P29
ABSTRACT
The current crisis in the mortgage securitization industry highlights significant failures in our models of how markets work and our political will, organizational capability, and ideological desire to intervene in markets. This article shows that one of the main sources of failure has been the lack of a coherent understanding of how these markets came into existence, how tactics and strategies of the principal firms in these markets have evolved over time, and how we ended up with the economic collapse of the main firms. It seeks to provide some insight into these processes by compiling both historical and quantitative data on the emergence and spread of these tactics across the largest investment banks and their principal competitors from the mortgage origination industry. It ends by offering some policy proscriptions based on the analysis.
P30
INTRODUCTION
The market to sell home mortgages as ‘‘securities,’’ what are called mortgage-backed securities (hereafter MBS) is at the core of the financial crisis. It is clear that the financial community, the policy analyst community, and government officials in charge of the Federal Reserve and Treasury Department mostly underestimated what was happening and optimistically thought at many points that the crisis could be contained. The crisis apparently even came as a great surprise to the people who ran the
companies that produced the meltdown. The president of Bear Stearns, James Cayne, for example, has argued that the company did nothing wrong and did not take risks they did not understand, even as his company was headed towards bankruptcy (Cohan, 2009). Even more disturbing, the actions of the regulators betrayed a deep ignorance of what was actually going on. Regulators mostly trusted in market actors to self-regulate. As a result, some regulators were convinced that there never was a mortgage bubble, because markets operated efficiently ( Kaufman, 2009, p. 235). Given regulators placed so much faith in the standard rational actor model, it is useful to ask, how well did it characterize the economic behavior of banks? The ongoing crisis suggests, ‘‘not very well.’’ But how that model fails is a complex question and one that we shed some light on in the research in this article. Our basic premise is that what was most lacking is a sense that there existed a system that structured the MBS business. We use the market model proposed by Fligstein (1996, 2001) to begin this reconstruction. This article tries to investigate how this system of relationships between regulators, mortgage originators, mortgage packagers (both commercial and investment banks), ratings agencies, and the holders of such bonds evolved through the rise and fall of the MBS market. We seek to uncover the inner logic of the mortgage business over time to understand why actors did not behave
‘‘rationally’’ in the narrow economic sense which regulators expected. We do so first by setting the basic facts of the rise and fall of the MBS market straight. We show that the success of the residential real estate market and the MBS business caused a rapid expansion of that business from 1993 until 2003. This rapid expansion brought the biggest banks into the market in a big way. Over time, the largest banks increasingly based their business on making fees off of selling mortgages to individuals,
packaging those mortgages into bonds, selling many of those bonds to investors, and perhaps most interestingly, retaining a significant portion to profit from the lucrative spreads on high yield bonds funded through capital acquired at low interest rates.
P31
Another important set of actors in the MBS field were the ratings
agencies. We document the complicity of the rating agencies in legitimating
the massive influx of subprime MBS which flooded the market from 2004 to
2007. We show that the three ratings companies inflated their ratings of all
MBS from 2003 to 2007, thereby potentially misleading investors as to the
underlying riskiness of the bonds. But, here too there was a system at work.
Ratings agencies were paid by the packagers of MBS. If they were not
cooperative in rating bonds highly, the banks would take their business
elsewhere. Given that the MBS business was highly concentrated, the MBS
issuers had power over the ratings companies. The ratings companies chose
to be active and willing participants in the rating of MBS in a generous
fashion, and they profited handsomely from this business up through 2007.
The final and perhaps most important player in the mortgage securitization
field is the federal government. The role of the government within the
mortgage securitization field has been two-fold. First, contrary to the popular
notion that government either inhibits the natural workings of the market or
alternatively stays away and lets the free market do its work, the MBS market
is a case where regulators and banks have co-evolved in a way that has put the
government in the position of creating the market, underwriting it, working to
make it expand, and allowing banks to take over key parts of the business,
thereby supporting the expansion of the production of new financial products.
Government created the first MBS in the late 1960s, and the private market
for MBS has grown up under and with the cooperation of the government-
sponsored enterprises (hereafter GSE) for the past 20 years. The GSEs
remained the dominant player in the mortgage market over this period.
Indeed, the government had to coax the banks into the MBS business.
Democratic and Republican presidents and Congresses pulled the banks into
the business by providing financial reforms that expanded the MBS market
and worked to allow the largest banks to do anything they wanted.
Second, despite being a central part of the mortgage-securitization
system, government regulators did not view it as a system. Instead,
paradoxically, they bought into the rational actor/financial model which
formed the dominant cultural-institutional infrastructure in which they were
embedded. Allen Greenspan, Chairman of the Federal Reserve, championed
a form of market fundamentalism in which he acted as if the government
had nothing to do with the creation of the market and should do little or
nothing to manage the riskiness of the market. Instead he assumed that the
banks would act in their own self-interest to protect their investments.
This disjuncture between the regulators’ actual role and their avowed role
led to a type of regulatory capture. As the activities of banks expanded and
P32
they invented more and more financial products, bankers were consistently
able to convince regulators and the executive and legislative branch of the
government to stay away from regulating the market. Their basic arguments
were that the financial innovations in the market were producing robust
economic growth and that at their core these innovations made a market that
was able to control its own risks. It is important to remember that the market
for MBS expanded from 1992 to 2007 which gave these arguments more
credence. In the run up to the crisis, Allen Greenspan denied that there was a
housing bubble. He bought into the argument that these products were
expanding rapidly because they were successfully controlling risk. As a result,
he prevented new regulations of financial products such as collateralized debt
obligations (CDO) and credit default swaps (CDS), and prevented regulators
from using existing regulations to stop the sale of subprime mortgages.
Looking back on what happened, it is hard not to conclude that all attempts
to regulate these markets were thwarted by the regulators themselves, who
bought into the arguments of the banks. This misunderstanding of the nature
of the market by regulators is central to the crisis.
We argue that the real beginnings of crisis emerged after 2003 when
unconventional nonprime markets rose from approximately 10% of the
market in 2003 to almost 70% of the market at their peak in 2007. This
transformation in the MBS field’s product mix was as rapid and dramatic as
the field’s overall growth.We show how this shift was sparked by a steep but
little-discussed decline in the supply of prime mortgages after the 2003
refinancing boom. In order to maintain high profits and volume, originators
and conduits aggressively pursued new sources of raw mortgages in various
‘‘unconventional’’ markets such as B/C, Alt-A, and home equity loans
(HELs). The result was that the subprime sector of the market, which had
formerly been marginal in size and dominated by specialist firms, quickly
became a linchpin of the financial sector.
In presenting this history, we empirically dispel several conventional
wisdoms that have taken hold. For example, one of the ‘‘facts’’ that is
already taken for granted in economic analyses of the meltdown is that the
banks that originated mortgages and packaged mortgage securitization
never held onto the securities themselves. It is asserted that this perverse
incentive made them more likely to take on larger risks. We show that
contrary to this view, every large originator and packager of mortgages held
onto substantial numbers of MBS and this increased dramatically after
2001. Simply put, they believed that they could control the amount of risk
they held. The result is that most of these firms are either out of business,
merged into larger banks, or owned by of the federal government. Another
p33
commonly voiced myth about the MBS market is that it was highly
dispersed, with too many players to control any facet of the market. On the
contrary, we show that over time all of the main markets connected to
MBS – the originators, the packagers, the wholesalers, the servicers, and the
rating companies – became not only larger, but more concentrated. By
the end, in every facet of the industry five firms controlled at least 40% of
the market (and in some cases closer to 90%). Separate market niches also
increasingly condensed around the same dominant firms. As a result, the
mortgage field was not an anonymous market scattered across the country,
but instead consisted of a few large firms. This concentration meant that
firms very much collaborated and competed in these various markets. Firms
would join together in MBS packages and assume different roles with each
other. This meant that they had a great deal of knowledge of the market and
what the others’ moves were.
Finally, we present data challenging the conventional wisdom that the
complexity and opacity of financial instruments – particularly CDOs – was
a chief contributor to the MBS bubble and subsequent meltdown.
A mortgage-related CDO (sometimes called ABS CDO) is a securitization
of existing MBS tranches. Although the complexity of pricing a CDO can be
very difficult due to the disparate income streams from which it is
constituted, at root it is simply a claim on MBS tranches, which are claims
on income from mortgage payments made by homebuyers. Although
investors lost huge amounts of money on CDOs, we provide evidence based
on ratings history data suggesting that CDOs were no more risky than the
subprime MBS from which they were constituted. CDOs did not exhibit
compositional ratings inflation (which indicates that their complexity did
not contribute to the rampant grade inflation by which subprime
mortgages were increasingly engineered into AAA tranches). Nor were they
downgraded at higher rates than more conventionally structured securities.
These findings suggest that evolution of financial instruments offers a less
germane explanatory axis than many scholars have suggested (e.g., Skreta &
Veldkamp, 2009).
THE HISTORY OF THE MORTGAGE
SECURITIZATION MARKET
Housing is at the core of the American economy. Indeed, owning a house
has been one of the linchpins of the American dream. The purchase of a
P34
house is the largest expense that most citizens ever make. Public policy in the
post-war era has recognized this as an admirable goal and governments of
all political persuasions have worked to make ownership a reality.
To understand the current crisis, one needs to step back and understand
the co-evolution of home financing between the public and private sectors
since the 1960s. Fig. 1 shows how most people got their mortgages until the
1980s. Individuals would find a house. They would go to their local bank
(most likely a savings and loan bank) and apply for a mortgage. The bank
would agree to lend the funds and then hold onto the mortgage until it was
paid off or the house was sold. During this historical period, the largest
holders of mortgage debt were private banks.
Fig. 2 describes the way in which the entire mortgage industry is currently
organized. Now, the borrower goes to a lending company (frequently a
bank, but not exclusively) that is called an ‘‘originator’’ because they make
the initial loan. Unlike the original savings and loan banks, these companies
do not want to hold on to the mortgages they sell, but instead want to sell
them off to others. Their business basically is organized to make fees off of
buying mortgages. If they hold onto the mortgages, then they are unable to
lend money again and their ability to generate fees goes away. So, they turn
around and sell the mortgages, thereby recapturing their capital, and move
back into the market to lend.
The mortgages are then packaged together into something called a special
purpose vehicle by underwriters who are either government sponsored
enterprises, investment banks, or commercial banks. This vehicle turns the
mortgages into an asset that pays a fixed rate of return based on the interest
rates being paid by the person who buys the house. These special purpose
vehicles divide up the mortgages into what are called ‘‘tranches.’’ Here, the
mortgages are separately rated by bond agencies in terms of their riskiness.
These bonds that contain these tranches are called CDOs. In this way,
investors can buy riskier bonds that pay a higher rate of return or less risky
bonds that pay a lower rate of return. The special purpose vehicles are
managed by firms called servicers who collect the monthly mortgage
payments and disburse them to the bond holder.
P35
Circa 1975, mortgages were highly geographically dispersed and held by
local banks. Now, after they are issued, they migrate to a few square miles of
Manhattan where in the offices of the major banks and GSEs they are
packaged into special purpose vehicles. They then are redispersed to
investors all over the world (although they are serviced from a few
locations). Investors are a heterogeneous group. The largest investors in
P36
these securities are the GSEs who hold onto lots of MBS. But, MBS are held
by banks, mutual funds, and private investors here and around the world.
The interesting question is how did we move from a world where the local
buyer went to their local bank to get a loan to one where most of the
mortgages in the United States are now packaged into MBS and sold into a
broad national and international market?
It will surprise most readers that the origins of the MBS and the complex
financial structure we just presented were not invented by the financial
wizards of Wall Street, but instead were invented by the Federal
government. It is probably even more surprising that this set of inventions
dates back to the 1960s. The federal government has been involved in the
mortgage market to some degree since at the least the 1930s. But the roots of
the modern industry begin in the 1960s.
Quinn (2008) argues that the idea to create MBS began during the
administration of President Lyndon Johnson. The Democratic Congress
and President had three goals: to increase the housing stock for the baby
boomer generation, to increase the rate of homeownership, and to help
lower income people to afford housing. Quinn (2008) shows that the
Johnson Administration did not think the fragmented savings and loan
industry was in the position to provide enough credit to rapidly expand the
housing market. But, federal officials interested in expanding homeowner-
ship were also worried about the size of the budget deficit. Because of the
Vietnam War and the Great Society expansion of Medicaid, Medicare,
and other social benefits, the government was running large and persistent
debts. An expensive housing program where the government provided
funds for mortgages would add to the deficit, because the government would
have to borrow money for the mortgages and hold those mortgages for up
to 30 years.
If the government was going to stimulate the housing market, the
Johnson Administration would need to do it in such a way as to not add to
the federal deficit. This caused them to reorganize the Federal National
Mortgage Association (Fannie Mae) as a quasi-private organization, called
a GSE, to lend money and hold mortgages. They also created a new entity,
the Federal Home Loan Mortgage Corporation (Freddie Mac) to compete
with Fannie Mae as well as a government-owned corporation to insure those
mortgages against risk of default, the Government National Mortgage
Association (Ginnie Mae).
But taking these mortgage granting entities private was not the only
innovation of the Johnson Administration. The government also pioneered
the creation of MBS (Sellon & VanNahmen, 1988). The government, even in
P37
the GSEs, did not want to ultimately hold the mortgages because this would
limit how many mortgages it could originate. Instead, it wanted to use its
capital to fund the mortgages and then offer the mortgages to investors as
bonds. It did so by offering and guaranteeing the first modern MBS. These
bonds could then be sold directly to investors and were sold by the GSEs or
through investment banks (Barmat, 1990). The first MBS was issued on
April 24, 1970 by Ginnie Mae (Wall Street Journal, 1970).
The private MBS market barely grew in the 1970s. There were several
issues. The savings and loan industry continued to have control over the
bulk of the mortgage market where they took deposits, lent money, and held
onto mortgages. But potential buyers of mortgage bonds were skeptical of
buying MBS because of prepayment risk. The problem was that if you
bought such a bond, people might pre-pay the mortgage before the end of
the mortgage term and bond holders would get their money back before
they made much of a profit. This was made worse by the fact that mortgage
holders were more likely to re-finance houses when interest rates were falling
thus leaving bondholders with money to invest at interest rates lower than
the original mortgages (Kendall, 1996).
This problem was ultimately solved through joint cooperation between the
GSEs and the investment banks. They created the system of ‘‘tranching,’’
described earlier, to let investors decide which level of risk of prepayment
they wanted (Brendsel, 1996). But there were also legal and regulatory issues
involved in the packaging of bonds (Quinn, 2008; Ranieri, 1996). The most
important was the problem of turning a mortgage into a security. The issue
of a loan originator selling the mortgage into a pool of mortgages required
changing the tax laws. The Tax Reform Act of 1986 cleared the way to the
rapid expansion of the MBS market.
The demise of the savings and loan banks was a fortuitous collapse
that hastened the growth of the MBS market. The general economic
crisis of the 1970–1980s produced very high interest rates. Savings and loan
banks relied on individual deposits for most of their funds. The regulation
known as Regulation Q fixed the rate that savings and loan banks could
pay on these deposits. Savers began to flee those accounts and the savings
and loan industry faced the crisis that they could not raise enough money
to make new loans. Moreover, they were holding on to a large number of
mortgages that were priced at very low interest rates. Congress responded
by passing the Garn-St. Germain Act. They repealed Regulation Q and
allowed the banks to pay whatever interest rate they chose. They also
allowed the banks to make riskier investments while still guaranteeing very
large deposits.
P38
The banks responded in several ways. First, they began to sell their
mortgage holdings at a great loss to raise capital. These mortgages were
repackaged into MBS by primarily Solomon Brothers ( Lewis, 1990). They
also began to pay high interest on government guaranteed bank accounts.
They then made very risky investments including many in commercial real
estate which helped create a commercial real estate bubble. This caused their
ultimate demise (Barth, 2004).
The basic outlines of the mortgage securitization industry producing
mortgage deals structured like those illustrated in Fig. 2 was in place by the
late 1980s. The MBS market grew enormously as the savings and loan banks
collapsed. Fig. 3 shows how the GSEs took up the slack from the savings
and loan banks during the 1980s. In 1980, the GSEs had only issued $200
billion of mortgages. This grew steadily to a peak of $4 trillion in 2006.
Fig. 4 shows the dramatic decline of the savings and loans and the rise of
the government-backed mortgage market. As late as 1978, the savings and
loan banks held almost 60% of the mortgage debt in the United States. But
beginning in the late 1970s their market share plummeted. By 1990, less than
15% of mortgages were held by savings and loans. The sharpest rise in
mortgage debt was now being packaged by GSEs into MBS pools.
Approximately 50% of the mortgages were in these pools. If one adds the
P39
P40
10% or so of the mortgages that were being held by GSEs, the GSEs were
involved in 60% of U.S. mortgages.
Beginning in the early 1990s, a financial revolution began among banks.
It is at this moment that the banks became more aggressive in all financial
markets. This revolution began with the idea of securitization, but quickly
created more complex and financial instruments to create new ways to invest
and control risk. It is to the story of the past 15 years that we now turn.
THE RISE OF THE MBS MARKET, 1993–2003
The GSEs remained the central actors in the mortgage securitization market
through the 1990s. But by the early 1990s, the people who worked for
investment banks came to see that mortgages could be profitably packaged
and sold as bonds in the same way as their other products ( Jungman, 1996).
Further, the potential size of these markets was huge. The market for
mortgages in the United States increased from $458 billion in 1990 to nearly
$4 trillion at its peak in 2003. Most of these mortgages were packaged into
MBS, and although most MBS were still sponsored by the GSEs,
commercial or investment banks played an increasingly prominent role in
putting these packages together and helping the government sell them.
As we elaborate further below, the big banks also created a massive market
segment for those unconventional mortgages that the GSEs would not back,
especially after 2003.
The growth of MBS during the 1990s occurred alongside (and contributed
to) an unprecedented period of growth in the housing market. In most of
our accounts of the crash, commentators rarely spend much time discussing
the rapid upsurge in housing sales first from 1990 to 1998 (almost tripling in
size in 8 years) and then in 2000–2003 (where the market rose almost 400%
between 2000 and 2003). Instead, they treat the housing market growth as a
background condition and focus their attention on the end of the bubble
from 2006 to 2008 and the role of such things as lack of regulation, the
growth of the use of financial instruments, and the greed and bonus pay of
bankers that caused them to engage in riskier and riskier ventures. But all of
these factors were present before the housing boom and they contributed as
much to the increase in mortgages over a 15-year period as they did to the
decline in 2006–2008.
It is useful to document the growth of the mortgage origination market
since the early 1990s. Fig. 5 presents data on total loan originations from
1990 to 2008. It also breaks down the loan types into various products. The
P41
American mortgage market was approximately $500 billion in 1990. During
the 1990s, it went up to nearly $1 trillion in 1993 and peaked in 1998 at
around $1.5 trillion. In 2000, it stood at $1 trillion a year. The real surge in
the mortgage market began in 2001 (the year of the stock market crash).
From 2000 to 2004, residential originations the United States climbed from
approximately $1 trillion to almost $4 trillion.
The second bubble had different causes. The year 2000 will be
remembered as the year of the crash in ‘‘dotcom’’ stocks. As that crash
began, the Federal Reserve, in response to the crash essentially lowered
interest rates to zero. Their actions were met by similar actions in central
banks around the world. The Federal Reserve did this to make sure that
there was substantial credit in the economy and that lending would
continue. But, the unintended effect of lowering interest rates so far was
that it encouraged the housing bubble in the United States. The rapid rise of
that bubble was astonishing: mortgage origination rose 400% in 4 years.
One of the critical facts, that the Federal Reserve knew this and did not
take any actions to stop it, is to be explained. Alan Greenspan has testified
that he did not believe this was a bubble because housing prices are a local
P42
affair in the United States. Hence, a house in Boston is not substitutable for
a house in Atlanta.
But what he chose to ignore in his analysis was that the cheap interest
rates across the country and strong demand from investors in the secondary
market encouraged banks to loan as much mortgage debt as they could.
Bankers could borrow money at approximately 1% and loan it at 5–7%.
This caused bankers to search far and wide for any local real estate market
where they could make loans. Where housing was scarce and population
was growing, prices rose and this encouraged banks to focus on those
markets. It is not surprising that Arizona, Florida, Nevada, and parts of
California were the ground zero of the housing crisis. In other words, the
housing bubble’s roots were systemic, and they were driven by the
governmental policies that encouraged homeownership and securitization.
The other major factor in both the first and second housing booms was
the proliferation of mortgage securitization tools and the increased
P43
participation of the bigger banks in these processes. The large banks entered
these markets with the goal of growing them and growing their market share
in them. They aggressively used securitization tools as a way to raise money
for mortgages and a way to sell them. Tables 1 and 2 show how the top
players in various parts of the mortgage market shifted over time as those
markets grew. Table 1 shows that in 1996, the largest players in the
mortgage market were mostly either mortgage specialists like Countrywide
or NW Mortgage or regional commercial banks like Fleet Financial or
PNC. But by the end of the second bubble, the identities of the largest loan
originators had changed to large national bank holding companies like
Wells Fargo, Citibank, and Bank of America. Countrywide had turned itself
into a national bank, as had Chase, Wachovia, and Washington Mutual.
These large players grew larger as the national market expanded.
Table 2 shows a similar process for the packagers of MBS, which in
the industry jargon are called ‘‘conduits.’’ In 1996, both Solomon Brothers
(now part of Citibank) and Merrill Lynch were on the list. But the packagers
of MBS were generally smaller firms that were more narrowly focused
financial firms than investment banks. In 2007, the list of mortgage conduits
was dominated by the investment banks like Lehman Brothers, Bear Stearns,
JP Morgan, Morgan Stanley, Deustche Bank, and Merrill Lynch. Note that
now several of the largest originators of mortgages – banks like Countrywide,
Washington Mutual, Indy Mac, and Wells Fargo – have taken advantage of
the changes in Glass-Steagall. They now not only make mortgage loans, but
also act as packagers of those loans. These tables imply that the major
beneficiaries and, indeed, the drivers of the growth in the mortgage origination
and the MBS business, were the largest investment and commercial banks.
One of the least remarked features of the mortgage origination market is
the degree to which it became concentrated over its history. To our
knowledge, the only other scholarly commentator who presents similar data
is Kaufman (2009). Table 1 shows that the market share of the top 5
originators stood at 16.3% in 1996, a remarkably low concentration ratio.
But in 2007, the top 5 originators accounted for 52.5% of a much larger
market. Fig. 6 shows how the concentration of lenders changed from 1990
to 2008. In 1990, the 25 largest lenders accounted for less than 30% of the
mortgage market. This rose steadily during the 1990s and by 2007, the top
25 originators controlled 90% of the market. This data directly contradicts
one of the standard tropes about the mortgage securitization business: that
the market for originators of loans was remarkably unconcentrated.
Although it is true that the number of independent mortgage grew in
number over this period, the origination market was dominated increasingly
by the big national banks. They either negotiated exclusive contracts with
so-called shelf originators or acquired formerly independent mortgage
originators to assure themselves of a supply of mortgages.
Table 2 shows a similar process for the conduit market. In 1996, the top 5
producers held a 24.5%market share whereas in 2007 this rose to 41%. If one
looks at the top 10 conduits in 2007, the total is 71%. So, there was not just a
rapid growth in the size of these markets, but also a rapid concentration of
activities among a few larger and more nationally oriented banks.
It is important to understand why the same large commercial and
investment banks were at the core of each MBS market segment. Investment
banks historically were involved in two related businesses. They helped
package and sell corporate and government bonds for investors. They also
helped firms issue stock and advised on mergers and acquisitions. These
activities placed them in the center of the financial system where they acted
as financial intermediaries. The Glass–Steagall Act forced banks to choose
P45
whether or not they wanted to be investment banks or commercial banks,
and enforced restrictions on the activities of each. During the past 15 years,
policymakers and bankers have worked to have this barrier broken down.
One of the reasons for this is because of the MBS business. As that
business became larger, commercial banks wanted to be able to sell loans
(be originators), package loans (be conduits), and hold onto loans (be
investors). Firms like Bank of America and Citibank saw that fees for
putting together these packages ended up with investment banks, and they
too wanted access to this lucrative business. Similarly, investment banks
looked to secure upstream supply of mortgages by taking over originators
so they could apply their deal structuring expertise on a larger scale. Both
investment and commercial banks got their wish. The Glass–Steagall Act
was rescinded in 1999 and banks were allowed to be in any business they
chose. This removal of market demarcations allowed large firms to integrate
and fully participate in every part of the MBS market.
P46
The repackaging of mortgages into bonds became the largest fee
generation business for many investment banks including Solomon
Brothers, Lehman Bros., Bear Stearns, Merrill Lynch, Morgan Stanley,
and Goldman Sachs. Of course, commercial banks and bank holding
companies like Bank of America, Wells Fargo, Citibank, and Countrywide
Financial also became deeply involved in all stages of the market, from
origination to packaging, to servicing.
The major firms employed strategies to profit from MBS in multiple ways
simultaneously earning money both from fees and income on retained MBS
assets. Bank originators could either use their own capital or cheap borrowed
capital to make loans to home buyers (Ashcroft & Schuermann, 2008 take up
this story). Then, they could turn around and sell those loans to conduits.
If they used someone else’s money (borrowed at say 1–2%), then they could
essentially do the entire transaction with very low cost and relatively high fees.
Conduit banks could also borrow money cheaply. They would then buy the
mortgages, package them, and sell them to investors. But, beginning
sometime around 2002, both commercial banks and investment banks began
to realize that they could borrow money for 1–2%, create MBS, and hold
onto the MBS which might pay as much as 6–7% in interest. This allowed
them to make a profit using other people’s money without risking their own
capital. The low interest rates in the United States and the world encouraged
banks of all kinds to make as many loans as they could and to hold onto MBS
because they were earning money on borrowed money.
But it was not just conduit banks that figured out that they could borrow
money cheaply and buy and hold MBS. Fig. 7 documents who owned MBS.
The conduits increased their holding of MBS from approximately $35
billion in 2002 to $175 billion in 2007, a more than 400% increase. But, at
the same time, commercial banks increased their holding from $650 billion
in 2002 to $1.1 trillion in 2007. Other private investors (including hedge
funds) increased their ownership of MBS during this period from $25 billion
to $700 billion. Mutual fund operators began to buy MBS as well and went
from approximately $400 billion to nearly $850 billion. Most spectacularly,
foreigners increased their investment from $200 billion to $1.2 trillion. This
shows that the world wide appetite for what were thought to be safe but
high yielding investments was one of the drivers of the crisis. The demand
for ‘‘safe’’ high yield bonds was not just being generated from the loose U.S.
monetary policy but from capital supply around the world.
The massive growth of MBS led to the development of two sorts of
secondary markets (Barmat, 1990). First, the CDO market allowed for
further pooling and tranching of risk by resecuritizing existing MBS
P47
tranches or other income streams including such things as corporate debt
and aircraft leases. Second, the credit default swap (CDS) market allowed
firms to insure the risks they held on CDO and other financial instruments
(see Tett, 2008 for a lucid discussion of the evolution of CDS).
As the mortgage securitization market has been negatively affected by the
collapse of the subprime lending market, this has put pressure on both the
CDO market and the CDS market. In the CDS market, losses have been
higher than expected. Since many firms bought CDS to protect themselves
against such risks, this has meant that holders of the CDS have come under
pressure to pay off their obligations. One of the main suppliers of CDS for
MBS was AIG, and their exposure in this market was the main cause of
their takeover by the government.
THE GROWTH OF SUBPRIME
By 2003, investors of all kinds – commercial banks, investment banks, hedge
funds, insurance companies, and other private investors – had figured out
how to use leverage by borrowing money cheaply to buy MBS. Investors
P48
who actually had cash – like pension funds, insurance companies, and
governments and banks around the world – were seeking out safe
investments that paid more than 1–2%, the going rate for government
debt. American mortgages seemed like a good bet. The underlying assets of
mortgages were houses and the MBS contained mortgages from all over the
country, thereby appearing to be diversified geographically. American
housing prices had risen steadily for as long as anyone could remember.
Finally, MBS were rated and it was possible to secure ‘‘AAA’’ rated bonds.
This made American mortgages seem like low risk, high yield investments.
Then, in 2004, the MBS market experienced a supply shock. The
reverberations from this event would lay the groundwork for the market
meltdown in 2008. As shown in Fig. 5, the supply of conventional mortgages
peaked in 2003 and began a rapid decline thereafter. Approximately $2.6
trillion worth of conventional or prime mortgages were bought in 2003 and
this dropped to $1.35 trillion, a drop of almost 50%, in 2004. The steep decline
in mortgage originations reflected neither weakness in the housing market nor
slackening demand from the secondary market. Rather, a saturated prime
market and an interest rate hike led to a significant drop off in the refinancings
which had driven the 2003 boom. So, while those who had money to buyMBS
were looking for product, those who were originating and packaging MBS
lacked enough to sell them. This meant that there was a huge incentive to
increase the number of mortgages. This incentive sent loan originators looking
for new mortgage markets to feed the securitization machine and led to the
rapid growth of the ‘‘subprime market,’’ i.e., the market to lend to people with
poor credit histories and little in down payments.
The rapidity with which firms gravitated toward subprime to make up for
the diminishing inventory of prime loans is remarkable. In 2001, the largest
conventional (prime, government-insured) originator did 91% of its
origination business in the conventional market. By 2005 the largest
conventional originator was doing less than half of its origination business
within the conventional sector.
It is useful to discuss Fig. 5 in more detail to understand the implications
of this transformation of the mortgage market. At the bottom of the graph
are home loans originated by the Federal Housing Administration (FHA)
and the Veteran’s Administration (VA). These were never a large part of
the total originated loans although they did increase slightly after 2001.
The largest parts of the market were conventional and conforming
mortgages. These are mortgages for people who put down 20% for their
house and whose loan value is not above a certain cutoff point where they
need to pay additional interest to protect the higher risk associated with a
P49
higher loan amount. The cutoff point was continuously raised as prices
increased. We can see that the bulk of the mortgage market from 1990 until
2003 consisted of these two categories of loans.
But beginning in 2003, we begin to see rapid increases in all forms of
nonconventional loans as banks began to search out customers. Home
equity loans refer to loans made against the value of the equity in a house.
These were frequently in the form of a line of credit or a second mortgage.
Alt-A and subprime mortgagees (sometimes called ‘‘B’’ and ‘‘C’’ mortgages
to denote their lower bond ratings) were people with poor credit history or
people who lacked the ability to make a large down payment (or sometimes
both). Jumbo loans have higher interest rates because the loan amount
exceeds a value set by the FHA each year.
In 2004, for the first time, these four categories of loans exceeded the
prime market or conventional market. In the peak of the mortgage craze in
2006, fully 70%of all loans that were made were unconventional mortgages.
This astounding change in the character of the mortgage market was noticed
by regulators and Congress. But, the Federal Reserve chose to ignore what
was going on. Alan Greenspan has famously testified before Congress that
the reason he did nothing to stop this rapid growth in unconventional
mortgages is that he did not believed that banks would not have made these
loans if they thought they were too risky. He is also on record as saying that
he clearly was mistaken on this point (Reuters, 2008).
To understand how fully the private sector banks were behind
the subprime crisis, it is useful to turn to Fig. 8. This figure contains the
amounts of MBS issuance by the type of mortgage from 1995 on for private
sector banks. During the 1990s and continuing through the first part of the
2000s, the conventional or prime MBS market was dominated by the GSEs.
But, beginning in 2001, the portion of the MBS market controlled by
‘‘nonagency’’ banks (i.e. private banks) rose dramatically. At the peak of the
subprime market, the private banks issued almost $1 trillion of unconven-
tional MBS each year. Subprime lending and subprime securitization had
long existed as a marginal niche market, but by 2005 this market had moved
to the center of the financial sector, massive in size and populated by the
biggest financial firms.
TURNING B/C MORTGAGES INTO AAA BONDS
Another key set of actors in legitimating this transformation were the rating
agencies. Although virtually all commentators agree over-inflated bond
P50
ratings were a central factor in producing the crisis, we currently know very
little about the sources and patterns of ratings inflation. What drove ratings
inflation? When did it begin? How was it related to the growth of subprime,
Alt-A, and other ‘‘nontraditional’’ sectors of the mortgage market? How
was it related to the increasingly complex structures issuers were using to
engineer securities? In this section, we provide an anatomy of MBS rating
dynamics and show how the patterns revealed by this anatomy offer
important insights into overarching explanations of what happened. 1
Fig. 9 chronicles the compositional shift in the initial ratings for
nonconforming mortgage MBS from 2003 to 2007, the core years of the
increases in the number of those forms of mortgages. Astonishingly, as the
period progressed, the percentage of issues receiving a AAA rating increased
from 15% to 42%. Almost 80% of the nonconforming MBS received an
A rating or above. So even as the number and size of these MBS increased,
their average ratings increased as well. Here we aggregate B/C, Alt-A, and
HEL MBS for the sake of space, but the same common trend toward
P51
fabricating increasingly prime-grade securities from nonprime mortgage
debt was virtually identical within each of these nontraditional asset classes.
This inflationary trend is all the more remarkable insofar as other evidence
suggests the credit composition of borrowers was moving in the opposite
direction during this period (Keys, Mukherjee, Seru, & Vig Vikrant, 2008).
One could argue that these ratings were legitimate. To show that
unconventional MBS were increasingly overrated, one would need to
examine how the ratings fared over time. A good measure of this is the
number of times that a bond is downgraded. Here the data indicate that
compositional inflation was coupled with considerable diminishment in
ratings quality/accuracy between 2003 and 2007, especially among the
most highly rated tranches. First, overrating became progressively more
widespread. Although approximately half of the B/C, Alt-A, and HEL
securities from the 2004 vintage were subjected to downgrades of at least
two notches after the meltdown, over 80% of the securities issued in 2006
were significantly downgraded. This pattern is repeated more or less
equivalently across each of the individual asset categories, indicating a
secular trend toward more extensive MBS overrating.
P52
Second, overrating also became progressively more intensive between
2003 and 2007. Fig. 10 shows the average magnitude of subsequent ratings
downgrades by vintage of the bonds through May 2009. The graph shows
that all bonds were subject to significant downgrades after the meltdown.
But it shows that bonds issued in 2005–2007, the height of the subprime
market, were particularly downgraded. A bond issued in 2006, for example,
was downgraded on average 4.6 steps whereas a bond issued in 2002 was
only downgraded 2.8 steps. Not only were bonds issued after 2004 more
highly rated, but they were also clearly more overrated as evidenced by
the large downgrades they took as the market deteriorated. Fig. 11 breaks
this out by the type of bond. Again we see a tendency for the bond
downgrades to affect all types of MBS. But generally, the riskier B/C, HEL,
and Alt-A-backed bonds experienced the most severe downgrades.
CAUSES OF THE CRISIS
We would argue that the proximate causes of the crisis can be found in two
shifts in the structure of the mortgage securitization field. First, the easy
credit available to all forms of financial investors after 2000 meant that
P53
money could be made by borrowing money at a low interest rate and then
turning around and buying MBS. This process of leveraging was the core
strategy of banks and many other financial institutions. Investors worldwide
who were not leveraged were also searching for higher, but safe returns and
American mortgages looked good to them. These strategies brought all of
the major banks aggressively into mortgage securitization, and made
mortgage securitization into a linchpin of the financial sector.
But, the second cause (which is not well understood) is as important as the
first. By 2004, there were simply not enough prime or conventional
mortgages left in the United States to package into MBS. This brought
about a search for new customers, many of whom had less money to put
down or worse credit. It was the dramatic growth of the subprime and Alt-A
market that came to replace the prime or conventional market. The
aggressive pursuit of that market by banks of all kinds has led us to the
current situation. The main role of the credit rating agencies was to allay
concerns by assuring participants that the transformation of the securitiza-
tion market was not a dangerous one. The main role that regulators played
was to refuse to intervene into these markets. The Federal Reserve was
dominated by people who believed that in spite of this dangerous shift in the
P54
market, market actors would not take on too much risk. We now know this
was wrong. The evolution of banks’ strategies since the early 2000s had left
them highly leveraged on assets which were largely junk.
It is useful to briefly outline the mechanics of how mortgage defaults
reverberated through the financial system. There were two main forces
that eroded the positions of banks and the GSE beginning in 2007. First,
the rate of foreclosures on ‘‘AAA’’ subprime MBS bonds turned out to be
higher than was predicted. By July 2007, delinquency and foreclosure
rates had risen to the point that even the reticent bond rating agencies were
forced to start mass downgrades of subprime tranches. As their price
dropped, banks that had taken loans to buy the MBS had to either pay off
those loans or put up more collateral to keep them. This was because most
of their loans contained covenants that required them to up their capital
investment if bond prices fell or the credit rating on the collateral was
downgraded.Most banks were very highly leveraged and eventually found it
impossible to raise enough capital to cover their loans. By mid-2007, it was
clear that subprime mortgages were undermining bond prices and pressure
was brought to bear on all of the banks. By spring of 2008, banks like Bear
Stearns began to fail.
It is useful to look at what has happened to the top banks that were leaders
in the mortgage securitization business ca. 2005. Tables 3 and 4 show how the
top 10 firms in the subprime mortgage origination business and the subprime
P55
MBS business fared. In 2005, 7 of the 10 largest subprime lenders are either
out of business or absorbed by merger. And in the same year 8 of the 10 top
subprime MBS firms are either out of business or merged into other entities.
The collapse of the subprime market essentially wiped out all of the firms that
had grown large on that business. The big investment banks at the core of the
subprime MBS market no longer exist with the exception of Morgan Stanley
and Goldman Sachs. Citibank, Bank of America, JP Morgan Chase, and
Wells Fargo have emerged as large conglomerate banks having absorbed
many of the subprime losers, whereas both Goldman Sachs and Morgan
Stanley have reorganized themselves to become commercial banks.
One can expect many more bank mergers and it is likely that a few large
conglomerate banks will end up dominating the American banking scene
across all products. The complete collapse of the investment banks shows
the folly of subprime mortgage securitization. The managers of the firms
that took this risk have lived by the sword and died by the sword. What has
saved the economy is the government takeover of the GSE and the
propping up of the rest of the banking system. The Federal Reserve now is
the largest purchaser of MBS. In an ironic way, the MBS market has
come full circle. The government began by attempting to stimulate the
housing market in the 1960s and 1970s. They were pleased to invent and
support the market and do what it took to bring in private investment. But
eventually, those banks expanded their activities into risky investments with
P56
borrowed capital. The government has returned full bore into the market
and now underpins it.
SUBPRIME MBS: SOME MYTHS AND REALITIES
There are several alternative perspectives on the crisis to the one we
presented earlier. In this section, we examine these competing accounts and
show that they are inconsistent with key pieces of evidence. One popular
account of the crisis focuses on the structure of transactional incentives.
Proponents of this account argue that perverse incentives and information
asymmetries between the sellers and buyers of MBS encouraged the former
to sell off risky assets. It also meant that those who bought the bonds did
not know how risky subprime mortgages were.
Because MBS are bonds, they are regulated by the Security and Exchange
Commission (SEC). When a conduit bank wants to issue an MBS, it has to
file a prospectus with the SEC. These prospectuses are public information
that can be accessed via the web in two or three clicks of a mouse. The term
‘‘subprime’’ actually has a set of formal definitions. To qualify for a prime
or conventional mortgage, a person needs 20% down and a credit FICO
score of 660 or above (the average score is 710 on a scale from 450 to 900).
Mortgagees who did not have these qualifications are not eligible for prime
or conventional mortgages. But, if they are willing to pay a higher interest
rate, they could qualify for an Alt-A or subprime mortgage.
It is useful to be explicit about what constitutes bad credit. Here are some
of the conditions that could qualify a mortgagee as subprime: two or more
delinquencies in the past 12 months; one or more 60-day delinquencies in the
past 24 months; judgment, foreclosure, or repossession in the prior 24
months; bankruptcy in the past 5 years; a FICO score less than 660; and
debt service to income ratio of 50% or greater (i.e. the monthly payment
was more than 40% of the gross income of the household). It should be
noted that these sets of characteristics were listed in all of the prospectuses
filed with the SEC.
Fig. 12 presents the key institutions that helped to create one special
purpose vehicle for subprime mortgages: GSAMP Trust 2006-NC2. The
originator for the trust was New Century Financial, one of the largest
subprime originators. Goldman Sachs acted as the conduit for the trust.
Moody’s rated the bonds in the trust. Ocwen acted as the servicer of the
trust. Deutsche Bank and Wells Fargo acted as advisers to the trust. This
trust is typical of the MBS packages that were done during this period.
P57
The following information is available in the prospectus for the trust. 2
There were 3,949 subprime mortgagees in the trust worth $881 million. 43.4%
were used to buy a new house whereas the rest were re-financing of existing
loans. 90.7% of the mortgagees were going to live in the house. 73.4% were
single family dwellings and the rest were condominiums. 38% of the homes
were in California and 10.5% in Florida. The average borrowers had a FICO
score of 626. 31.4% had a score below 600, 51.9% between 600 and 660, and
only 16.7% above 660. The ratio of total debt to income was 42% in the
whole set of mortgages. Table 5 shows the tranches and the bond ratings of
the tranches. Approximately 79%of the bond offering was rated ‘‘AAA,’’ the
highest ratings. Less than 5% were rated ‘‘B’’ which should be more typical of
a subprime rating. It is clear from the detail of this information that anyone
who wanted to understand what they were buying in GSAMP 2006-NC2
could not claim to not understand what they were viewing.
A second myth about subprime MBS is that neither the issuers nor the
conduits held onto the bonds. We have already briefly discussed Fig. 7
which presents data on the largest holders of MBS from 2002 to 2008. The
GSE are the largest holders of bonds. From 2002 to 2008, when the supply
of subprime mortgages increased most rapidly, commercial banks and bank
holding companies increased their holdings of MBS from $650 billion to
$1.1 trillion. As we noted earlier, conduits who packaged the MBS deals
increased their MBS holdings 400% from $30 billion to $175 billion between
2002 and 2007, even as bonds became progressively riskier over those
P58
vintages (Inside Mortgage Finance Publications, 2009). The conduits and
commercial banks were deeply involved in the production of MBS; the fact
that they retained such a significant portion of the assets (and often went
bankrupt as a result) casts doubt on the notion that the crisis occurred
because intermediaries strategically sold off all the riskiest assets to
unwitting investors.
A third place where commentators have identified problematic incentives
is in the ratings process. The basic problem here was that the security issuers
paid the rating agencies rather than buyers. This put the agencies in the
position of having to inflate ratings to satisfy their customers (the issuers),
who would otherwise take their business elsewhere in a practice known as
ratings shopping. The result is that issuers are able to bid up the ratings for
their securities. Proponents of this explanation point to the fact that so-
called shopped ratings for a given security tend to be higher than unsolicited
ratings. It is also consistent with the data we presented earlier, which shows
the agencies greatly inflated bond ratings for new MBS issues as the market
grew. The problem with this explanation, however, is that this same perverse
incentive structure had been in place for some time ( Tett, 2008). Only after
P59
2003 did it lead to rampant ratings inflation. 3 Although perhaps necessary,
perverse credit rating agencies incentives are insufficient to explain the over-
time change in ratings composition. What exactly happened here remains a
bit of a puzzle.
Overall, an explanation of the credit crisis based on the structure of
incentives fails to account for the behavior of the main players. It is true that
the structure of incentives suggested that cunning, rational firms should have
securitized risky subprime mortgages and then sold them all off to unwitting
investors who could not possibly understand the risky assets contained
within their opaque structures. But empirically this is not what happened in
the case of subprime MBS. The firms that originated, packaged, and issued
subprime MBS did hold on to a considerable portion, largely because they
were yielding high short-term profits. Secondly, those investors who
purchased the remaining portion of subprime MBS could easily access
information on the underlying mortgage collateral. Rather than a scenario in
which each actor was fleecing the next guy down the line, the data is more
consistent with a theoretical imagery of collective, field-wide delusion.
In particular, the fact that the banks held on to so much of their own junk
poses a problem for the rational actor model. One important question raises
is if banks knew the assets they held were extremely risky, why did they hold
onto them? Even more important, why did almost all of the banks in the
core of the market do it? Indeed, with the exception of Goldman Sachs and
to a lesser degree, JP Morgan, all of the large banks continued to hold on to
the subprime MBS to the very end. Answering this question is beyond the
scope of this article, but we return to this issue below as an important area
for future research.
Another competing perspective on the credit crisis focuses on financial
engineering innovations and the resulting complexity of securitized assets.
Although not mutually exclusive, this explanation differs from ours by
focusing explanatory attention on the technology of financial instruments
over the strategies of financial actors. For instance, MacKenzie (2009) writes
that ‘‘The roots of the crisis lie deep in the socio-technical core of the
financial system’’ (p. 10).
There are several reasons why the growing complexity of financial assets
may have heightened risk or served to conceal the risk of subprime MBS.
Although standard MBS allowed issuers to construct predominantly AAA
tranches from subprime mortgages, CDOs essentially allowed for a double
upgrade by taking the mezzanine-level (BBB) tranches from conventional
MBS securities (those tranches that were first to lose in the event of default),
and repackaging them as AAA CDO tranches. MacKenzie argues this made
P60
CDOs especially dangerous while simultaneously making them appear less
risky and more palatable. However, there has been no empirical evidence
that CDO structures contributed to the meltdown above and beyond the
MBS assets from which they were constituted. In other words, proponents
of the complexity explanation have provided little evidence that complexity
had an independent effect.
One way to empirically test the complexity argument is to examine how
ratings dynamics for credit default options differed from more convention-
ally structured MBS assets. If these arguments are correct, ratings inflation
should be most pronounced among the most complex instrument classes
(CDOs), and CDOs instrument’s increased riskiness should be evident in the
subsequent downgrades they experienced after the bubble burst.
Contrary to the hypothesis that financial engineering drove ratings
inflation, the data shows that the most highly complex and innovative
CDO instruments actually displayed the greatest constancy in their overall
ratings composition. As shown in Fig. 11, the initial ratings for newly issued
CDOs remain remarkably stable over time compared to the inflationary
trends evident in the MBS assets. What this means is the repackaging of
MBS tranches into CDOs did not further contribute to ratings inflation.
Despite a great deal of talk about the especially acute dangers of complex
CDO instruments, our data show that these instruments were no more
dangerous than the underlying MBS on which they were built, at least as
measured by subsequent downgrades. In fact, as Fig. 13 shows, CDOs,
along with Whole Loan (i.e. nonconforming ‘‘jumbo’’) securities, actually
tended to be somewhat less overrated than B/C, Alt-A, or HEL securities.
This suggests that variations in overrating were related more to the
underlying quality of the mortgage debt than the complexity of the bond
structure. It also suggests that those who pin the sources of the crisis on
the growth of more complex securitization structures must present better
evidence for how CDOs heightened the riskiness of the underlying MBS on
which they were based.
CONCLUSIONS
We began this article by suggesting that there has been a symbiotic
relationship between government regulation, the housing market, and
the main private sector players in that market for the past 40 years.
Beginning in the 1960s, the government wanted to stimulate the market in
order to expand the opportunities that citizens would have to own homes.
P61
Because of budgetary problems stemming from the Great Society and the
War on Poverty in the 1960s, the Johnson Administration innovated a
set of tactics to increase the incentives of the private sector to loan money
for mortgages. They created the government sponsored enterprises Fannie
Mae, Freddie Mac, and Ginnie Mae to buy mortgages and repackage them
into MBS.
The private part of this market struggled to take off in the 1970s and early
1980s due to a number of problems. Some of these were regulatory and
some were due to the lack of interest on the part of private capital to invest
in mortgages. Investment banks like Solomon Brothers and Morgan Stanley
convinced the federal officials to support changes to the tax laws to make
MBS possible and attractive to investors. The deregulation of the savings
and loan banks inadvertently further promoted the mortgage securitization
revolution by pushing the savings and loan banks out of their traditional
Main Street mortgage finance role. This left a void that Wall Street
progressively filled. Government regulators were always friendly toward the
investment bankers and commercial bankers who worked with the GSE to
push forward MBS. For example, they worked to let banks introduce
variable rate mortgages. These products protected banks against sudden
P62
increases in interest rates while forcing consumers to pay more if interest
rates went up.
The mortgage market expanded rapidly during the 1990s and securitiza-
tion tools allowed consumers to get mortgages and all other forms of credit
more easily and cheaply. The mortgage market shifted from a $500 billion to
a $1.5 trillion a year market by the end of the decade. This attracted the
largest banks in the country to enter the expanding market. It also caused
them to lobby the government to allow banks to be in all parts of the
mortgage business. Once again, both the executive and legislative branch
accommodated the wishes of the financial sector. The largest commercial
banks which had become holding companies and the investment banks came
to dominate the market by the turn of the 21st century.
After the stock market crash of 2000, the Federal Reserve dropped
interest rates dramatically. This created the conditions for a rapid expansion
of the mortgage securitization market. One fact that we have noted that has
not been widely noted elsewhere is that this created a huge demand for
MBS. To fill this demand, originator banks and conduit banks (sometimes
the same people) needed to find a new mortgage market. The market they
found was the subprime market. The Federal Reserve consistently refused to
reign in the fast growing subprime part of the market. In the end, almost all
of the large players in the financial system came to own lots of MBS. The
ones who did so by borrowing money cheaply found themselves in a
liquidity crisis beginning in 2007.
It is common to hear discussions of the subprime market in two sorts of
terms. First, commentators often view subprime mortgagees as either ill
informed about what they were buying or else profligate for trying to live
beyond their means. Second, critics of this perspective argue that the goal of
subprime mortgages was to expand homeownership to poorer people or
people with less stellar credit. Since homeownership is one core part of the
American dream, there are commentators who praise the rapid expansion of
such credit to those consumers.
Although sympathetic to the second claim, our discussion suggests that
the speed with which the subprime mortgage market was expanding should
have alarmed government officials. The rapid decrease in the conventional
mortgage share of the market coupled with the rapid increase in the
nonconventional part of the market should have alerted government
officials to a potential problem. Moreover, the increasing use of leverage to
buy and hold MBS was a clear sign that low interest rates were driving
banks to find people to lend money so that the banks might borrow money
to loan to those consumers and then hold onto the bonds. Finally, the rapid
P63
inflation in the ratings for subprime mortgages as they grew to comprise an
ever larger portion of the market should have raised many questions. The
entire structure of the core of the U.S. banking system quickly found its way
in making its main profits from putting together MBS deals for
nonconventional mortgages.
We began this essay by suggesting that one of the main problems of
regulators is that they actually lacked a view of how the financial market is a
system. By system, we mean a set of markets that have a small set of
participants, all of whom know one another, watch one another, and imitate
one another’s moves. These same players came to want to occupy positions
in all parts of the mortgage market to expand their businesses and make fees
off of selling mortgages to consumers, packaging those mortgages for
investors, selling those mortgages to investors, and holding onto a sizeable
part of the MBS that they bought with borrowed money. In this system,
every bank could make money in multiple ways.
Regulators facilitated the creation of this structure in many ways. They
helped pioneer the instruments and they supported the market by creating
GSE that offered government guarantees to mortgage investors. They
passed friendly legislation whenever the industry asked for it. Most
importantly, they declined to intervene into the markets almost under any
circumstances. They did so trusting that the bankers knew what they were
doing. We would argue that this was a kind of passive form of regulatory
capture. Bankers made their wishes known to regulators and they eventually
got favorable legislation and lax regulation. Bankers also took advantage of
regulators by shopping for regulators who would treat them favorably
(Davis, 2009). Government officials essentially trusted the market and
believed that the private incentives were in place to prevent any form of
meltdown.
If we are correct, this implies that any changes in regulation must
begin with the idea that the mortgage securitization industry is a system,
a set of interconnected markets dominated by a few players with a set
of tactics that every player knew. In the introduction, we argue that
regulators, legislatures, the executive branch, the banks themselves, and
the bond ratings agencies have all been part of the MBS story. We note
that until 2003, the MBS market appeared to be working fine. It provided
large amounts of loans to a large number of consumers and it grew
dramatically. It is our belief that this success made regulators after 2003
wary of thinking that anything might have been wrong. But, with hindsight,
we think there are some important lessons to learn and some obvious
reforms to be made.
P64
WHAT DO NOT WE KNOW?
The story we have told has outlined the anatomy of how mortgage
securitization rose and fell. We show how consistent with a view of
markets as fields, the government played a critical role in the founding of
the market and its subsequent evolution. We have also showed how the
government over time also wanted the private sector to become more
and more involved in providing mortgages to more and more Americans.
This meant that Presidents, Congress, and regulators generally acted to
give private firms what they wanted to get them to be more involved in
the market.
This raises two interesting questions. First, government officials certainly
broadly accepted what Kaufman has called ‘‘economic libertarianism’’
(2009, p. 235). By this he means that they viewed the role of government as
mostly needing to stay out of the way of financial market processes. In other
words, financial markets were efficient and they worked. They had this view
certainly from the mid-1980s on and it caused them to seem to forget that
the government had initiated the market and continued to underwrite
it through the GSEs. Second, how did policymakers understand the link
between their more overt roles in helping to produce and sustain this market
and their more prosaic view that the market was self-regulating?
A related question is the role of the banking industry in this process.
We have two different images of this process. One is that individual banks
appear to use the system ruthlessly to their advantage by working to escape
regulations and shopping for the easiest regulation they can get. The other is
that the industry got what it wanted through regulatory capture: allowing
for variable mortgages, changing tax laws so that MBS could be held by
pension funds and insurance companies, repealing Glass–Steagall, and
affecting the accounting involved with MBS they held. Was there regulatory
capture or did firms really have to regulation shop to get what they wanted?
There are a number of other unanswered questions. We have just begun
to explore the degree to which the mortgage business changed, particularly
from 1990 onwards. The conventional wisdom is that 1980s financial
deregulation and the decline of Savings and Loan associations led to vertical
disintegration of the mortgage finance value chain. On the contrary, our
data suggest that banks grew bigger and more integrated around mortgage
finance. It also implies that banks viewed it as necessary to partake in
each transaction along the chain, from selling mortgages to individuals, to
packaging mortgages into MBS, to servicing MBS, and finally to retaining a
portion of their MBS as investments. It is important to understand better this
P65
vertically integrated market structure. One argument is that banks shifted
from client-based banking to transaction or fee based banking in the 1990s.
This shift in their underlying conception of control is not well documented
nor well understood. To the degree that it happened, it explains why banks
found it irresistible to capture all parts of the mortgage transaction chain. In
order to understand this process better, it is necessary to undertake studies
that incorporate archival, interview, and quantitative data.
The question of what investors really knew about the MBS they were
buying also remains not totally understood. Behavioral finance claims that
actors are greedy, tend to underestimate risk, tend to be too optimistic, tend
not to have good knowledge of risk, and finally tend to follow herds. The
dominant explanation for this and other bubbles rests on these assertions
(for examples of this kind of argument see Schiller, 2008; Akerlof & Schiller,
2009; Krugman, 2009 ).
We are skeptical of strong forms of this argument for a couple of reasons.
There is evidence that the buyers and sellers of MBS did believe that
mortgages were relatively safe and that this was based on historical data and
the use of CDS as insurance. We see little evidence that information about
the riskiness of subprime loans/MBS was unavailable to investors. However,
we do not know how much they sought such evidence out or whether they
were duped by high-credit ratings and convincing bond salesmen. We also do
not really know the degree to which higher level executives did or did not
know how risky the investments their underlings were making were or if they
just did not care because the housing market seemed like a safe bet and they
assumed the government would bail them out if there were truly a disaster.
There are several possible answers future research might explore. First, there
is a great deal of anecdotal evidence that bank executives understood CDS as
a form of insurance which they thought would protect them in a downturn.
Second, all of the banks were making lots of money off of borrowing money
at low interest rates and holding on to subprime MBS investments. Many of
these firms probably thought that if they had to sell off the MBS, they could
do so into what seemed like a relatively liquid market. Since they were
making essentially free money off of these investments, losses would have to
be prohibitively large to offset the gains to date.
Another related issue concerns the rationality of the main actors after the
housing market began to stall. How rational were banks in understanding
what was going on once these markets began to unravel? Were they in
denial, as James Cayne of Bear Stearns appeared to be, or were they caught
in a downward spiral that they understood, but could do little about given
the tight interconnectedness of the financial system?
P66
It is clear that the markets and most of the dominant firms registered little
response when housing prices started to stall out and foreclosure rates began
to rise as early as late 2006. Several large banks such as Merrill Lynch
continued to aggressively expand their nonprime businesses in early 2007,
and signs of troubles in the subprime market were widely portrayed in the
financial news media as an isolated issue. Only once Bear Stearns could no
longer fund their obligations did the markets also begin to look for other
banks that were similarly vulnerable. It was this cascade of information that
eventually brought the whole show down.
Finally, we know little about the people who used subprime mortgages to
buy their homes and why they did so. We have multiple images of such
buyers. First, we are told that poor people were sold a bill of goods whereby
they got homeownership but with a predatory loan. These were people with
no jobs, no income, and no down payment. Second, we are also told that
about one quarter of loans at the end of the mortgage bubble were going to
speculators who never intended to live in their homes but instead intended
to sell the properties as soon as prices rose. Third, we have the image of
people who live in parts of the country where home prices are rising so fast,
that they no longer can afford to buy with a conventional mortgage. These
middle class people seeking out good neighborhoods and good schools for
their families end up deeply in debt. They are forced into the subprime
market because they cannot afford a conventional loan. It is important to
sort out the degree to which each of these stories is true in order to protect
consumers and regulate mortgage markets more effectively.
POLICY RECOMMENDATIONS
Even without knowing everything about what happened in great detail, it is
possible to consider what types of regulation might make sense. There are a
number of areas where government regulators either failed to use the powers
they had or Congress refused to add additional regulation. We list those
below and suggest regulatory actions.
1. Subprime borrowers were the victims of predatory lending practices.There is sufficient evidence for this point (e.g.Albers, 2008; Sanders,2008; not presented here). What we have presented here is that banksrapidly moved to sell mortgages to people who probably have credit veryfar beyond what made sense. There is evidence suggesting thatfrequently, these people did not know what they were buying.
Solution: Laws should be passed and enforced prohibiting such lendingpractices. Such laws would have protected consumers and tempered therapid growth of the subprime bubble, without altogether inhibiting theprovision of housing finance to those underserved populations wholegitimately needed it.2. The credit rating agencies are paid by the packagers of the bonds. Thiscreates a conflict of interest whereby the bond rating agency does nothelp the buyer of the bond but the seller.Solution: Credit rating agencies should be held responsible for thetransparency of their rating schemes. It would also be a good idea tofigure out how to have the buyers of MBS pay for the credit ratings andnot the sellers of loans. There is a clear conflict of interest here and onethat obviously had an effect.3. Investment banks and other investors used loans to buy mortgagesecurities and formulas that assumed inappropriately low default ratesand inappropriately high prices. This meant that they were leveragingtheir capital substantially. When the underlying assets began to fall, theywere unable to back them up and raise additional capital.Solutions: Banks should be regulated more closely to insure that theircapital is adequate to cover their potential losses. Banks should also haveto put all of their financial risks on their books and not be allowed tohave special purpose vehicles that remain off books. They should alsohave to use more realistic risk assessment models for those securities.4. The Federal Reserve decided to keep interest rates low in the 2001–2005.This encouraged the housing bubble and it encouraged banks to borrowmoney cheaply to buy MBS.Solutions: The Federal Reserve should take more seriously its role inproviding the fuel for bubbles whether they are stock market or housingmarket related. Had the Federal Reserve raised interest rates earlier, theinvestment banks would not have had the incentive to borrow moneycheaply to buy MBS. They should have also been wary of rapid shifts inmarkets such as that took place from 2003 to 2007 in the MBS marketand not assume that market actors understood the risks.5. The largest banks that emerged were allowed to use their highly regulatedactivities to provide capital for their riskier activities. This meant thatwhen their loans went bad, they were deemed ‘‘too big to fail’’ andrequired large bailouts.Solution: Bank regulation should provide for the separation ofactivities of bank holding companies such that the riskiness of the banksendeavors is properly accounted for in their capital.
6. CDS were viewed as a form of insurance. Since the CDS market was nottransparent, it was not clear who was insured against risk and how muchinsurance they had.Solution: If the CDS market is to be an insurance market, it should beregulated like insurance. This would prevent companies like AIG fromnot properly exposing their risk to CDO products.If these sets of regulations were in place, would they have been enoughto prevent the current financial crisis? We think that if regulators hadeffectively enforced such a set of regulations, they would have made it moredifficult for subprime lending to expand so dramatically after 2003. Thegeneral level of the housing bubble meant that not just subprime borrowerswere at fault. Here, the Federal Reserve probably should have undertakenefforts to keep the housing bubble under control. They could have done soby raising interest rates which would have had the effect of making it lessattractive for investment banks to borrow money to buy mortgage-basedsecurities. Similarly, if the bond rating agencies had done their jobs moretransparently and effectively customers would have had a better idea aboutwhat they were buying.
NOTES
沒有留言:
張貼留言