,
that HBCUs pay about 20% higher gross spreads
than non-HBCUs. In terms of basis points,
that’s about 17 basis points. So you might say,
well, you know, is this really plausibly
race discrimination? I agree with you. You’ve probably
ruled out the fact that it’s differences
in credit risk, or the ability of one school to
pay better than another school. But there are still
reasons to be cautious. The municipal bond
market is $3.7 trillion. That’s how big it is. There are a lot of buyers,
a lot of active individuals, participating in this market. Typically, we think in
economics that large markets will remove the effects
of race discrimination because if you can’t
find one buyer, there’s another one waiting in
line who is willing to transact in the marketplace. The other issue is that
in the bond market, unlike the labor market,
where an employer is worrying about interacting with a black
individual, where they just don’t even want to be around
them from a race discrimination standpoint, there is
no real interaction between the investor
and the school. It’s just bond
payments showing up. There’s nothing more to it. And so you might be
cautious to suggest that our results
so far are purely due to race discrimination. We’ve ruled out other things. It’s unclear what will be left
other than race discrimination, but perhaps we can do better. So here’s what we do to
rule in race discrimination. When we talk to bond traders
about this issue, what they say is it’s hard to find a buyer
for HBCUs because of state tax benefits. And what that means
is a municipal bond carries with it a state tax
exemption, meaning that if you live in the state of Alabama,
and you buy a municipal bond from either Alabama
State or North Alabama, you will get a tax exemption. It’s beneficial for
you to buy bonds from the same state
in which you reside. You wouldn’t get a tax
benefit from, for example, buying a Duke University bond
or University of North Carolina bond if you live in Alabama. So what happens is
schools are what we call geographically captive. It means the marketplace for
them is limited by their state boundaries. The individuals who really
want to buy the bond are going to be
in the same state. It doesn’t mean that outsiders
can’t buy those bonds, but they’re just not
incentivized to based on state tax benefits. It turns out that’s a
great feature for us in terms of trying to rule
in race discrimination. This is because we can
score individual states based on racial animus. So there are scores out there. For example, what was
the white vote share for Barack Obama in 2008
versus John Kerry in 2004? We can also look at the
number of racist tweets that happened in each state
when Barack Obama was elected. We can look at a
state’s opposition to affirmative
action, etc., etc. There are many metrics you
can use, and what we do is we score all the states
across many metrics, and then we sort them. And we figure out which
are the most racist. And so what we find is
in the far deep south, the most racist states turn
out to be Alabama, Mississippi, and Louisiana. So what we conjecture is that
if our results are really due to racism, then
the effects should be much stronger in these
high-racism areas than in non-high-racism areas. So here’s the example,
going back to our schools in Alabama– Alabama State and North Alabama. What we found before
is overall on average, that there was about 17
basis points more that had to be paid by HBCUs
relative to non-HBCUs. Our prediction
would be that this would be much stronger
in high-racism states and much lower in
low-racism states. So let’s add in two
new universities. So instead of just considering
Alabama State and North Alabama, who are in
the far deep south, let’s also consider
American University and Howard University in
Washington D.C. Howard University is an HBCU. American is not. What we would predict is
that if this is due to race discrimination, the
difference between schools– HBCU versus non-HBCU–
should be much, much larger in the far deep south
than it is, for example, in Washington D.C. That’s
precisely what we find. So when you’re doing the
comparison of American and Howard University in D.C.,
the difference between them is only 11 basis points. But when you go down
to the deep south, and you focus only on schools
like Alabama State versus North Alabama, what you find is a
difference of 29 basis points. So it’s about three times larger
when you get to the deep south. It’s hard to explain that
result with anything else other than race discrimination. And so that is an important
finding in our paper that helps us rule in that
what we’re documenting is due to racial animus in
the municipal bond market. So you might say,
OK, that’s great. How do we fix this? Well, the reason
that this occurs is because of this
geographic captivity. If we could somehow expand
the market of individuals who would be willing
to buy the bond, then we perhaps could break
the setting, or the race discrimination, that’s
inflicted on HBCUs. One way to do that is to make
the bonds triple tax exempt. Now what that means is it’s
exempt from tax at all levels, including the state level. If you designated HBCU
bonds as triple tax exempt, then anyone in the
United States would get the tax exemption
at the state level, not just the investors
in the same state. That would break the captivity. So we forwarded this
as a possible solution to this problem,
and it was picked up and written into a bill– H.R. 6048, titled the HBCU
Investment Expansion Act. And that was introduced to
the House of Representatives back in September of 2016. That same bill was referenced
recently in late February of 2017 by the Congressional
Black Caucus in their letter to the Trump
administration, suggesting that they get behind this
bill as one of the many ways to help HBCUs as they try
to finance their needs as educational institutions. So to conclude, what
we’ve learned today is race discrimination obviously
remains an important issue in our society. It can even exist
in large capital markets such as the
municipal bond market. Separating race discrimination
from statistical discrimination is quite difficult.
And if you’re going to make progress on
solving race discrimination issues, you have to be
very careful to make sure you can rule out
statistical discrimination.