代写留学生金融硕士论文-Not What They Had in Mind:A History of Policies

发布时间:2012-02-22 09:13:22 论文编辑:代写留学生金融硕
代写留学生硕士论文Electronic copy available Not What They Had in Mind:A History of Policies that Produced the Financial Crisis of 2008

Executive Summary
The financial crisis of 2007 to 2008 will go down as one of the most significantevents in economic history. Large financial institutions such as Bear Stearns and
Lehman Brothers failed, and stock prices plummeted. This major crisis affected thereal economy, culminating in the current recession, and many analysts predict a long
road to economic recovery for the United States.The severity of the current crisis raises many questions about its root causes. Any
attempt to understand these root causes, however, requires the placement of policies
and regulations in the appropriate context.
This paper looks at the roots of the current crisis through an analytical framework ofbad bets, excessive leverage, domino effects, and 21st-century bank runs. The paper
shows that broad policy areas—including housing policy, capital regulations for banks,industry structure and competition, autonomous financial innovation, and monetary
policy—affected elements of this framework to varying, but important, degrees. Whileconsidering alternative points of view concerning the causes of the financial crisis, the
paper concludes that bank capital regulations were the most important causal factorin the crisis and that the policy “solutions” to previous financial and economic crises
sowed the seeds for this current crisis.
To fully understand the current crisis, one must account for the complex history, evolution,and integrated nature of financial regulations. Without this evolutionary history,
there will be no meaningful lessons for today’s policy makers. Unless the UnitedStates comes to terms with the fact that the actions of policy makers and regulators
contribute to financial fragility, it has little hope of moving in the direction of a lessfragile system for the future.
I would like to thank Ben Klutsey for research assistance. I would also like to thankLawrence J. White, Tyler Cowen, Russ Roberts, Brian Hooks, and Rob Raffety for
helpful comments. Errors that remain are my own.
The ideas in this paper do not represent an official position of George Mason Universityor the Mercatus Center. The Mercatus Center wishes to acknowledge the support ofthe Legatum Institute in making this project possible.
Not What They Had in Mind: A History of Policies that Produced the Financial Crisis of 2008
2
Mercatus Center at George Mason University
3
contents
Executive Summary 1
1. Introduction 5
2. Framework for Understanding the Financial Crisis 6
2.A. Bad Bets 6
2.B. Excessive Leverage 6
2.C. Domino Effects 6
2.D. 21st-Century Bank Runs 7
2.E. The Four Elements Together 8
3. The Matrix of Causal Factors 9
Figure 1. Policy Importance 9
4. Past Crises Make Bad Policy: Housing Policy and Capital Regulation
11
5. Housing Policy 13
Figure 2. Housing Policy Timeline 14
5.A. CRA and the Under-Served Housing Market 17
Figure 3. Changes to Capital Rules Timeline 20
6. Bank Capital Regulations 22
Figure 4. Changes in Capital Requirements 25
7. Erosion of Competitive Boundaries 29
Figure 5. Competitive Boundaries Timeline 30
8. Financial Innovation 33
9. Monetary Policy and Low Interest Rates 38
10. Domino Effects and Bank Runs—Revisited 39
11. Easy to Fix vs. Hard to Break 41
12. Conclusion 42
Appendix: The Shadow Regulatory Committee on Barriers to Entry 45
Not What They Had in Mind: A History of Policies that Produced the Financial Crisis of 2008
4
Mercatus Center at George Mason University
5
“Those who cannot remember the past are condemned
to repeat it.”
—George Santayana
1. INTRO DUCTION
Many Americans who lived through the financial
crisis of 2008 will remember the stunning events
that took place: large, famous financial institutions
suddenly unable to survive as independent entities;
policy makers scrambling to prevent what they saw
as a potential catastrophe; massive taxpayer-funded
bailouts; plummeting stock prices; “toxic assets” with
exotic initials like CDO and CDS. Representatives of
credit rating agencies excoriated by congressional
committees. Executives at firms like AIG Insurance
and Merrill Lynch accused of excessive short-term
greed and risk-taking.
But those who only remember the headlines of 2008
will fail to heed Santayana’s warning. For the roots
of the crisis go back many decades, and if we are to
avoid repetition, we have to fully understand the
context in which decisions were made in the years
leading up to the crisis.
As this paper will illustrate, the seeds for much of the
current crisis were sown in the policy “solutions” to
previous financial and economic crises. Any attempt
to dissect and understand the current crisis that does
not account for the complex history, evolution, and
integrated nature of financial regulations will not
yield meaningful lessons for today’s policy makers.1
What made the crisis possible were the illusions that
key participants held during the years that preceded
the meltdown. Financial executives had excessive
confidence in mathematical models of risk, in financial
engineering, and in the “AAA” designation of
credit rating agencies. However, it is misleading to
simply write, in the words of one prominent white
paper, that “Market discipline broke down as investors
relied excessively on credit rating agencies.”2
What this formulation overlooks is the fact that regulators
themselves encouraged the reliance on agency
ratings, particularly for compliance with bank capital
requirements. In fact, we will see that the regulatory
impetus to use agency ratings dates back to the
1930s, was reinforced in the 1970s, and was significantly
enhanced as recently as January 1, 2002. To
ignore these regulatory policies and instead assert
that agency ratings were relied on because “market
discipline broke down” is to present a distorted view
of history.3
The fact is that the regulatory community shared in
the illusions of key market participants. Regulators,
too, placed too much confidence in financial engineering.
Regulators, too, thought that the dispersal
of risk into the “shadow banking system” helped
make the core financial system safer. Regulators,
too, thought that securitization was a superior form
of mortgage finance.
This paper examines the history of the evolution
of financial markets and financial regulation as it
pertains to the financial crisis. While it considers
alternative points of view concerning the causes of
the crisis, it takes a particular position, based on my
experience in looking at competition in the market
for mortgage credit risk. Specifically, it emphasizes
the role played by bank capital regulations in
promoting
the practices that produced an unstable
financial system.
The next section presents a framework for looking at
the crisis as a combination of four elements: bad bets,
excessive leverage, domino effects, and 21st-century
1. To find poor historical analysis, one need only examine Department of the Treasury, Financial Regulatory Reform—A New Foundation:
Rebuilding Financial Supervision and Regulation (Washington, DC: GPO, 2009), http://financialstability.gov/docs/regs/FinalReport_web.
pdf/. This 89-page white paper overlooks many key historical factors.
2. Ibid., 3.
3. Ibid.
Not What They Had in Mind: A History of Policies that Produced the Financial Crisis of 2008
6
bank runs. This in turn allows one to assess the relative
importance of five broad policy areas:
• housing policy;
• capital regulation for banks;
• industry structure and competition;
• autonomous financial innovation (not driven
by capital regulation); and
• monetary policy.
To understand how policies in these areas might
have contributed to the crisis, we need to have a
framework that describes the crisis. Once we know
how the crisis came about, we can start to allocate
responsibility to various policy areas.
2. A FRAM EWOR K FOR UN DERSTAN DIN
G THE FINANCIA L CRI SIS
The financial crisis can be thought of as consisting
of four components:
1. bad bets;
2. excessive leverage;
3. domino effects; and
4. 21st-century bank runs.
2.A. Bad Bets
Bad bets were the investment decisions that individuals
and firms made that they later came to regret.
They were the speculative investments that drove the
housing bubble. When consumers in 2005 through
2007 purchased houses primarily on the expectation
that prices would rise, those investments turned out
to be bad bets. When lenders held securities backed
by mortgage loans made to borrowers who lacked
the equity or the income to keep their payments current
during a downturn, those were bad bets. When
AIG insurance sold credit default swaps (CDS) on
mortgage securities, giving AIG the obligation to pay
insurance claims to security investors in the event of
widespread mortgage defaults, those were bad bets.
One way to estimate the significance of bad bets is
to estimate the loss in the value of owner-occupied
housing. The peak value was roughly $22 trillion, and
if house prices declined by 25 percent, this is roughly
a $5 trillion loss. This is a reasonable estimate of the
order of magnitude of the losses from bad bets.
2.B. Excessive Leverage
Banks and other financial institutions took on
significant risks without commensurate capital
reserves. As a result, declines in asset values forced
these institutions either to sell hard-to-value assets
or face bankruptcy. Commercial banks had insufficient
capital to cover losses in their mortgage
security portfolios. Freddie Mac and Fannie Mae
had insufficient capital to cover the guarantees that
they had issued on mortgage securities. Investment
banks, such as Merrill Lynch, had insufficient capital
to cover losses on mortgage securities and derivatives.
AIG insurance had insufficient capital to cover
the decline in value of its CDS portfolio.
In hindsight, large financial institutions were far
too fragile. They were unable to withstand the drop
in value of mortgage-backed securities that in turn
stemmed from falling house prices.
2.C. Domino Effects
Domino effects are the connections in the financial
system that made it difficult to confine the crisis
to only those firms that had made bad bets. Healthy
institutions could be brought down by the actions of
unhealthy institutions. For example, when Lehman
Brothers declared bankruptcy, a money market fund
known as Reserve Prime, which held a lot of Lehman
debt, indicated that it would have to mark the value
of its money market fund shares to less than $1 each
(“breaking the buck” in financial parlance).
Of course, one could argue that Reserve Prime was
not so much the victim of a domino effect as it was a
bad bettor. Financial professionals had been aware
for months that Lehman was in difficulty, and keeping
a large position in Lehman debt can be viewed
as making a bet that the government would treat
Lehman as “too big to fail.”
Another domino effect potentially comes from sales
of hard-to-value assets. Suppose that Bank B holds
rarely traded securities and that the most recent market
prices indicate a value of $X for those securities.
However, Bank A is in distress and so must sell similar
assets at a low price. This causes Bank B to mark
its assets down below $X. As a result, Bank B falls
below regulatory capital requirements and must sell
some of these assets. This depresses their price further,
causing Bank C to mark down its assets and fall
below its minimum capital requirements, and so on.
We may never know how serious domino effects
might have been in the financial crisis because the
federal government took such strong steps to prop
up institutions. For example, we do not know what
would have happened if the government had allowed
Freddie Mac and Fannie Mae to go into bankruptcy.
Presumably, institutions with large holdings of
government-sponsored enterprise (GSE) securities
would have suffered major losses.
2.D. 21st-Century Bank Runs
In a traditional bank run, depositors who wait
to withdraw their money from an uninsured bank
might find that the bank is out of funds by the time
they reach the teller. That creates an incentive for
a depositor to run to the bank so as to be the first in
line—hence a bank run. By 21st-century bank runs,
I mean the financial stress created by situations in
which the first creditor that attempts to liquidate its
claim has an advantage over creditors that wait.
The incentives for bank runs come from a structure
of financial claims that leads individual agents to
form mutually incompatible contingency plans. In
the case of an uninsured bank, each depositor’s contingency
plan may be to withdraw funds at the first
sign of trouble. Such plans are incompatible because
if too many depositors attempt to execute their plans
at once, they cannot all succeed. Instead the bank
will fail.
For AIG Insurance, credit default swaps resulted in a
21st-century bank run carried out by counterparties.
Banks that had purchased protection on mortgage
securities from AIG were not sure that AIG had the
resources to make good on its swap contracts. These
counterparties exercised clauses in their contracts
that allowed them to demand good-faith collateral
from AIG in the form of low-risk securities, even for
credit default swaps on securities that had not yet
defaulted. The demands for collateral soon exceeded
the available liquid assets at AIG, which might have
forced AIG either to liquidate valuable assets hurriedly
or to declare bankruptcy. It was at that point,
in late September of 2008, that the government
stepped in to provide the low-risk assets that enabled
AIG to meet its collateral obligations in exchange for
the government taking over most of the equity value
of AIG.
These 21st-century bank runs caused the failures of
the large investment banks. They held portfolios of
illiquid securities, including tranches of mortgagebacked
securities, that they financed in the “repo”
market, meaning that they borrowed funds and used
the illiquid securities as collateral.4 When investors
developed concerns about the value of mortgage
securities, they greatly reduced their willingness
to make “repo” loans to institutions offering those
illiquid securities as collateral. For investment banks
with large inventories of securities to finance, this
created a shortage of liquidity. For such institutions,
the situation felt like a bank run.
Mercatus Center at George Mason University
7
Suppose 4. that institution A holds a mortgage-backed security, which it wants to carry using short-term financing. Institution A sells the
security to institution B, but institution A commits to repurchase the security in one week at a slightly higher price that reflects the short-term
interest rate. Institution B is said to make a “repo” loan to institution A with the security as collateral. If institution A were to default on the loan,
institution B would retain possession of the security.
Similarly, the structured investment vehicles (SIVs),
created by commercial banks, were attempting
to carry long-term, mortgage-backed securities
financed with short-term commercial paper. When
investors became concerned about the value of the
mortgage securities, the commercial paper market
dried up. This created conditions among the SIVs
that were similar to a bank run.
The 21st-century bank runs suggest multiple equilibria.
An institution in the good equilibrium can hold
onto its long-term positions by rolling over shortterm
funding at low interest rates: The institution
proves solvent. In the bad equilibrium, the institution’s
creditors panic; it cannot roll over its shortterm
funding except at very high interest rates, and
the institution collapses. With domino effects, the
bad equilibrium spreads from one firm to another.
Domino effects and 21st-century bank runs exposed
a weakness in the ability of regulators and courts to
handle failures of large institutions. If bankruptcy
or some other form of resolution could take place
quickly
with clear rules for determining the priorities
of various creditors, then there would be less
incentive for creditors to rush to exercise claims
on troubled institutions. In addition, this practice
would limit the domino effects because creditors
could obtain quickly whatever assets to which they
were entitled, rather than face months of legal uncertainty.
Finally, with an effective resolution authority
in place, government officials would not feel so compelled
to bail out troubled institutions.
2.E. The Four Elements Together
It is important to keep in mind that the financial
crisis required all four elements. Without the bad
bets, financial institutions would not have come
under stress. Without the excess leverage, the
bad bets would not have caused a financial crisis.5
Without the potential domino effects and the 21stcentury
bank runs, policy makers in 2008 would have
been less frustrated and frightened, and they would
have been hard pressed to justify the emergency
financial measures, including unprecedented financial
bailouts, if the crisis had been limited just to bad
bets and excessive leverage.
The government presumably designed the emergency
response to forestall domino effects and bank
runs. However, in the process of propping up troubled
institutions, policy makers also put themselves
in the position of insulating key firms from some of
their losses on bad bets. The ideal objective might
be to prevent domino effects and bank runs without
forcing taxpayers to absorb losses from bad bets.
However, that is a difficult needle to thread.
Because policy makers took such extensive measures,
it is difficult to gauge the significance of domino
effects and bank runs. As a result of the bailouts and
other policies, we presumably did not observe the
worst of what might have happened had the domino
effects and bank runs been allowed to play out. It is
impossible to know exactly how serious the consequences
would have been had those phenomena proceeded
unchecked.
To the extent that a financial institution was the victim
of bad bets and excessive leverage, one is tempted
to argue that those were its own choices and its managers
and shareholders should suffer the consequences.
These are losses due to bad decisions. On
the other hand, to the extent that an institution was
squeezed mostly by domino effects and bank runs,
one is tempted to argue that government action might
correct this bad equilibrium, as these are problems of
loss of confidence.
The regulatory response was focused on loss of confidence.
The Federal Reserve and the Treasury placed
more importance on loss of confidence than on bad
Not What They Had in Mind: A History of Policies that Produced the Financial Crisis of 2008
8
The collapse of stock prices in 2000 5. at the end of the dot com bubble illustrates how bad bets alone need not have catastrophic consequences
for the financial system or for the economy. Because the bad bets took place in the equity market, the stock market crash was fairly
self-contained, and the resulting recession was mild.
decisions. Both their actions during the crisis and
the reform proposals that they floated in 2009 were
focused mostly on issues related to domino effects
and bank runs.
In this respect, the financial regulators probably
reflected the views of the financial institutions. The
institutions saw themselves as victims of a loss of
confidence. In that regard, they reacted like executives
of other businesses under adversity. In general,
if you ask the CEO of a failed business what caused
the failure, the CEO will cite loss of confidence
rather
than bad decisions. As far as the oil wildcatter
is concerned, he was just about to strike oil when
his financing gave out. The founder of a startup that
burned through all of its cash will argue that he was
making great progress until his investors lost their
nerve. The retailer or real estate developer that goes
bankrupt will blame the banks for their unwillingness
to stretch out loans. Similarly, executives at
Citigroup or AIG will claim that the problem is not
the severity of their losses but the loss of confidence
on the part of their creditors and counterparties.
Accordingly, one has to be somewhat skeptical of the
claims that the financial crisis was primarily due to
an unwarranted loss of confidence.
The evidence for bad decisions includes the large
number of mortgage defaults and the large number of
downgrades of mortgage securities. It also includes
the fact that private hedge funds did not see much
opportunity in picking up distressed assets. If loss
of confidence were important, then on a temporary
basis assets would have been driven far below fundamental
values, and other firms would have found it
profitable to buy illiquid assets or to take over troubled
banks. As it turned out, only the government was
willing to try to take advantage of this profit opportunity.
If loss of confidence was the primary problem,
then the government’s investments in banks ought to
earn profits for the taxpayers. Even the AIG bailout
should ultimately provide taxpayers with a windfall
return. It is too early to say, but my guess is that this
will not prove to be the case.
3. THE MATRIX OF CAU SAL FACTOR S
The next step in understanding the historical evolution
of the financial crisis is to map policy areas to
the four elements of the crisis in terms of causal relationships.
As stated earlier, the five policy areas are
housing policy, capital regulation for banks, competitive
boundaries in financial intermediation, response
to financial innovation, and monetary policy. Below
is a matrix that includes my weights on the importance
of each of these factors relative to the column
heading. For example, I assign housing policy a high
weight in leading to bad bets and no weight in creating
bank runs. The remainder of this section will
present my rationale for these weights.
FIGURE 1: Policy Importance
Policy
Are a
Bad
Bets
Le ver age
Domino
Effects
Runs
Housing
Policy
High
weight
No weight No weight No weight
Capital
Regulation
Very high
weight
Very high
weight
Very high
weight
Very high
weight
Industry
Structure
No weight
Very low
weight
Low weight
Low
weight
Innovation
Low
weight
Low weight Low weight
Low
weight
Monetary
Policy
Low
weight
Low weight No weight No weight
As this matrix conveys, capital regulations were the
most important causal factor in the crisis. Capital
regulations encouraged banks and other financial
institutions to make bad bets, to finance those bets
with excessive leverage, and to set up financial structures
that were subject to domino effects and to 21stcentury
runs.
Bad bets were caused primarily by capital regulations
and by housing policy. As will be explained below,
Mercatus Center at George Mason University
9
Not What They Had in Mind: A History of Policies that Produced the Financial Crisis of 2008
10
capital regulations distorted mortgage finance away
from traditional lending and toward securitization.
Capital regulations specifically referenced credit
rating agency grades of securities, and these grades
proved faulty. Thus, banks were steered toward making
bad bets.
Another contributor to bad bets was housing policy.
Housing policy consistently encouraged more home
ownership and subsidized mortgage indebtedness.
This policy contributed to an unsustainable speculative
surge in home purchases.
It is worth noting that property bubbles took place
at around the same time in many other countries,
including the United Kingdom and Spain. These
property bubbles cannot be blamed on U.S. housing
policy. Thus, policy alone is not entirely responsible
for the bad bets. Clearly, there were other factors,
such as the apparent flow of savings from China or
other rapidly growing countries into Western property
markets.
Excess leverage should be blamed largely on the perverse
nature of capital regulations. These regulations,
which were supposed to constrain leverage,
instead were implemented in ways that encouraged
risk-taking. For commercial banks, regulators sanctioned
banks’ use of securitization, credit default
swaps, and off-balance-sheet entities to hold large
amounts of mortgage risk with little capital. For
investment banks, the SEC voted in 2004 to ease
capital requirements. For Freddie Mac and Fannie
Mae, the low capital ratios that had historically been
applied to investments in low-risk mortgages came
to be applied to the firms’ forays into subprime mortgage
securities. AIG Insurance, as a major seller of
credit default swaps, was effectively writing insurance
without being required to set aside either loss
reserves or capital. Thus, every major financial institution
was given the green light to pile on mortgage
credit risk with very little capital.
Regulators understood most of the reasons for the
increase in leverage, but they did fail to appreciate
some innovations. For example, it is unlikely that
the Office of Thrift Supervision, which had nominal
oversight of the AIG Insurance unit that sold credit
default swaps, understood the nature of the leverage
in AIG’s positions. Thus, I give a low but nonzero
weight to autonomous innovation in creating
excess leverage.
In explaining bad bets and excessive leverage, there
are those who place a higher weight than I do on the
monetary policy of the Federal Reserve. The argument
is that the Fed kept short-term interest rates
too low for too long, and this encouraged institutions
to fund risky mortgage securities with short-term
debt.6 As I will explain below, I believe that monetary
policy was not such a large culprit in creating the
housing bubble and the expansion in leverage.
I also believe that capital regulations set the stage for
domino effects and bank runs because the regulations
skewed incentives away from traditional mortgage
lending and toward securitization and risky financial
structures that incorporated mortgage securities.
Financial engineers created collateralized debt
obligations (CDOs), credit default swaps (CDSs), and
other esoteric products largely to exploit opportunities
for regulatory capital arbitrage. Compared with
traditional mortgage lending financed by deposits,
these financial instruments increased the financial
interdependence and vulnerability to runs of the
financial system.
For domino effects and bank runs, intuition may
suggest that a large role was played by changes to
industry structure due to mergers, acquisitions,
and the erosion of boundaries between investment
banking and commercial banking. The Obama
Administration’s white paper7 is among many analyses
that stress the significance of the growth of the
“shadow banking system.” This shadow banking
6. See, for example, John Taylor, Getting off Track: How Government Actions and Interventions Caused, Prolonged, and Worsened the
Financial Crisis (Stanford, CA: Hoover Institution Press, 2009).
7. U.S. Department of Treasury, Financial Regulatory Reform—A New Foundation, 2009.
Mercatus Center at George Mason University
11
system refers to off-balance-sheet entities (such as
SIVs) and portfolios of investment banks and other
non-bank institutions, which together amounted to
trillions of dollars.
However, much of what is now called “shadow banking”
emerged in response to capital regulations. The
consequent fragility of the financial system reflected
above all the risk allocation created by the structured
transactions and the leverage at individual institutions,
rather than new relationships between institutions
of different types. If we could conduct an
alternate history with capital regulations that did
not favor securitization and off-balance-sheet entities,
then the shadow banking system would not have
been an issue, and no crisis would have occurred.
Conversely, consider an alternate history where
institutions had to maintain a strict, Glass-Steagall
separation of commercial from investment banking
yet continued to operate under capital regulations
that blessed securitization, off-balance-sheet financing,
and other complex transactions. In that case, I
believe that the crisis would have unfolded pretty
much as it did.
Apart from practices that were developed for the
purpose of regulatory capital arbitrage, financial
innovation played a small role in the crisis. CDOs,
CDSs on mortgage securities, and SIVs are examples
of innovations that took advantage of regulatory
capital arbitrage. On the other hand, mortgage credit
scoring is an example of what I call an autonomous
innovation, meaning an innovation that was created
for reasons other than regulatory capital arbitrage. It
seems that overconfidence in credit scoring helped
fuel the bad bets in mortgage lending. However, on
the whole, most of the dangerous innovation seems
to have been driven by regulatory capital arbitrage.
4. PAST CRI SES MA KE BAD POLICY :
HOU SIN G POLICY AN D CA PITA L
R
EGULATION
Before proceeding to a more detailed look at the
evolution of policy in the five areas, it is worth pointing
out that housing policy and bank regulatory policy
evolved out of previous crises. The lesson is that
financial regulation is not like a math problem, where
once you solve it the problem stays solved. Instead, a
regulatory regime elicits responses from firms in the
private sector. As financial institutions adapt to regulations,
they seek to maximize returns within the
regulatory constraints. This takes the institutions in
the direction of constantly seeking to reduce the regulatory
“tax” by pushing to amend rules and by coming
up with practices that are within the letter of the
rules but contrary to their spirit. This natural process
of seeking to maximize profits places any regulatory
regime under continual assault, so that over time the
regime’s ability to prevent crises degrades.
The U.S. government made its first attempt to
reshape the mortgage market in the 1930s. When the
Great Depression hit, the typical mortgage loan was a
five-year balloon: The borrower paid interest only for
five years, at which point the entire mortgage came
due. The borrower either had to obtain a new loan
or pay off the existing loan. Under the Depression’s
circumstances of declining prices and incomes and
closing banks, many homes went into foreclosure. In
the absence of reliable deposit insurance, banks were
subject to runs, and thousands of banks closed.
In response to these problems, policy makers pressed
for two major reforms. One was the advent of the
thirty-
year fixed-rate mortgage, promoted by new agencies,
including the Federal Housing Administration
(FHA) and the Federal National Mortgage Association
(FNMA), which was created in 1938. Another was the
creation of federal deposit
insurance.
Fast forward forty years. From the late 1970s through
the late 1980s, the savings and loan industry in the
United States collapsed, with many institutions
becoming insolvent. Because the savings and loans
associations (S&Ls) were holding thirty-year, fixedrate
mortgages, their assets plummeted in value with
rising inflation and interest rates. Largely funded
with insured deposits, they had little incentive to
avoid taking risks, and indeed with deregulation they
made bad bets in a number of areas, including junk
bonds and commercial real estate, in a desperate
Not What They Had in Mind: A History of Policies that Produced the Financial Crisis of 2008
12
attempt to restore profitability. Thus, the combination
of thirty-year, fixed-rate mortgages and insured
deposits, which were the solutions to the 1930s mortgage
crisis, ended up producing the 1970s crisis.
Through the 1970s, banks and S&Ls were subject to
regulation Q, which placed ceilings on the interest
that these institutions could pay on various forms of
deposits. As a result of regulation Q, when inflation
and interest rates increased in the 1970s, the interest
rates on deposits were artificially low, causing savers
to seek higher returns elsewhere. The result was disintermediation,
in which depositors bypassed banks
and S&Ls for money market funds.
Disintermediation posed a dilemma for depository
institutions and their regulators. If regulators
did not lift the regulation Q ceilings, then the volume
of deposits would shrink. However, lifting the
ceilings would raise the cost of funds for banks and
S&Ls. Because their assets were long-term, fixedrate
mortgages, the S&Ls were in trouble with or
without regulation Q. With regulation Q, they lost
funds. Without regulation Q, they suffered a negative
spread between the earnings on their assets and the
cost of their liabilities.
Regulation Q ceilings were phased out in the early
1980s. At the same time, interest rates were at record
levels, as the Fed attempted to bring down inflation.
Holding thirty-year fixed-rate mortgages funded by
short-term deposits, the S&Ls were being squeezed
to death. Ultimately, many of the institutions were
closed, and taxpayers took losses of over $100 billion
in order to cover deposit insurance.
In the aftermath of the S&L crisis, policy makers
drew three conclusions. One was that securitization
of mortgages was better than traditional mortgage
lending. The thinking was that pension funds, insurance
companies, and other institutions with longterm
liabilities were better positioned to bear the
interest-rate risk associated with thirty-year fixedrate
mortgages than were banks and S&Ls that relied
on short-term deposits.
Another lesson of the S&L crisis was that regulators
should not rely on book-value accounting. By not marking
to market their economically depreciated mortgage
assets, S&Ls were able to stay in business even though
they were insolvent, taking on more risk and adding to
the ultimate cost of the taxpayer bailout.
A final lesson of the S&L crisis was that capital
requirements needed to be formal and based on risk.
Policy makers wanted private investors, not taxpayers,
to be the primary suppliers of risk capital to
banks. The concept of risk-based capital was embedded
in the Basel Accords in 1989, an international
set of standards adapted and implemented by bank
regulators in countries across the world, including
the United States.
Thus, the regulators responded to the S&L crisis by
promoting securitization, market-value accounting,
and risk-based capital, all of which contributed to or
exacerbated the most recent crisis. Mortgage securities
became the “toxic assets” at the core of the crisis.
Risk-based capital regulations promoted the use and
abuse of these instruments. The combination of riskbased
capital and market-value accounting served to
exacerbate both the boom and the bust.
During the crisis, risk-based capital and market-value
accounting contributed to domino effects. When a
bank was forced to sell mortgage-backed securities,
this lowered the market value of these securities,
triggering write-downs at other banks under market-
value accounting. This put other banks below
the regulatory minimum for capital.
This history suggests that as policy makers respond
to one crisis, their solutions can set the stage for the
next crisis. There is a significant difference between
hindsight and foresight, a fact that I wish to emphasize
when looking at the evolution of policy in the
five main areas: housing policy, capital requirements,
industry structure and competition, innovation, and
monetary policy.
In discussing each of these five policy areas, my goal
is to provide a historical narrative that explains how
Mercatus Center at George Mason University
13
the issues appeared to policy makers. What factors
made their decisions seem reasonable at the time?
What factors were overlooked? What lessons might
we learn?
5. HOU SIN G POLICY
Housing policy was close to the center of the financial
crisis. The U.S. government’s policy has been to
encourage as many people as possible to purchase
homes. The use of mortgage credit has been particularly
subsidized. The culmination of this policy was
a wild spiral of increasing home purchases, higher
home prices, and increased housing debt-to-equity
ratios, until these trends reached their limit and the
process went into reverse.
From 2000 to 2005, the total value of residential real
estate in the United States rose by 81 percent.8 The
total value of household mortgage debt rose even
faster.9 Over that same period, the GDP price index
for residential construction increased 29 percent.10
Thus, even after adjustment for changes in the cost of
construction, real-estate values and mortgage indebtedness
increased by more than 50 percent in just five
years. The home ownership rate, a politically salient
figure, reached 69 percent, up 5 percentage points
from a decade earlier.11
Between 2005 and 2008, household mortgage debt
continued to rise, by a total of 18 percent. However,
the value of residential real estate declined by 14
percent. As a result, over these three years the average
ratio of home equity to real-estate value plunged
from 58 percent to 43 percent.12
Policies that encouraged home ownership in the
past decade include: the mortgage interest deduction,
the capital gains tax exclusion, federal programs
that guarantee mortgage loans (such as the
Federal Housing Authority (FHA) and Veterans
Administration (VA)) and federal programs that
guarantee some liabilities of some mortgage lenders
(deposits of savings loans, debt and securities
of Freddie Mac and Fannie Mae), the Community
Reinvestment Act, and “affordable housing goals”
for Freddie Mac and Fannie Mae.
The mortgage interest deduction has been in place
since the income tax was first enacted in the United
States. It probably had its greatest impact in the
1970s, when marginal tax brackets and nominal
interest rates were higher than they are today. At the
margin, the mortgage interest deduction probably
played little role in encouraging the recent surge in
home ownership. Many of the marginal home buyers
had low income tax rates. For home buyers in
higher
tax brackets, the effect of the mortgage interest
deduction may have been to increase the demand
for larger and higher quality homes.
What the mortgage interest deduction may have
affected in recent years was the amount of debt consumers
were willing to have on their homes. The
tax deduction reduced the incentive of owners to
pay off or pay down their mortgages. By the same
token, it gave homeowners a reason to believe that
8. Board of Governors of the Federal Reserve System, Flow of Funds Accounts of the United States: Flows and Outstandings First Quarter
2005, 2005, table B100, line 4.
9. Ibid., line 33.
10. Bureau of Economic Analysis, National Income and Product Accounts Table: Table 1.1.4 Price Indexes for Gross Domestic Product, July
31, 2009, http://www.bea.gov/national/nipaweb/TablePrint.asp?FirstYear=2000&LastYear=2005&Freq=Qtr&SelectedTable=4&ViewSeries=
NO&Java=no&MaxValue=112.283&MaxChars=7&Request3Place=N&3Place=N&FromView=YES&Legal=&Land=.
11. U.S. Census Bureau, Housing Vacancies and Homeownership Table 15: Homeownership Rates of the United States, by Age of
Householder and by Family Status, 2005, http://www.census.gov/hhes/www/housing/hvs/annual05/ann05t15.html.
12. Federal Reserve Board, Flow of Funds Accounts, 2005 and Board of Governors of the Federal Reserve System, Flow of Funds Accounts of
the United States: Flows and Outstandings Fourth Quarter 2008, 2008.
Not What They Had in Mind: A History of Policies that Produced the Financial Crisis of 2008
14
1934
1986 1987
1938
1989
1954
1990 1992
1964
1993 1995 1997
1968
National
Housing Act
Emergency Low
Income Housing
Preservation Act
of 1987
Federal National
Mortgage Association
established by statute
Financial Institutions
Reform, Recovery,
and Enforcement Act
of 1989
Housing and Urban
Development Act of
Housing Act of 1964 1968, Title VIII
Housing Interim Goals
set for 1996–2000
Federal National
Mortgage Association
Charter Act (part of the
Housing Act of 1954)
Multifamily Housing
Finance Improvement Act
(part of the Housing and
Community Development
Act of 1992)
Federal Housing Enterprises
Financial Safety and Soundness
Act of 1992 (part of the Housing
and Community Development
Act of 1992)
Cranston-Gonzalez
National Affordable
Housing Act
Taxpayer Relief Act
Departments of Veterans
Affairs and Housing and
Urban Development and
Independent Agencies
Appropriations Act, 1997
Housing Interim Goals
set for 1993–1994
Tax Reform
Act of 1986
Figure 2: Housing Policy timeline
Mercatus Center at George Mason University
15
2000
1970 1971
2003 2004
1972 1973 1974 1975 1977 1979 1980 1983 1984
Housing Interim
Goals set for
2001–2004
American Dream
Downpayment Act
Housing and
Community
Development Act
of 1980
Housing and Urban-
Rural Recovery Act
of 1983
Community
Reinvestment
Act
Housing and
Community
Development
Act of 1977
Emergency
Home Finance
Act of 1970
Emergency
Housing Act of
1975
Ginnie Mae
authorized to
purchase mortgages
Housing Interim Goals
set for 2005–2008
President announces
mortgage assistance
plan
Secondary
Mortgage Market
Enhancement Act
of 1984
Minimum requirement
of loans to low-income
residents set
Emergency Home
Purchase Act of 1974
Housing and
Community Development
Act of 1974
Freddie Mac’s obligations
granted same standing as
government securities
Low-income
housing program
authorized
Office of Emergency
Preparedness abolished
President proposes
legislation to increase
homeownership
Not What They Had in Mind: A History of Policies that Produced the Financial Crisis of 2008
16
home equity loans were the cheapest form of credit
available, particularly after the deductibility of other
forms of consumer interest was ended in 1997.
The capital gains tax exclusion was changed in 1997.
Prior to that, homeowners over age 55 could exclude
up to $125,000 in capital gains on the sale of their primary
residences. Before age 55, a homeowner could
avoid capital gains tax by “rolling over” into a more
expensive home.
In 1997, this was changed to a straight exclusion of
$500,000 for married couples ($250,000 for single
individuals), regardless of age. Under some conditions,
second homes also could be eligible for this capital
gains tax exclusion. The more liberal capital
gains
tax exclusion rewarded housing speculators and thus
may have contributed to the housing bubble.
From the 1930s onward, mortgage lending was undertaken
by institutions whose liabilities were guaranteed
by the federal government. In addition to Fannie Mae,
which was chartered in 1938, there were the savings
and loans, which had federal deposit insurance.
By the late 1960s, restrictions on interstate banking
and regulation Q (which set regulatory ceilings on
the interest rates that thrifts could pay depositors)
created a shortage of mortgage funds in fast-growing
regions, particularly in California. Rather than fix
this problem by addressing the regulatory causes,
Congress chartered Freddie Mac to do what it had
forbidden the S&Ls to do: Raise funds in one part of
the country to finance mortgage lending elsewhere.
Freddie Mac created a secondary market in mortgages,
in which mortgages could be pooled together
and sold as securities.
In fact, the mortgage securities market was initially
a government-created phenomenon. In 1968,
Congress created the Government National Mortgage
Association (Ginnie Mae) to sell securities backed
by mortgages guaranteed through government programs
of the Federal Housing Administration (FHA)
and the Veterans Administration (VA). One purpose
was to get these mortgages off the books of the federal
government so that the administration would
not have to keep coming back to Congress to request
increases in the debt ceiling, for these requests
created
opportunities for Congress to express frustration
with the Vietnam War. As part of this process
of trying to trim the government’s balance sheet,
Fannie Mae was sold to private investors.
By the early 1980s, S&Ls needed a new source of
funds. They could not sell their mortgages without
incurring losses that would have exposed their
insolvency. Instead, with the approval of regulators,
investment bankers concocted a scheme under
which a savings and loan would pool mortgages into
securities that would be guaranteed by Freddie Mac.
The S&L would retain the security and use it as collateral
to borrow in the capital market. However,
unlike an outright sale of the mortgages, the securitized
mortgage
transaction would not trigger a
write-down of the mortgage assets to market values.
The accounting treatment of mortgage securities, in
which they were maintained at fictional book-market
values, enabled the S&Ls to keep a pretense of viability
as they borrowed against their mortgage assets.
Fannie Mae soon joined Freddie Mac in undertaking
these transactions.
Thus, from the 1960s through the early 1980s, mortgage
securitization was driven largely by anomalies in
accounting treatment and regulation. Ginnie Mae was
developed in order to move mortgages off the government’s
books, even though government was still providing
guarantees against default. Congress created
Freddie Mac to work around the problems caused
by regulation Q and interstate banking restrictions.
And the growth in securitization by Freddie Mac and
Fannie Mae was fueled by the desire of regulators to
allow S&Ls to raise funds using their mortgage assets
without having to recognize the loss in market value on
those assets. Mortgage securitization did not emerge
organically from the market. Instead, it was used by
policy makers to solve various short-term problems.
Securitization failed to prop up the S&L industry.
When that industry collapsed, Freddie Mac and Fannie
Mae were poised to dominate the housing finance
Mercatus Center at George Mason University
17
market. They did so from the late 1980s until the late
stages of the homeownership boom. By 2003, Freddie
and Fannie together held half of all mortgage debt outstanding.
However, from 2003 through 2005, many
buyers could not qualify for the “investment quality”
mortgages that Freddie and Fannie were focused on
purchasing. Consequently, the market share of these
GSEs actually declined over this period. The GSEs
became much more active in the subprime market in
2006 and 2007, in part to try to recover market share.
5.A. CRA and the Under-Served
Housing Market
In 1995, Congress revised the Community
Reinvestment Act (CRA), first enacted in 1977, to give
banks a stronger impetus to raise the portion of consumer
loans (including mortgages) going to low-income
borrowers. Both the Clinton Administration and
the Bush Administration also gave Freddie Mac and
Fannie Mae quotas for supporting low-income housing.
In order to meet these quotas and to try to stop
the erosion in market share, the GSEs set aside some
of their “investment quality” requirements and found
ways to participate in the subprime mortgage market.
Many mortgage loans that met the standards for CRA
were of much higher quality than the worst of the
mortgage loans that were made from 2004–2007.
Thus, one must be careful about assigning too much
blame to CRA for the decline in underwriting standards.
It is possible that, even in the absence of CRA,
many lenders would have pursued the market for
low-quality mortgages simply in pursuit of profits.
Careful research would be needed in order to determine
the marginal impact of CRA.
In the mortgage market as a whole, the quality of
loans deteriorated along many dimensions:
• The share of loans for non-occupant owners
(speculators) rose from 5 percent in the
early 1990s to 15 percent in 2005 and 2006.
Moreover, official data may understate the
growth in housing speculation since a buyer of
an investment property may claim an intent to
occupy the home when she applies for a loan.
• The loan products became riskier. More loans
were adjustable-rate loans with low initial
“teaser” rates. A number of loan products incorporated
features that reduced or eliminated the
automatic amortization of principal.
• Down payment requirements were loosened.
Loans with down payments of 3 percent, 2 percent,
or even zero became common. Borrowers
were allowed to take out “refinance” loans for
100 percent of the appraised value of their
homes (and sometimes even more).
• Lenders waived requirements that borrowers
document their incomes, assets, and employment
information on their mortgage applications.
In traditional mortgage lending, borrowers were
asked to provide proof of income, employment, and
assets. The lender might call the company where the
borrower worked to verify employment. The borrower
might be asked to supply pay stubs to verify
income. And the borrower might be asked to supply
bank statements to verify assets.
Most of the time, this documentation was redundant.
Mortgage originators, trying to compete for business
by offering greater convenience, would try to make
exceptions to the documentation requirements. They
then would negotiate with Freddie Mac and Fannie
Mae to allow these exceptions.
For the vast majority of mortgage loans, reduced
documentation saved on costs without any adverse
effect on loan quality. However, a program of
reduced documentation becomes a magnet for fraud.
Under such programs, swindlers operating as mortgage
originators can concoct remarkable schemes
to sell mortgage loans and abscond with millions of
dollars. The GSEs experienced this sort of fraud in
the late 1980s, and that is why in 1990, when a trend
toward reduced documentation of mortgage loans
was building, Freddie Mac and Fannie Mae issued a
joint policy against purchasing “low-doc” loans. For
a time, this put a halt to the trend.
Not What They Had in Mind: A History of Policies that Produced the Financial Crisis of 2008
18
However, fifteen years later, another move toward
“low-doc” lending emerged. The newer “NINJA”
loans (“no income, no job, no assets”) were motivated
less by a desire to provide convenience to ordinary
borrowers and more by a desire to reach out to
new borrowers by focusing on housing appreciation
and credit scores as the primary tools for controlling
credit risk. This time, the GSEs were not able to take
a stand against the dangerous trends in mortgage
origination. Their market shares had been eroded
by private-label mortgage securitization. They were
under pressure from their regulators to increase their
support of low-income borrowers. Finally, they had
been stained by accounting scandals in which they
had allegedly manipulated earnings, so that they had
little political capital to throw into a fight to maintain
underwriting standards.
The weakening of mortgage credit standards was
destabilizing for the housing market. This was particularly
the case with the trend toward lower down
payments and innovative mortgage designs that
required less repayment of principal. As a result,
many homeowners relied on house price appreciation
for the equity in their homes. As long as prices
were rising, home purchases could be sustained at
high levels, including speculative purchases and
homes that were too expensive for the borrowers to
afford. Once prices stopped rising, however, there
was no equity cushion to prevent defaults and foreclosures,
so that a rapid and severe downward spiral
took place.
At the time that mortgage credit quality was deteriorating,
the main regulatory concern was with consumer
protection. Those who had this concern, such
as Edward Gramlich of the Federal Reserve Board,
thought that lenders were exploiting consumers by
providing loans that were dangerous, costly, and
poorly understood by borrowers.
The danger to financial firms of poor mortgage credit
quality went largely unnoticed. However, the issue
was raised in an article written by FDIC economist
Cynthia Angell in 2004. She concluded:
In summary, because home prices have
appreciated briskly over the past several
years and outpaced income growth, concerns
have been voiced about the possibility
of a nationwide home price bubble.
However, it is unlikely that home prices
are poised to plunge nationwide, even
when mortgage rates rise. Housing markets
by nature are local, and significant price
declines historically have been observed
only in markets experiencing serious economic
distress. Furthermore, housing markets
have characteristics not inherent in
other assets that temper speculative tendencies
and generally mitigate against price collapse.
Because most of the factors affecting
home prices are local in nature, it is highly
unlikely that home prices would decline
simultaneously and uniformly in different
cities as a result of some shift such as a rise
in interest rates.
The greater risk to insured institutions is
the potential for increased credit delinquencies
and losses among highly leveraged,
subprime, and ARM borrowers. These
high-risk segments of mortgage lending may
drive overall mortgage loss rates higher
if
home prices decline or interest rates rise.
Credit losses may, in turn, spill over to
nonmortgage consumer credit products
if households prioritize debt repayment
to give preference to mortgage payment.
Residential construction lending in markets
where there is significant speculative
building, as well as an abundance of thinly
capitalized builders, also may be of concern,
especially when the current housing boom
inevitably cools.13
13. Cynthia Angell, “Housing Bubble Concerns and the Outlook for Mortgage Credit Quality,” FDIC Outlook, February 2004, http://www.
fdic.gov/bank/analytical/regional/ro20041q/na/infocus.html/.
Mercatus Center at George Mason University
19
After this was published, home prices continued
climbing for nearly three years. Mortgage credit quality
deteriorated further. However, regulators did not
focus on the potential impact for the financial system.
The common assumption was that profit-driven
financial institutions knew what they were doing.
As noted above, regulatory concern with mortgage
origination practices was largely limited to worries
about individual borrowers not understanding the
risks they were assuming. In any case, regulators did
little or nothing about even these latter worries.
With homeownership rising, household wealth
increasing, and financial sector profits robust, policy
makers were much more inclined to view mortgage
trends as benign rather than as a threat. The overall
policy of encouraging home purchases with mortgage
debt seemed to be working, and it had powerful
support from the various interest groups that benefited
from the boom.
In hindsight, the government had an opportunity to
avert the crisis by changing housing policy in 2003 or
2004. It could have forced Fannie Mae, Freddie Mac,
and banks to hold more capital to back their expansion
into subprime mortgage loans. Better yet, regulators
could have recognized the risks of trying to expand
home ownership to weaker and weaker borrowers
in an environment of high house prices. Instead of
encouraging the GSEs and the banks to make more
loans to low-income borrowers, the regulators could
have leaned on those firms to maintain prudent lending
standards, particularly for down payments.
Regulators, like their private-sector counterparts,
failed to imagine the potential financial cataclysm
that was developing in the mortgage market. Even
if they could have envisioned the scenario of a bursting
of the housing bubble and anticipated the consequences
for institutions involved in the mortgage
financing system, regulators would have had to convince
politicians of the validity of their concerns.
Former Federal Reserve Board Chairman William
McChesney Martin once described the Fed’s job as
“taking away the punchbowl just when the party is
getting good.” From a political perspective, a regulatory
crackdown on loose mortgage underwriting
standards in 2004 would have meant taking away a
punch bowl filled with more home ownership—particularly
among minorities—as well as expansion and
profits in the businesses of home building, real estate
brokerage, mortgage origination, and Wall Street
financial engineering. Whether the political process
would have accepted taking away that punch bowl
is questionable.
To the extent that there was a trade-off between
expanding the availability of mortgage credit and
maintaining safety and soundness, the political pressure
appeared to be toward expanding credit availability
as opposed to worrying about safety and soundness.
This can be seen in the way that Congress rejected
efforts by both the Clinton and Bush Administrations
to restrain the growth of Fannie Mae and Freddie
Mac. Various economists, including a group calling
itself the Shadow Regulatory Committee, were worried
by the rapid growth of the GSEs, but, for the most
part, these economists expressed fears that the GSEs
would take on too much interest-rate risk. Credit risk,
which proved to be their downfall, was not the focus
of much concern.
14
The housing lobby has been one of the most powerful
coalitions in Washington. It includes real-estate
agents, community action groups that advocate for
expanded home ownership, home builders, mortgage
originators, mortgage financing firms, and securities
trading firms—all interest groups that benefit from
expanding the demand for housing and for mortgage
loans. When it came to mortgage lending, the
political pressure on policy makers all went in one
direction—for more subsidies and fewer restrictions.
Thus while in theory, the most logical and straightforward
way to avert the financial crisis would have
The 14. GSEs take credit risk when they guarantee mortgage securities against any defaults on the underlying mortgages. They take interestrate
risk when they themselves hold mortgage securities in portfolio. It was curbs on the size of the GSEs’ security portfolios that economists
both inside and outside the Clinton and Bush Administrations sought.
Not What They Had in Mind: A History of Policies that Produced the Financial Crisis of 2008
20
1931 1936 1975 1979 1986 1988
Ratings assessments
used in bank portfolios
Bank purchases
restricted based on
ratings
Guidelines on
managing banks’
off-balance-sheet
exposures
The consolidated
supervision of banks’
international activities
Rating agencies’ judgments
used in regulatory decisions
(Basel I ) International
convergence of capital
measurement and
capital standards
Figure 3. Ch ange s to Capi tal Rules timeline
Mercatus Center at George Mason University
21
1994 1995 1996 2001 2004 2006 2007 2008 2009
Credit Rating
Agency Reform
Act of 2006
BCBS consultative
document on
computing capital
Ratings-based
approach applied to
asset- and mortgagebacked
securities
U.S. moved
further to adopt
Basel standards
Supervisory guidance
for credit derivatives
Changes made to capital
standards of Special
Purpose Vehicles
U.S. regulators
attempt to further
incorporate Basel
standards
Revisions
to Basel II
market risk
framework
(Basel II) International
convergence of capital
measurement and capital
standards: a revised
framework
Some broker dealers permitted
alternative means to
compute capital
Treatment of credit
risk associated with
certain off-balancesheet
items
Amendment to the
1988 capital accord
to broaden
the recognition of
collateral
Not What They Had in Mind: A History of Policies that Produced the Financial Crisis of 2008
22
been to adjust housing policy, in practice, the political
landscape made such an approach very unlikely
to be attempted.
6. BAN K CA PITA L REGULATION S
The most important regulatory failure contributing
to the financial crisis was in the arena of safety
and soundness. Bank capital regulations were the
primary culprit. In addition, regulators permitted
Fannie Mae, Freddie Mac, AIG, and many investment
banks to take too much risk with too little capital.
In fact, it will be seen below that the risk-based bank
capital regulations had perverse effects. The regulations
created an incentive for banks to take highly
levered positions in securities backed by risky mortgage
loans.
The financial tactics that ultimately were at the heart of
the financial crisis emerged in order to achieve regulatory
capital arbitrage—gaming the system in order to
minimize capital while retaining risk. These tactics
included securitization, off-balance-sheet financing,
the use of credit derivatives such as credit default
swaps, and the reliance on ratings of credit agencies.15
The capital requirements were part of a regime
known as the Basel Accords. The problems with
the Basel regulations, and especially with the use
of credit rating agencies, were anticipated by many
economists. In particular, the Shadow Regulatory
Committee, a group of economists offering independent
opinion on bank regulation, issued timely and
accurate criticisms of the approach that regulators
were taking toward capital regulation.
By incorporating Nationally Recognized Statistical
Rating Organization (NRSRO) ratings into formal
capital requirements, bank regulators effectively outsourced
critical oversight functions to the credit rating
agencies.16 However, as it turned out, the credit
rating agencies did not serve well the regulators’ purpose.
Instead, they rated mortgage-backed securities
too generously, under assumptions about house prices
that were too optimistic. This problem was foreseen
by critics at Fannie Mae and in the Shadow Regulatory
Committee, who pointed out that when securities
were being rated for regulatory purposes rather than
for trading purposes, the rating agencies would face
less market incentive to rate conservatively.
The Basel Accords were created in stages. The first
stage was the initial agreement, which was issued
in 1988. The latest stage, known as Basel II, was
scheduled
to be implemented in the United States in
2008. In between, there were a number of modifications
to Basel I. Some of the modifications had a
significant impact on the treatment of mortgages and
mortgage securities.
15. The regulatory use of credit rating agencies dates back to the 1930s. Flandreau, et al., pointed out that
In the midst of a wave of defaults and plummeting bond prices in 1931, the OCC instituted formulae based on credit ratings to book the
value of US national banks’ bond portfolios. The role of rating agencies was extended in 1936 when the OCC restricted the purchase by
banks of securities with lower credit ratings.
[In September of 1931], time bond prices were plummeting in the wake of the German financial crisis and a run on Sterling. The OCC
ruling was reported to state that all Federal, State, and Municipal U.S. securities, as well as other domestic and foreign securities belonging
to any of the top four categories of ratings, could be booked by banks at face value (Harold 1938), while other securities and
defaulted bonds should continue to be marked to market.
Marc Flandreau, Norbert Gaillard, and Frank Packer, “Ratings Performance, Regulation and the Great Depression: Lessons from Foreign
Government Securities,” CEPR Discussion Paper 7328, 2009, http://www.graduateinstitute.ch/webdav/site/iheid/shared/publicationsNEW/
publications_GCI/working_paper_ratings_gci.pdf/.
16. In 1975, the Securities and Exchange Commission designated a small, select subset of these credit rating agencies as Nationally
Recognized Statistical Rating Organizations (NRSROs). In recent years, the only NRSROs were Moody, Fitch, and Standard and Poor.
Mercatus Center at George Mason University
23
The initial Basel agreement called for banks to hold
8 percent capital against risk-weighed assets. At
least half of this capital had to consist of equity or
published reserves. The rest could be in undisclosed
reserves, preferred stock, subordinated debt, and
other categories.
The risk weights of assets were as follows:
• Claims on OECD governments and central
banks had zero risk weight. At the margin,
these assets required no capital.
• Claims on other OECD public-sector entities
(such as U.S. state governments or Fannie
Mae and Freddie Mac) and short-term claims
on banks had a 20 percent risk weight. At the
margin, these assets required (.08)(.20) = 1.6
percent capital.
• All home mortgages, regardless of risk characteristics,
carried a 50 percent risk weight.
At the margin, mortgages required 4 percent
capital.
• All other assets, including ordinary commercial
loans, had a 100 percent risk weight. At the margin,
these assets required 8 percent capital.
Among other effects, these risk weights created an
advantage for mortgage securitization because the
bank capital standards for low-risk mortgage loans
were overly onerous while Freddie Mac and Fannie
Mae faced lower capital standards.17 Recall that the
Basel agreement created an effective 4 percent capital
requirement (2 percent tier one or equity capital)
for all mortgages, regardless of risk. However,
for mortgage securities guaranteed by Freddie Mac
or Fannie Mae, the capital requirement would have
been 1.6 percent (0.8 percent tier one). Thus, it was
capital-efficient to securitize mortgage loans with
Freddie Mac or Fannie Mae.
The late 1990s saw the emergence of collateralized
debt obligations (CDOs). These enabled mortgage
securities to be deemed low risk for capital purposes,
even though they were not guaranteed by Freddie
Mac or Fannie Mae. These so-called “private label”
securities now became eligible for regulatory capital
arbitrage. The financial engineers carved CDOs
into tranches, with junior tranches bearing the risk
of the first loans to default, insulating senior tranches
from all but the most unlikely default scenarios. Once
regulators endorsed the use of credit rating agency
evaluations, CDO tranches could earn high ratings,
which meant low capital requirements. At that point,
private-label securitization really took off.
Capital requirements could be reduced further
by moving CDOs off a bank’s balance sheet into
a structured investment vehicle (SIV). As long as
the bank only offered a short-term line of credit
(less than one year) to the SIV, the assets of the SIV
did not have to be included in the calculation of capital
requirements.
The phenomenon of regulatory capital arbitrage
was well understood by the Federal Reserve Board.
Although papers in academic journals written by
Federal Reserve Board employees routinely carry a
disclaimer that they do not represent the opinions of
the board or its staff, a paper published in 2000 by
17. Economists Paul Calem and Michael Lacour-Little calculated capital requirements for banks to have a BBB solvency standard. Using this
approach, they pointed out,
newly originated loans with 80 percent loan-to-value ratios and a prime borrower credit score of 700 require very little capital to cover
credit risk: no more than 0.51 percent in a well-diversified portfolio and 0.90 percent in a regionally concentrated portfolio, assuming a
BBB solvency standard and an eight year horizon.
. . . current rules may encourage regulatory capital arbitrage, including increased rates of securitization of mortgage assets.
Paul S. Calem and Michael Lacour-Little, “Risk-Based Capital Requirements for Mortgage Loans,” (FEDS Working Paper no. 2001-60,
November 2001), 3, http://papers.ssrn.com/sol3/papers.cfm?abstract_id=295633/.
Not What They Had in Mind: A History of Policies that Produced the Financial Crisis of 2008
24
Fed researcher David Jones provides clear evidence
that the Fed knew that regulatory arbitrage relative
to capital requirements was taking place. Moreover,
the tone of the paper was generally sympathetic to
the phenomenon.
In recent years, securitization and other
financial innovations have provided
unprecedented opportunities for banks to
reduce substantially their regulatory measures
of risk, with little or no corresponding
reduction in the overall economic risks—a
process
termed “regulatory capital arbitrage”
(RCA).
. . . Ultimately, RCA is driven by large
divergences that frequently arise between
underlying economic risks and the notions
and measures of risk embodied in regulatory
capital ratios. As discussed below,
such divergences create opportunities to
unbundle and repackage a portfolio’s risks
in ways that can reduce dramatically the
effective capital requirement per dollar of
economic risk retained by a bank. Efforts to
stem RCA without narrowing or eliminating
these divergences—for example, by limiting
banks’ use of securitization and other risk
unbundling technologies—would be counterproductive
and perhaps untenable. In
some circumstances, RCA is an important
“safety-valve” that permits banks to compete
代写留学生论文effectively (with nonbanks) in lowrisk
businesses they would otherwise be
forced to exit owing to unreasonably high
regulatory capital requirements. Moreover,
as evidenced through their widespread use
by nonbanks, securitization and other risk
unbundling technologies appear to provide
genuine economic benefits to banks,
quite apart from their role in RCA. Lastly,
the same shortcomings giving rise to RCA
under the Accord also distort bank behavior
in other ways, such as discouraging the true
hedging of economic risks.
. . . when capital standards are not based
on any consistent economic soundness
standard (e.g., probability of insolvency),
through securitization and other techniques
it is often possible to restructure portfolios
to have basically similar risks, but much
lower regulatory capital requirements.
. . . Federal Reserve staff have estimated the
outstanding (non-mortgage related) ABSs
[asset-backed securities] and ABCP [assetbacked
commercial paper] issued through
programs sponsored by the 10 largest US
bank holding companies. Even excluding
mortgage securitizations, these estimates
reveal that the securitization activities of
these companies loom large in relation to
their on-balance sheet exposures. As of
March 1998, outstanding non-mortgage
ABSs and ABCP issuance through securitization
programs sponsored by these
institutions
exceeded US$200 billion, or
more than 25% of the institutions’ total riskweighted
loans.
. . . Since the underlying securitized assets
tend to be of relatively high quality, a strong
case can be made that the low capital
requirements against these retained risks
actually may be appropriate.
. . . Unless these economic and regulatory
measures of risk are brought into closer
alignment, the underlying factors driving
RCA are likely to remain unabated. Without
addressing these underlying factors, supervisors
may have little practical scope for
limiting RCA other than by, in effect, imposing
more or less arbitrary restrictions on
banks’ use of risk unbundling and repackaging
technologies, including securitization
and credit derivatives.
Such an approach, however, would be counterproductive
(and politically unacceptable).
Mercatus Center at George Mason University
25
. . . By reducing banks’ effective capital requirements
against such activities to levels more
consistent with the underlying economic
risks, RCA may permit banks to compete
efficiently in relatively safe businesses they
would otherwise be forced to abandon.18
In essence, the author argued:
• The Basel risk buckets were arbitrary.
• The risk classifications may have been overly
conservative for certain types of loans.
• Regulatory Capital Arbitrage (RCA) enabled
banks to reduce the capital requirements for
these loans.
• RCA was difficult to stop politically.
• RCA did not necessarily harm safety and soundness
if it kept banks competitive in markets to
make low-risk loans.
What is striking about the paper is the degree to
which the regulator shows understanding and support
for the banks’ use of securitization and off-balance-
sheet entities to reduce capital requirements.
Because we know what happened subsequently (the
paper was published in 2000), reading the Jones
paper is like watching a movie in which we see how
a jailer becomes sympathetic to the plight of a prisoner,
while we know that eventually the prisoner is
going to escape and go on a vicious crime spree.
A key modification of the Basel regulations was developed
from 1997–2001 and put into place by U.S. banking
regulators with an effective date of January 1,
2002. This new rule broadened the definition of lowrisk
securities to include securities rated double-A or
higher by NRSROs.19 This meant that they had a risk
weight of 20 percent, which put them on par with
securities issued by Freddie Mac or Fannie Mae. This
in turn drew the attention of the GSEs, which recognized
that their competitive role could be undermined
by the more lenient bank capital requirements.
In a comment on the proposed rules, Freddie Mac
showed what would happen to the capital requirement
on a representative structured financing of a
$100 million pool of mortgages owned by the bank.
Recall that under the original Basel agreement, the
capital requirement would be $4 million ($100 million
times a 50 percent risk weight times the 8 percent
capital requirement).
FIGUrE 4: change s in capi tal re quire ments
Mortgage
tranche rating
and supp ort level
Capi tal require ments
AAA $94 million $1.504 million (1.6%)
AA $ 2 million $.032 million (1.6%)
A $ 2 million $0.080 million (4%)
BBB $ 1 million $0.080 million (8%)
BB $ 0.5 million $0.080 million (16%)
Unrated $0.5 million $0.5 million (gross-up)
TOTAL $100 million
$2.276 million (vs. $4 million
unsecuritized)
Source: Memorandum from Freddie Mac, June 7, 2000
Thus, the new rule dramatically lowered the capital
banks needed in order to hold mortgage assets. For
mortgages, the rule had the exact same effect as lowering
the generic capital requirement from 8 percent
to something closer to 4.5 percent.20
18. David Jones, “Emerging problems with the Basel Capital Accord: Regulatory capital arbitrage and related issues,” Journal of Banking and
Finance, 2000, 35–58.
19. See Michael J. Zamorski, “Final Rule to Amend the Regulatory Capital Treatment of Recourse Arrangements, Direct Credit Substitutes,
Residual Interests in Asset Securitizations, and Asset-Backed and Mortgage-Backed Securities,” November 29, 2001, http://www.fdic.gov/
news/news/financial/2001/fil0199.html/.
20. Memorandum from Freddie Mac to the bank regulatory agencies, June 7, 2000. Reproduced in Corine Hegland, “Why the Financial
System Collapsed,” National Journal, April 11, 2009, http://www.nationaljournal.com/njmagazine/cs_20090411_7855.php/.
Not What They Had in Mind: A History of Policies that Produced the Financial Crisis of 2008
26
Fannie Mae offered similar examples. In addition, it
pointed out that the new rules would create incentives
to undermine the integrity of NRSRO ratings.
Banks would shop for ratings. Moreover, if the securities
were not traded, and instead were only rated
for regulatory purposes, then the NRSROs would
have little incentive to worry about the reputations
of their ratings.
The criticisms made by the GSEs might have been
dismissed as self-serving. Protecting their own
advantages in terms of low capital requirements
was critical to maintaining the franchise value of
Freddie Mac and Fannie Mae. However, the Shadow
Regulatory Committee—a group of market-friendly
economists offering independent opinion on bank
regulation and no friend of the GSEs, which the committee
thought were far too large and excessively
exposed to risk—weighed in with similar concerns.
Referring to a Basel Committee proposal along the
lines of the U.S. regulators’ proposal, the Shadow
Regulatory Committee’s statement number 160,
written in March of 2000, said in part,
the use of private credit ratings to measure
loan risk may adversely affect the quality
of ratings. If regulators shift the burden of
assessing the quality of bank loans to ratings
agencies, those regulators risk undermining
the quality of credit ratings to investors.
Ratings agencies would have incentives to
engage in the financial equivalent of “grade
inflation” by supplying favorable ratings to
banks seeking to lower their capital requirements.
If the ratings agencies debase the
level of ratings, while maintaining ordinal
rankings of issuers’ risks, the agencies may
be able to avoid a loss in revenue because
investors still find their ratings useful . . . In
short, if the primary constituency for new
ratings is banks for regulatory purposes
rather than investors, standards are likely
to deteriorate.21
In this instance, events proved the Shadow Regulatory
Committee correct. The rating agencies, undisciplined
by investors and seeking only to meet the demands of
banks, who in turn were motivated solely by the desire
to reduce regulatory capital, were generous with their
AAA and AA ratings. The optimism in the ratings
emerged as a central scandal of the financial crisis.
The 2002 rule thus had several deleterious effects.
First, it created opportunities for banks to lower their
ratio of capital to assets through structured financing.
Second, it created the incentive for rating agencies
to provide overly optimistic assessment of the risk in
mortgage pools. Finally, the change in the competitive
environment adversely affected Freddie Mac and
Fannie Mae, which saw their market shares plummet
in 2004 and 2005. The GSEs responded by lowering
their own credit standards in order to maintain a
presence in the market and to meet their affordable
housing goals. Thus, the 2002 rule unleashed the final
stages of the mortgage boom: the expansion in privatelabel
securities and subprime lending.
The drive to hold mortgage assets backed by as little
capital as possible proceeded well beyond the initial
structured finance mechanisms sketched in the table
above. Other tactics for minimizing regulatory capital
included:
• bundling and re-bundling mortgage-backed
securities (Wall Street terminology included
“CDO” for “collateralized debt obligation” and
“CDO-squared” for a CDO collateralized by
CDOs);
• “renting” AIG’s triple-A rating by obtaining
credit default swaps from that insurance
company;
and
• putting mortgage-backed securities into offbalance-
sheet entities called special purpose
vehicles (SPVs) and structured investment
vehicles (SIVs).
21. Shadow Financial Regulatory Committee, “Reforming Bank Capital Regulation,” statement number 160, March 2, 2000, http://www.aei.
org/docLib/20051114_ShadowStatement166.pdf/. statement number 160, http://www.aei.org/article/16542/.
Mercatus Center at George Mason University
27
Supposedly, the off-balance-sheet entities were selfcontained,
primarily relying on commercial paper
for funding. However, once investors lost confidence
in the soundness of the underlying assets, they were
no longer willing to invest in the commercial paper.
The banks were obligated (or at least felt obligated)
to put the assets in these entities back onto their
books. This damaged the banks in terms of liquidity,
because short-term funding for mortgage-backed
securities was no longer available. It also damaged
them in terms of capital adequacy, because the assets
now counted against their capital requirements.
After the crisis, the Financial Accounting Standards
Board (FASB) took steps to change the treatment of
special purpose vehicles.22
In hindsight, one wonders how the banks were able
to obtain regulatory permission to move mortgage
securities off their balance sheets, effectively evading
capital requirements altogether. In view of the
fact that banks later took possession of these assets,
it is clear in retrospect that the banks had not offloaded
the risk of those mortgage securities.
Regulators were thinking that the original Basel
rules were keeping banks from expanding their holdings
of mortgage assets, which regulators viewed as
relatively safe. The regulators were concerned with
the rigidity of the Basel rules and the slow pace at
which these could be changed. As a result, regulators
had to choose between giving the SPVs and SIVs
on-balance-sheet treatment, under which the riskbucket
approach would have demanded too much
capital (or so it was thought at the time) or giving
them off-balance-sheet treatment, which demanded
no capital.
Step by step, innovation by innovation, the process
of regulatory arbitrage became more efficient.
Financial engineers squeezed more and more assets
into banks with less and less required regulatory capital.
Investors who purchased the securities issued
by banks, Fannie Mae, Freddie Mac, and other companies
saw attractive returns on liquid assets that
apparently carried no risk. However, behind these
securities were risky, long-term mortgages without
a sufficient capital cushion.
What emerged was a highly leveraged financial
structure that was vulnerable to an adverse shift in
the housing market. When some mortgage securities
firms ran into trouble in 2007 due to excessive loan
defaults, investors took steps to assess and then limit
their exposure to mortgage assets. The commercial
paper market for the banks’ off-balance-sheet entities
collapsed. The holders of credit default swaps
from AIG demanded collateral in the form of shortterm,
risk-free assets.
In fact, the whole dynamic of the financial sector
went into reverse. Financial institutions had been
loading up on long-term, risky assets, while issuing
short-term liabilities and minimizing on capital.
Now, every institution needed to boost its liquidity
and its capital position, and few firms were interested
in buying mortgage securities.
In hindsight, many observers have faulted the rise
of the “shadow banking system,” meaning the various
investment banks and off-balance-sheet entities
that became involved in mortgage finance. However,
at the time, most regulators were pleased with the
way that mortgage credit risk was allocated by these
transactions. For example, the annual report of the
International Monetary Fund in 2006 stated that
financial innovation “has helped to make the banking
and overall financial system more resilient.”23 At the
time, in the view of many regulators, securitization
and credit derivatives helped to disperse risk in ways
that made the financial market safer.24
22. See Binyamin Appelbaum, “Board to Ban Accounting Practice That Helped Lending Proliferate,” The Washington Post, May 18, 2009,
http://www.washingtonpost.com/wp-dyn/content/article/2009/05/17/AR2009051701779.html/.
23. International Monetary Fund, Annual Report of the Executive Board for the Financial Year Ended April 30, 2006, August 3, 2006, 11.
24. See Gillian Tett, Fool’s Gold: How the Bold Dream of a Small Tribe at J.P. Morgan Was Corrupted by Wall Street Greed and Unleashed a
Catastrophe (New York: Free Press, 2009).
Not What They Had in Mind: A History of Policies that Produced the Financial Crisis of 2008
28
Another key policy maker, Federal Reserve Chairman
Ben Bernanke, said in June of 2006:
The evolution of risk management as a
discipline has thus been driven by market
forces on the one hand and developments in
banking supervision on the other, each side
operating with the other in complementary
and mutually reinforcing ways. Banks and
other market participants have made many
of the key innovations in risk measurement
and risk management, but supervisors have
often helped to adapt and disseminate best
practices to a broader array of financial
institutions.
. . . The interaction between the private and
public sectors in the development of riskmanagement
techniques has been particularly
extensive in the field of bank capital
regulation, especially for the banking organizations
that are the largest, most complex,
and most internationally active.
. . . Moreover, the development of new technologies
for buying and selling risks has
allowed many banks to move away from the
traditional book-and-hold lending practice
in favor of a more active strategy that seeks
the best mix of assets in light of the prevailing
credit environment, market conditions,
and business opportunities. Much more
so than in the past, banks today are able to
manage and control obligor and portfolio
concentrations, maturities, and loan sizes,
and to address and even eliminate problem
assets before they create losses. Many banks
also stress-test their portfolios on a business-
line basis to help inform their overall
risk management.
To an important degree, banks can be more
active in their management of credit risks
and other portfolio risks because of the
increased availability of financial instruments
and activities such as loan syndications,
loan trading, credit derivatives, and
securitization. For example, trading in
credit derivatives has grown rapidly over
the last decade, reaching $18 trillion (in
notional terms) in 2005. The notional value
of trading in credit default swaps on many
well-known corporate names now exceeds
the value of trading in the primary debt
securities of the same obligors.25
Thus, regulators were well aware of the innovations
in credit risk management. However, they viewed
these developments with sympathy and approval.
In retrospect, given the failure of the Basel regime,
what might have worked better? The Shadow
Regulatory Committee warned of flaws in the
approach to safety and soundness embodied in the
Basel capital standards even before the first version of
those standards became official in 1989.26 In a number
of statements that the Shadow Regulatory Committee
issued in the early 1990s, it recommended the use
of subordinated debt as an alternative to the Basel
approach of trying to manage safety and soundness
by classifying assets according to regulators’ determination
of risk.27 The idea behind requiring banks
to issue subordinated debt is that creditors would
require interest rates based on their perception of
the risk of the bank. The size of this risk premium
would in turn provide a market signal to regulators of
Ben 25. Bernanke, “Modern Risk Management and Bank Supervision” (speech at the Stonier Graduate School of Banking, June 12, 2006)
http://www.federalreserve.gov/newsevents/speech/Bernanke20060612a.htm/.
26. In May 1987, the economists warned that fixed risk weights for assets would distort credit allocation while failing to protect deposit insurance
funds. Shadow Financial Regulatory Committee, “Regulatory Proposals for Risk-Related Capital Standards,” statement no. 18, May 1987,
http://fic.wharton.upenn.edu/fic/Policy%20page/20051230_18%5B1%5D.pdf.
27. See for example, Shadow Financial Regulatory Committee, “Mergers and Acquisitions in the Banking Industry,” statement no. 147, May
4, 1998, http://www.aei.org/docLib/20051114_ShadowStatement147.pdf.
Mercatus Center at George Mason University
29
where to look for problems. Moreover, the subordinated
debt would provide an additional layer of protection
for taxpayers. Many economists continue to
believe that subordinated debt would be useful. For
example, economists Susan Woodward and Robert
Hall expressed support for a proposal made by a team
of banking experts called the Squam Lake Working
Group for subordinated debt that could convert automatically
to equity in a systemic crisis for banks that
fell below their capital requirements.
28
Both the Squam Lake Working Group and the
Shadow Regulatory Committee recognize that
much of the challenge in bank regulation involves
what economists call “the time inconsistency problem.”
That is, prior to a crisis, regulators wish to
convey to unsecured bank creditors that they will
not be bailed out, so that market discipline will be
exercised. However, at the time of a crisis, regulators
will face political pressure to bail out unsecured
creditors. Knowing this, creditors may assume that
their unsecured claims really have a high probability
of being protected by regulators, and this assumption
could undermine market discipline.
For example, Freddie Mac and Fannie Mae each
were limited to a $2.25 billion line of credit from the
Treasury prior to the crisis. However, in September
of 2008, the GSEs were placed under government
“conservatorship,” under which all of their debt was
effectively covered by the taxpayers. This demonstrated
the time inconsistency problem.
The problem of time inconsistency illustrates that it is
difficult to establish and to maintain a clear boundary
between the responsibilities of the private sector and
the responsibilities of government authorities for preventing
and resolving financial crises. If government
tries to let private creditors suffer the consequences
of the risks that they take, the political fallout can be
severe. On the other hand, if government bails out
private creditors, this creates moral hazard, leading
private creditors to take excess risks.
Another major challenge with financial regulation is
that the natural evolution of banks as they seek to
maximize return on equity tends to undermine any
regulatory regime. As we saw earlier, the solutions to
any given crisis have an eerie tendency to come back
as the causes of the next crisis. It would be relatively
easy to devise rules that would prevent an exact repetition
of what occurred in 2008. However, in view
of history one has to wonder whether new regulations
will fail to prevent—or perhaps help to cause—
some future crisis.
7. EROSION OF COM PETITIV E
BOUN DARI ES
Much of the regulatory change that took place
over the past forty years consisted of the informal
erosion and formal elimination of barriers to entry in
financial services. The prohibition against interstate
banking was relaxed and finally ended. The separation
between commercial and investment banking,
established by the Glass-Steagall Act of 1933, was
breached by a number of financial innovations (such
as money market funds) and by regulatory rulings.
The final elimination of Glass-Steagall functional
boundaries through the Gramm-Leach-Bliley Act of
1999 served more to ratify ongoing trends than to create
a dramatic shift in the competitive environment.
The original restrictions on interstate banking and
on the functional boundaries of banks were enacted
under the theory that banks would be too powerful if
they operated nationwide or engaged in a full range of
financial services. By the 1980s though, many economists
viewed the policies to restrict bank operations
28. Susan Woodward and Robert Hall, “Financial Policy: Looking Forward,” Financial Crisis and Recession, May 11, 2009, http://woodwardhall.
wordpress.com/2009/05/11/financial-policy-looking-forward/ writing in support of Squam Lake Working Group on Financial Regulation,
“An Expedited Resolution Mechanism for Distressed Financial Firms: Regulatory Hybrid Securities” (Squam Lake Working Group Paper,
Council on Foreign Relations, April 2009), http://www.cfr.org/publication/19002/expedited_resolution_mechanism_for_distressed_financial_
firms.html/.
Not What They Had in Mind: A History of Policies that Produced the Financial Crisis of 2008
30
1967 1970 1978 1980 1982
Limits imposed on nonbanking
activities of S&L
companies
Bank Holding Company
Act restrictions
expanded to
one-bank holding
companies
Garn-St. Germain
Depository
Institutions Act
National Bank Act
extended to state banks
and savings associations
Depository Institutions
Deregulation and Monetary
Control Act (DIDMCA)
OCC changed national
bank charter standards
Federal mutual charters
granted to state-chartered
savings banks
Figure 5. Compe titive Boundarie s timeline
Mercatus Center at George Mason University
31
1984 1985 1986 1987 1989 1991 1996 1997 1999
Financial Institutions
Reform Recovery
and Enforcement Act
(FIRREA)
Federal Deposit
Insurance C
orporation
Improvement Act
(FDICIA)
Commercial bank
involvement in
investment services
Gramm-
Leach-
Bliley Act
OCC changed
national bank charter
standards, requiring
institutions to provide
statements on formal
lending policies and
fund-management
strategies
Commercial bank
involvement in
securities
Reverse tying
practices repealed
Commercial bank
involvement in underwriting
Investment revenue
limits raised
The Competitive Equality
Banking Act (CEBA)
Subsidiary
involvements
in securities
OCC changed
national bank
charter standards,
requiring CEOs
to be designated
before charter
approval
Not What They Had in Mind: A History of Policies that Produced the Financial Crisis of 2008
32
as anachronistic. Instead, they thought that consumers
would benefit from more vigorous competition in
financial services and that restrictions only protected
inefficient suppliers of those services. There were
three factors that worked to change the competitive
environment in financial services: financial innovation,
regulatory rulings, and legislation. The latter
was probably the least important, in part because of
the long lags involved in enacting banking laws.29
One way to summarize the legislative history of the
Gramm-Leach-Bliley Act, which formally ended
the restrictions on commercial banks engaging in
investment banking and vice-versa, is that it was the
culmination of a long process. For over thirty years,
the competitive structure contemplated in the Glass-
Steagall Act and the Bank Holding Company Act had
been eroded by innovation and regulatory rulings.
The legislative effort to remove barriers to entry
was stalled for many years, because of “turf wars”
involving various interest groups within the financial
services
industry and their regulators. What is important
to recognize is that safety and soundness were
not primary concerns in the debate over the competitive
boundaries within financial services. The chief
legislative challenge was addressing the concerns of
the various interest groups, with each sector trying to
gain entry into other niches while restricting entry to
its native niche. When the legislation finally passed,
it appeared that the banks and their regulators had
won: Banks entered other markets while suffering
relatively little new entry into banking.30
The erosion of competitive boundaries did have consequences
for the structure of the banking system.
Banks became larger and more complex. Non-bank
financial firms became critical to the functioning of
the financial system and closely intertwined with
banks. In retrospect, the complexity and interconnectedness
of the system seemed to play a role in
making the financial system vulnerable to domino
effects and runs. However, given the environment
created by new financial instruments and technologies,
retaining Glass-Steagall and/or the restrictions
on interstate banking would have done little or
nothing to preserve simplicity in financial services.
If anything, retaining the antiquated legislative
framework in the context of ongoing financial innovation
might have resulted in ever more opacity in
the financial system, as institutions continued their
relentless searches for ways to fit the square pegs of
new technology
into the round holes of antiquated
statutory language.
It helps to distinguish two issues: barriers to entry
and safety and soundness. Glass-Steagall and restrictions
on interstate banking were regulatory barriers
to entry. They were attempts to restrict the ways in
which banks could compete with one another and
to restrict entry by certain types of financial institutions
into the markets of other financial institutions.
Economists are predisposed to dislike barriers to
entry. Moreover, innovation and technological
change were constantly undermining the barriers to
entry. Other things being equal, the case for removing
barriers to entry is a sound one.
The safety and soundness issue concerns the fact
that the stability of certain financial institutions has
become a matter of public policy, particularly with
the use of deposit insurance. It is taken as given that
policy makers ought to try to forestall domino effects
and bank runs. To the extent that removing barriers
to entry allows financial institutions to expand
their scope in ways that make them more difficult to
regulate or to stabilize, one can argue that barriers
to entry represent a component of safety and soundness.
When banks are prohibited from undertaking
profitable activities, this does not necessarily preclude
those activities from taking place: Non-bank
In reconstructing the 29. history of the competitive environment in financial services, I have found it highly instructive to review the statements
issued over two decades by the Shadow Financial Regulatory Committee. Some relevant quotes from the committee’s statements are
included in the appendix. The full statements are on the Web site of the American Enterprise Institute at http://www.aei.org/research/shadow/
publications/pageID.888,projectID.15/default.asp.
30. However, this may reflect the fact that Wall Street had already succeeded, with money market funds and mortgage securitization, in penetrating
the most profitable segments within banking services.
Mercatus Center at George Mason University
33
financial firms can expand into those areas. At this
point, regulators face a dilemma. If they allow regulated
banks to expand into previously forbidden
activities, supervisors and examiners may lack the
expertise to assess risk accurately, particularly as balance
sheets become more complex and opaque. On
the other hand, if banks are restricted in their activities,
a “shadow banking system” can grow in these
restricted areas, and that, too, may pose problems
for the safety of the financial system. In retrospect, it
appears that regulators faced both problems—banks
with complex and opaque structures as well as a
large “shadow banking system.”
Over most of the last four decades, the Shadow
Regulatory Committee and many regulatory agency
staff came to view barriers to entry as providing
little or no benefit for promoting safety and soundness.
Today, we can observe that these barriers were
eroded and that safety and soundness was not maintained.
However, it is not necessarily the case that
the barriers could have been retained in the face of
technological change.
If barriers to entry had been retained, this might
have indirectly enhanced safety and soundness by
strengthening the franchise value of financial institutions.
Effective barriers to entry create excess
profits (economists call these “rents”). When a firm
earns rents, it has an incentive to protect those rents
by avoiding risks. In a paper written after the financial
crisis, Gary Gorton makes the point that prior to
the erosion of barriers to entry:
bank charters were valuable because of
subsidies, in the form of limited entry into
banking, local deposit monopolies, interest-
rate ceilings, and underpriced deposit
insurance. In other words, bank regulation
not only involved the “stick” of restrictions
(reserve requirements, capital requirements,
limitations on activities), but also the
“carrot,” that is, the subsidies.31
Any regulation that creates excess profits for financial
firms therefore has the indirect effect of enhancing
safety and soundness. In general, economists have
not advocated using regulations to create excess profits
for this purpose, because barriers to entry create
inefficiency. However, as Gorton suggests,
the inefficiency
might be a price worth paying
if there were no
better way to enhance safety and soundness. Gorton
suggests that this might be worth considering.
8. FINANCIA L INNOVATION
“As to new financial instruments, experience establishes
a firm rule . . . that financial operations do not
lend themselves to innovation. What is recurrently
so described and celebrated is, without exception,
a small variation on an established design, one that
owes its distinctive character to the aforementioned
brevity of the financial memory. The world of finance
hails the invention of the wheel over and over again,
often in a slightly more unstable version. All financial
innovation involves, in one form or another, the creation
of debt secured in greater or lesser adequacy
by real assets. . . . All crises have involved debt that,
in one fashion or another, has become dangerously
out of scale in relation to the underlying means
of payment.”
—John Kenneth Galbraith,
A Short History of Financial Euphoria
Notwithstanding Galbraith’s curmudgeonly
observations, there is much to be said for financial
innovation over the past forty years. There is little
reason to be nostalgic for the financial services industry
of 1960. We would not like to do without automated
teller machines. Not many of us would like to
see minorities shut out of mortgage markets, as they
were to a large extent until recent decades. Few of
us would like to see mainstream financial services
kept out of reach of people with low incomes, forcing
them to rely on pawn shops and the like. There seems
to be little to be said for returning to the high brokerage
commissions on stock trades that prevailed forty
Gary Gorton, “Slapped in 31. the Face by the Invisible Hand: Banking and the Panic of 2007” (NBER working paper, May 9, 2009), http://
papers.ssrn.com/sol3/papers.cfm?abstract_id=1401882/.
Not What They Had in Mind: A History of Policies that Produced the Financial Crisis of 2008
34
years ago. Prior to the advent of money market funds
and mortgage securitization, consumers earned less
on their deposits and paid more for their mortgages.
Without the growth of financial derivatives, it would
not have been possible for institutions to issue fixedrate
mortgages without taking on substantial interest-
rate risk, the perils of which were demonstrated
by the savings and loan industry in the 1970s. Of all
of the financial innovations that emerged in the past
forty years, the overwhelming majority were not
implicated in the crisis. However, a few innovations
clearly were at the center of the turmoil.
Mortgage credit scoring largely replaced human
underwriting in the 1990s. This automated part of the
mortgage application makes processing routine, perhaps
saving consumers one or two hundred dollars
in fees. More importantly, mortgage credit scoring
changed the approach to credit risk in the market.
The rules of thumb in human underwriting served
to segment the market into essentially three categories:
investment quality (meeting the strict credit
standards of Freddie Mac and Fannie Mae), below
investment quality, and unqualified. Credit scoring
allowed for finer gradation of risk, with many risk
buckets. Lenders priced for risk by charging different
interest rates for loans in the various risk buckets.
Many formerly non-investment-quality borrowers
could be charged interest rates closer to that
on an investment-quality loan. Furthermore, many
formerly
unqualified borrowers could be accommodated
at an appropriate interest rate (or so it was
thought). Credit scoring also facilitated securitization
of mortgages, giving purchasers of mortgage
pools objective data with which to measure the credit
risk of the underlying mortgages.
Credit scoring was adopted at a time when there were
no major imbalances in housing markets. In the 1980s
there were regional housing slumps in Texas, New
England, and California. However, from the mid-
1990s through 2005, house prices rose everywhere.
This probably caused many investors to take an overly
optimistic view of the effectiveness of credit scoring.
Some of the apparent success of credit scoring reflected
the favorable trends in house prices, rather than the
reliability of the scoring methodology.
Another important innovation in this period was
private-label mortgage securities. These were securities
not guaranteed by Freddie Mac or Fannie Mae.
Instead, the credit risk was retained by private investors.
Credit scoring helped to give investors guidance
concerning the risk of the underlying mortgages.
Growth of private-label securities was propelled by
another innovation known as structured finance, in
which the credit risk in a given mortgage pool was split
unevenly among various tranches. The most junior
tranche would take the first losses. The next losses
would go to the next tranche. Other tranches, called
senior tranches, were insulated from taking losses
except under the most unlikely catastrophic scenarios.
Senior tranches were able to obtain ratings
of AA and
AAA from the national credit rating agencies.
A further innovation that helped enlarge the mortgage
securities market was the use of credit default
swaps. A credit default swap can be thought of as a
form of insurance against the default of a security.
Default insurance has long been in use to broaden
the market for municipal bonds, allowing cities and
states with imperfect credit ratings to sell bonds
to investors that are required to hold only low-risk
securities. Similarly, with the protection of credit
default swaps, mortgage securities could be sold to
institutions that otherwise might be precluded from
holding or reluctant to hold them.
The thinking behind credit default swaps is that they
are comparable to other financial derivatives, such as
options on foreign currencies or on Treasury securities.
Derivatives create a liquid market for trading
risk, and they can provide a public measure of the
price of risk. Thus, many market participants view
the changes in the prices of credit default swaps as
indicators of changes in the probability of default of
the underlying instruments.
However, credit risk is unlike interest-rate risk
or currency risk in that it is highly asymmetric.
Mercatus Center at George Mason University
35
Currencies and interest rates move up or down with
approximately equal probability. Taking a position
on currencies or interest rates is a bit like betting on a
coin flip. In contrast, mortgages and corporate bonds
default with a very low probability, but the severity of
loss is high. The seller of credit default swaps is positioned
like a property insurance company with a lot
of exposure along the Gulf Coast. Most of the time,
the seller just collects premium income. However,
if a severe hurricane strikes, the losses could be
very large.
Credit default swaps played a major role in one of
the main acts of the crisis: the downfall of AIG insurance.
In the period 2003–2005, AIG was the insurance
seller for billions of dollars of credit default
swaps on what were presumed to be safe securities.
By 2008, when the outlook for the underlying securities
was becoming much more treacherous, AIG’s
counterparties were demanding that AIG post collateral
to ensure that it would not default on the credit
default swaps. These collateral calls taxed AIG’s ability
to raise liquid funds, forcing the company to borrow
heavily from the Federal Reserve and from the
U.S. Treasury.
Credit default swaps also helped produce the interinstitutional
entanglement that made government
officials fear domino effects. Because credit default
swaps were traded over-the-counter, rather than in
an organized exchange, there was a prospect that if
a major seller of credit default swaps went bankrupt,
its counterparties could be in legal limbo until the
bankruptcy was resolved by the courts.
In the late 1990s, the head of the Commodities Futures
Trading Commission (CFTC), which oversees derivatives
trading on organized futures markets, argued
that the CFTC should have regulatory authority over
credit default swaps. Today, many economists believe
that credit default swaps would be safer if they were
standardized and traded on an organized exchange,
rather than traded over-the-counter.32 Another point
to note is that AIG’s subsidiary that sold credit default
swaps operated under the umbrella of a savings and
loan, which was subject to the regulatory jurisdiction
of the Office of Thrift Supervision (OTS). In hindsight,
it does not appear that OTS exercised sufficient
oversight over the risks that AIG accumulated by selling
credit default swaps.
It is not clear what would have been the result had
Congress chosen to encourage or to require that
credit
default swaps be traded on an organized
exchange. The following issues arise:
1. Standardized credit default swaps would not
have served the mortgage securities market.
Holders of mortgage securities are not
looking to buy an insurance policy that pays
32. The Shadow Regulatory Committee did not support the earlier proposals to regulate credit default swaps. In September of 2000, these
economists wrote:
The Committee also recommends that over-the-counter derivative transactions between sophisticated investors be exempt from CFTC
and SEC regulation.
. . . The dominant players in the OTC markets are banks, and federal banking regulators already exercise regulatory control over those
institutions and their derivatives activities that renders unnecessary additional regulatory oversight of the OTC market. The Committee,
therefore, recommends clarifying legislation to exempt OTC derivatives bought and sold by sophisticated investors from regulation by the
SEC and CFTC. The Committee recognizes that the exemption of OTC derivatives from CFTC regulation raises some substantive
competitive issues about the structure and regulation of derivatives products traded on exchanges. The Committee believes that
serious thought should be given to reducing federal regulation on all derivative products that are bought or sold by sophisticated
investors, whether traded over-the-counter or on an exchange.
Shadow Financial Regulatory Committee, “The Regulation of Derivative Instruments,” statement no. 163, May 2000, http://fic.wharton.
upenn.edu/fic/Policy%20page/20051114_ShadowStatement163%5B1%5D.pdf.
Not What They Had in Mind: A History of Policies that Produced the Financial Crisis of 2008
36
off in the event that some generic mortgage
bond defaults. They want to buy protection in
case their specific bonds default. Because the
demand for insurance is specific rather than
generic, it is not clear how anything other than
an over-the-counter market could have served
the purpose.
AIG was an enormous player 2. in the credit
default swap market. It is not clear how an
organized exchange could manage its exposure
relative to a single, dominant participant.
3. Credit default swaps start out as deep, outof-
the-money options. That is, when the
underlying
securities are first issued, the probability
of default is very low. Generally speaking,
options traded on organized exchanges are
much closer to being at-the-money. At-themoney
options behave much better than out-ofthe-
money options. The latter are worth zero
under most scenarios, but under extreme conditions
they can be worth a fortune. This highly
nonlinear behavior makes it very difficult for an
exchange to manage its counterparty risk to sellers
of deep, out-of-the-money options. Hence,
such organized exchanges do not offer such
options ordinarily.
Structured finance and credit default swaps emerged
in order to feed the appetite of institutions for AAArated
assets. This appetite was stimulated by riskbased
capital rules. In fact, the question of whether
generic credit default swaps could substitute for
over-the-counter credit default swaps depends in
part on capital regulations. If a bank could get the
same reduction in risk-based capital required for
holding a mortgage security protected by a generic
credit default swap as it could for holding that security
protected by an over-the-counter credit default
swap, then that would improve the viability of trading
CDS on an organized exchange. However, such a
policy would greatly complicate the administration
of risk-based capital regulations.
The demand for credit default swaps on mortgagebacked
securities was closely related to risk-based
capital regulations at banks. Thus, the growth of
credit default swaps, particularly in AIG’s portfolio,
was not autonomous. It was part of the process
of regulatory capital arbitrage. Rather than blame
financial innovation per se, it may be more appropriate
to fault the regulatory framework that created
incentives for these particular innovations to take
off and to be abused.
As we have seen, risk-based capital regulations, particularly
beginning in January of 2002, put a premium
on AAA-rated assets: banks could hold such
assets with very little capital. Obtaining protection
from AIG insurance, with its AAA rating, enabled
banks to expand their holdings of mortgage securities.
Risk was transferred from the banks to AIG. As a
result, capital left the banks, but it did not go to AIG.
AIG used its AAA rating, not actual capital, to back
its positions. Or, to put this another way, the Office of
Thrift Supervision, which regulated the unit at AIG
that traded CDS, did not require AIG to add capital
in proportion to the amount of capital that AIG’s
counterparties were able to subtract. The result was
a net increase in the ratio of risk to capital in mortgage
finance.
The rigidity of the risk buckets in the Basel Accords
may have played a role in stimulating the growth of
credit default swaps. The risk buckets measure the
risk of each asset individually, rather than treating
assets as a portfolio. Suppose that a diversified portfolio
of B-rated bonds will be as safe as a single bond
that is rated AA. With rigid capital requirements, a
bank would have to hold more capital to hold the
B-rated bonds. However, by buying credit default
swaps from a highly rated insurance company, the
bank could hold the B-rated bonds without having
to hold additional capital.
Of course, if the diversified bond portfolio really is
low risk, then the bank should be allowed to reduce
its capital without having to purchase a credit default
swap. On the other hand, if the diversified bond portfolio
is not really low risk, then when the insurance
company sells the credit default swap, its regulator
should require higher capital. The credit default
Mercatus Center at George Mason University
37
swap does not change the underlying risk of the bond
portfolio. Allowing capital to leave the financial system
because of the credit default swap reflects a flaw
in the design of capital regulations. One can blame
this on innovation, but it goes back to the design and
implementation of capital requirements.
Credit default swaps on corporate bonds might be
a source of 21st-century bank runs if the sellers of
such swaps use what is known as dynamic hedging
to protect their positions. The analogy would be with
portfolio insurance, which was a phenomenon that
emerged two decades ago. Portfolio insurance created
synthetic put options on stock portfolios, just as
credit default swaps create synthetic put options on
interest-bearing securities.
On October 19, 1987, stock prices in the United States
fell by more than 20 percent—the largest one-day
percentage drop in history—without significant
news. Many institutional investors had obtained
“portfolio insurance,” which guaranteed their stock
portfolios against large losses. The sellers of portfolio
insurance planned to execute stock sales in order
to back their insurance promises. Selling stocks as
prices fall in order to create a synthetic put option is
known as dynamic hedging. It works in a liquid market
when it is attempted in low volume. However, not
everyone can execute dynamic hedging at the same
time. Hence the contingency plans of the sellers of
portfolio insurance were not mutually compatible.
In some instances, credit default swaps may have
been sold under the same contingency plans as portfolio
insurance. A credit default swap is like a put
option or insurance. The buyer of a credit default
swap is obtaining insurance against a default on the
security. The seller is providing such insurance.
In theory, the sellers of credit default swaps on
individual firms may have planned to implement
dynamic hedging. If I have sold a credit default swap
on debt from company A, my plan might be that if
company A starts to get into trouble I will short the
stock or other debts of company A in order to create
a synthetic put option to offset my sale of the credit
default swap. However, if many other investors have
the same plan, then we cannot all sell at once without
driving down the prices of the bonds and shares
of company A faster than dynamic hedging can
be executed.
In theory, credit default swaps create inherent instability
by leading sellers of CDS to form contingency
plans for aggressive short-selling that cannot all be
executed when desired. However, I cannot provide
evidence that this problem manifested itself in practice.
Although there was widespread concern over
short-selling in the latter half of 2008, we did not
observe the sort of rapid, overwhelming selling that
took place in the October 1987 stock-market crash.
Like portfolio insurance, credit default swaps represent
put options that start out deep out of the
money. If you sell me a put option on a security with
an exercise price of, say $80, then I have the option
to sell you that security for $80. If the current price
of that security is $100, then the option is deep out
of the money, because the price would have to fall
by at least $20 before I would want to exercise that
option. Sellers of such options expect to earn small
premiums in most scenarios, but they stand to lose
substantial amounts in rare scenarios.
Regulating financial innovation is much easier after
the fact than before. Many innovations, such as the
growth of hedge funds and private equity firms, were
feared to pose risks but were not implicated in the
recent crisis. On the other hand, mortgage credit
scoring seemed to be a relatively benign innovation—
lowering the transaction costs in obtaining a mortgage
and broadening the availability of mortgage
credit—yet it helped to contribute to the excesses in
sub-prime lending and securitization. It is difficult
to have confidence that regulators will be able to distinguish
ex ante the dangerous innovations from the
benign ones.
Not What They Had in Mind: A History of Policies that Produced the Financial Crisis of 2008
38
9. MON ETARY POLICY AN D LOW
INT EREST RAT ES
In retrospect, it can be argued that expansionary
monetary policy in 2001–2003 set the stage for the
housing bubble. Low interest rates were an enabling
factor in the increase in home purchases and the
expansion of mortgage lending.33 Moreover, the
excesses of the bubble from 2004–2006 might have
been curtailed by tightening monetary policy sooner
and more aggressively than was done. Therefore, it
is worth providing a brief outline of how the conventional
wisdom on monetary policy evolved over the
past forty years.
In the late 1960s, the conventional view of macroeconomic
stabilization policy focused on fiscal policy.
The standard view emphasized a trade-off between
inflation and unemployment (the Phillips Curve),
with an additional causal factor known as “costpush”
inflation, reflecting the conflict over income
shares between labor and capital. The problem of
cost-push inflation was thought to require “incomes
policies,” which were government efforts to limit
wage and price increases.
In the 1970s, the Nixon Administration implemented
wage and price controls in an effort to control inflation.
Although these policies met with initial success, by
the late 1970s inflation was approaching 10 percent
per year, with high unemployment. The conventional
wisdom began to shift in favor of the views of Milton
Friedman, who argued that (a) there was no permanent
trade-off between inflation and unemployment
and (b) inflation is always a monetary phenomenon.
In 1979, President Carter appointed Paul Volcker
to be chairman of the Federal Reserve Board, and
Volcker was given a mandate by Carter, as well as
by Carter’s successor, President Reagan, to slow the
rate of money growth in order to curb inflation. For
the next twenty-five years, inflation declined while
unemployment, after rising sharply during a recession
in 1980–1982 caused by Volcker’s contractionary
monetary policy, dropped to low levels.
The period from 1983 through 2007, during which
the U.S. economy experienced low unemployment,
low inflation, and only shallow recessions, was often
described as the Great Moderation. The conventional
wisdom was that monetary policy played a big role in
achieving these outcomes. This reinforced the view
that monetary policy should be the dominant tool for
macroeconomic stabilization. The focus was on maintaining
a low rate of inflation, with the presumption
that fluctuations in employment would be moderate.
During the Great Moderation, a number of financial
crises took place—a stock market crash in August of
1987, a series of sovereign debt crises in the 1980s and
1990s, and the dot com crash in 2000. However, in
each case, any potential impact on economic growth
and employment was apparently mitigated by monetary
expansion. Thus, the conventional wisdom was
that because monetary authorities could mitigate the
effects of financial crashes, there was no need for
monetary policy to focus on identifying or stopping
financial bubbles in order to prevent such crashes.
This conventional wisdom would be less well
accepted today. In contrast with previous financial
crises, the current crisis led to a sharp recession that
could not be mitigated with monetary expansion.
Essentially, the old wisdom would say that expansionary
monetary policy, as the Fed has been pursuing
since the fall of 2008, should be sufficient to
prevent a recession. This is not the case, as shown by
the fact that (a) we are also trying fiscal stimulus and
(b) even so, we are having a severe recession. This
suggests that in hindsight more should have been
done to prevent the housing bubble from expanding
as much as it did. This in turn suggests that the
Another factor that held 33. down interest rates was the large demand for U.S. securities. Federal Reserve officials referred to a “global savings
glut” as a possible explanation for low rates. Ben Bernanke, “The Global Saving Glut and the U.S. Current Account Deficit” (speech,
Virginia Association of Economics, Richmond, VA, March 10, 2005).
Mercatus Center at George Mason University
39
monetary easing that took place from 2001–2003
was excessive.
However, at the time, the sluggish growth in employment
(the 2001–2003 period was commonly referred
to as a “jobless recovery”) was thought to justify the
monetary expansion and low levels of interest rates.
Indeed, in August of 2002, Paul Krugman wrote a
column on the sluggishness of the economy, in which
he passed along a joke that proved to be prophetic.
To fight this recession the Fed needs more
than a snapback; it needs soaring household
spending to offset moribund business
investment. And to do that, as Paul
McCulley of Pimco put it, Alan Greenspan
needs to create a housing bubble to replace
the NASDAQ bubble.34
Krugman and others were frustrated that
Greenspan’s Fed was keeping short-term interest
rates too high. However, at the same time, long-term
interest rates had been falling. In fact, the differing
behavior of long-term and short-term interest rates
should raise questions of just how much control the
Fed really has over the mortgage market.
Before the recent crisis, the conventional wisdom
was that monetary policy should focus on aggregate
economic performance and that it was not wise to
put the entire economy through a recession merely
to stop a housing bubble. That view looks less compelling
today. However, if there are other regulatory
tools available for addressing financial safety and
asset market bubbles, then it would still seem better
to use those tools to stabilize financial markets while
reserving monetary policy for stabilizing the growth
rate in nominal gross domestic product (GDP).
10. DOMINO EFFECT S AN D BAN K
RUN S—REVI SIT ED
A number of economists, including Hyman Minsky
and John Kenneth Galbraith, suggest that instability
is a characteristic of financial markets. In this view,
finance is naturally subject to waves of euphoria and
pessimism. There may be an inherent tendency for
financial institutions to become vulnerable to domino
effects and runs. Fundamentally, the nonfinancial
sector wants to hold short-term, riskless assets
(think of demand deposits) and to issue long-term,
risky liabilities (think of long-term debt to finance
purchasing a home or planting fruit trees). The financial
sector fills a need by having a balance sheet with
the opposite characteristics: risky, long-term assets,
financed by issuing short-term riskless liabilities.
Financial intermediation can work through three
mechanisms: diversification, risk selection and
monitoring, and signaling. The systematic instability
tends to come from signaling.
Diversification can be on the asset side or on the
liability side of the intermediary’s balance sheet. On
the asset side, investing in a great many fields of fruit
trees or home mortgages reduces the risk that any
one adverse event will bankrupt the intermediary.
On the liability side, having many depositors reduces
the risk that the demand for withdrawals at any one
time will be more than the bank can handle. Risk
selection and monitoring allows the bank to specialize
in the collection of information about the risks.
In our simple examples, a bank could study different
fields to know where fruit trees are more likely to
thrive. It could underwrite individual mortgage borrowers
in order to select loans that are most likely to
be repaid.
Finally, there is signaling, which is the most likely
to contribute to systematic instability. Given that a
34. Paul Krugman, “Dubya’s Double Dip? The Conscience of a Liberal,” New York Times, August 2, 2002, http://www.nytimes.
com/2002/08/02/opinion/dubya-s-double-dip.html.
Not What They Had in Mind: A History of Policies that Produced the Financial Crisis of 2008
40
financial intermediary knows more than others about
the nature of the risks on its balance sheet, and given
that its balance sheet consists of long-term risky
assets and short-term, riskless liabilities, the intermediary
depends on the trust of its creditors. This
makes signaling very important. A bank needs to send
signals to depositors that it is sound. Traditional signals
included expensive lobbies and conservatively
dressed employees. Recognized brand names and
long histories of profitability can also be signals that
appeal to consumers.
No matter how many pleas are made for greater transparency,
signaling will always be a part of financial
intermediation. If an intermediary were perfectly
transparent, then the investor would know exactly
what risks it is taking. If the investor knew everything
about the underlying risks, then the investor could
select the risks for herself—she would not need the
intermediary. Invariably, some of the diversification,
risk selection, and risk monitoring is going to be opaque
to the investor. Given that opacity, investors will rely
on signals to decide where to entrust their funds.
Signals of government backing can be extremely
valuable. Banks in the United States put the symbol
of FDIC insurance on their front doors. Even after
they were sold to private shareholders, Freddie Mac
and Fannie Mae kept their original names (Federal
Home Loan Mortgage Corporation and Federal
National Mortgage Association) in order to signal
their government backing.
Until recently, the ratings issued by NRSROs were considered
valuable signals. In part, this was due to the fact
that government regulators, particularly after January
1, 2002, allowed AA- and AAA-rated securities to have
lower risk weights in bank capital requirements.
As economic circumstances improve, signals tend
to have upward momentum. If a signal was trusted
yesterday, it will be trusted slightly more today. And
if it is still trusted today, it will be trusted slightly
more tomorrow.
On the other hand, signals can lose value suddenly.
Highly-rated mortgage securities went from being
trusted to “toxic” in very short order. Few investors
seemed willing or able to sift through these securities
to
determine which ones might be less risky than others.
A major reason that signals lose value so quickly is
that a slight adversity can trigger a downward spiral.
In a classic case of uninsured banks, this is a
bank run. Once bad news circulates about the bank,
it is in the interest of every depositor to withdraw
funds. This weakens the bank further, leading to
more withdrawal until the bank is either bailed out
or has to be closed. Although consumer bank runs
were mostly avoided during the most recent crisis,
there were institutional equivalents. For example,
as Freddie Mac and Fannie Mae in 2008 announced
large credit losses for preceding quarters, their debt
began to include a large risk premium charged by
investors. This in turn made the two firms less viable,
and they were taken into conservatorship by the
Treasury. Another example was AIG, whose counterparties
began to be concerned about its ability to
back its portfolio of credit default swaps. The large
counterparties, including major investment banks,
demanded that AIG post collateral. This forced AIG
to sell assets in order to obtain low-risk securities.
An increased demand for collateral also took place
in the market for repurchase agreements. In the
“repo” market, as risk premiums increased, investment
banks and the trading accounts of commercial
banks were compelled to post more collateral or to
sell assets.
One of the problems with the idea of using subordinated
debt as a market-based tool for regulating
financial institutions is that investors lose confidence
quickly rather than gradually. One month,
Freddie Mac and Fannie Mae were borrowing at
interest rates less than one-quarter of one percent
above comparable Treasuries. A few months later,
they had to pay over one percentage point above
Treasuries. The GSEs no longer signified safety
and soundness to investors, so that in order to keep
Mercatus Center at George Mason University
41
them operating the Treasury had to take the firms
under conservatorship.
The unstable behavior of financial signals poses
another problem for regulators. Just like private
investors, regulators have imperfect knowledge of
the exact risks embedded in the balance sheet positions
of regulated institutions. The regulators, too,
must rely on signals, and they mistakenly relied on
NRSRO ratings of securities as signals.
In principle, what regulators want is for the signals
issued by financial intermediaries to be successful
at convincing investors of soundness—but not too
successful. If signals are too successful, then intermediaries
will expand too much, as they did during
the mortgage securitization boom. If signals are too
distrusted, then intermediation will be overly curtailed,
reducing economic activity. Given the natural
instability of trust in signals, it would seem that the
regulators’ goal of maintaining risk at a level that is
“just right” is not easy to achieve. Instead, it seems
more likely that signals will gradually become more
and more trusted, until the trust is excessive and an
event triggers a crash.
This theory of financial instability has two implications
for regulators. One implication is that regulators
have to figure out how to take away the punch
bowl when the party is getting good. This means recognizing
the point where financial complacency and
euphoria are too high. It means devising policies to
try to curb excess without causing a severe economic
slump. Finally, it means overcoming bureaucratic
and political obstacles in order to execute policy.
As with many aspects of financial regulation, the goal
of taking away the punch bowl at the right time can
lead to two types of errors. One type of error, which
we might call Type I, is taking away the punch bowl
too late. The other type of error, which we might call
Type II, is taking away the punch bowl before it is
necessary or perhaps when it is not necessary at all.
A Type I error results in financial intermediation
expanding too much, leading to excessive risk-taking.
When the risks start to become apparent to market
participants, a vicious downward spiral takes place.
Bad investments have to be written off. Moreover,
trust in the existing financial intermediation practices
and signals is broken, which further exacerbates the
economic costs of the financial collapse.
The economic cost of a Type II error is more difficult
to assess. Once regulators crack down it is not
possible to observe what might have happened had
they allowed financial intermediaries to expand
further. We can never know if the crackdown was
premature or unwarranted. However, the political
cost of a Type II error can be high, because it puts
the regulator in a position of restricting a practice
that appears to be generating profits for firms and
benefits for consumers.
The second implication of this theory of financial
instability is that regulations designed with the
knowledge of previous financial euphorias will not
necessarily be able to stop the next euphoria. In
fact, as this paper has detailed, each era of regulation
seems to contribute to the next era of euphoria.
Thus, after the Great Depression, when uninsured
banks and short-term “balloon” mortgages were the
problem, policy makers produced a mortgage finance
system dominated by thirty-year, fixed-rate mortgages
held by savings and loans. These S&Ls were
precisely the institutions that blew up in the next crisis,
as the high inflation and interest rates of the late
1970s and early 1980s made them insolvent.
Next, given the role that book-value accounting, lack
of formal capital requirements, and interest-rate risk
played in the S&L crisis, policy makers promoted
market-value accounting, risk-based capital, and
securitization. These were precisely the features
that blew up in the most recent crisis.
11. EASY TO FIX VS. HAR D TO BREAK
If economic stability inevitably gives way to financial
euphoria, then it may not be possible to devise
a fool-proof regulatory regime. Instead, it may be
Not What They Had in Mind: A History of Policies that Produced the Financial Crisis of 2008
42
more effective to aim for a system that is easy to fix
than a system that is hard to break. This means trying
to encourage financial structures that involve less
debt, so that resolution of failures is less complicated.
It also means trying to foster a set of small, diverse
financial institutions.
In the United States, tax policies tend to encourage
debt financing. Higher leverage in financial structures
makes a system difficult to repair when investments
founder. If tax policy encouraged equity financing
instead, investment failures would not cause so much
difficulty. For example, the crash of the dot com bubble
in 2000 caused much less economic dislocation
than the more recent housing crisis.
Another way to make a financial system easy to fix
would be to have small institutions with only weakly
correlated risks. If that were the case, then the closure
of one institution would not be a major event for
the economy. Of course, arranging for risks to be only
weakly correlated is easier said than done.
From the standpoint of making the regulatory system
harder to break, it may make sense to have a
neat regulatory organization chart, without gaps or
overlaps. However, such a well-ordered regulatory
system might result in a situation where all of the
institutions performing a particular function, such
as mortgage lending, fail together. With a messier
structure, the failure of some firms might be overcome
by other, overlapping firms taking a larger role.
Thus, instead of aiming to bring all mortgage lending
under a single regulatory regime, it might be easier to
fix a system if there were a variety of mortgage lenders,
regulated differently.
Of course, one problem with multiple regulators
is that there can be a competitive “race to the bottom,”
as each type of institution asks its regulator for
relief from its perceived regulatory disadvantages. It
appears that bank regulators felt sympathy toward
banks because of the low capital requirements for
taking mortgage credit risk enjoyed by Freddie Mac
and Fannie Mae. The regulators leveled the playing
field not by raising the capital requirements for GSEs
but by lowering the capital requirements for banks. If
a system of multiple regulators is to be retained, then
they need to respond to complaints about tilted playing
fields by tightening up on the favored institutions
at least as readily as they loosen regulations for the
disadvantaged institutions.
12. CONC LUSION
The core of the financial crisis of 2008 consisted
of unsound practices in mortgage underwriting and
mortgage finance. A number of regulatory developments
helped to stimulate the boom in mortgage
lending and securitization.
The Basel Accord on risk-• based capital set
up crude risk buckets that initially favored
Freddie Mac and Fannie Mae, because capital
requirements
were lower for mortgages securitized
by the GSEs than for loans originated and
held by banks.
• The January 2002 modification to the risk
weights allowed NRSRO ratings to substitute
for GSE guarantees. This reduced the relative
advantage of the GSEs, but it increased the
relative advantage of mortgage securitization.
Private-label securities, consisting of pools of
low-quality mortgages, expanded dramatically
from 2002 through 2005.
• From the mid-1990s onward, the government
pressured mortgage lenders to increase lending
to low-income borrowers. Freddie Mac and
Fannie Mae lowered credit underwriting standards
considerably in response to this pressure,
taking on significant sub-prime mortgage
exposure in 2006 and 2007, just as house prices
were poised to fall.
• The incentives to hold AAA- and AA-rated assets
stimulated various financial innovations that
had unfortunate consequences. Among many
examples,
AIG insurance used credit default
swaps on mortgage securities to “rent” its AAA
rating to banks.
Mercatus Center at George Mason University
43
• Monetary policy that was intended to stabilize
inflation and employment kept interest rates
low from 2002 through 2004, which contributed
to the housing boom.
• Regulators lacked the will and the ability to
enforce competitive boundaries in the financial
sector. These boundaries eroded over a fortyyear
period, primarily as a result of innovation
but also as a result of regulatory decisions and
legislation. Consequently, institutions became
large and complex. These “too big to fail” firms
posed major challenges to policy makers during
the crisis, because they were subject to domino
effects and 21st-century bank runs.
In this paper, I have stressed the differences between
the way that policies were viewed at adoption and the
way that they are viewed in retrospect. For example,
basing capital requirements on risk and on the market
value of assets made sense in light of the S&L crisis,
but such policies are now recognized to be procyclical.
They should not be abandoned altogether, but they
need to be modified. Other policies that are now recognized
as harmful, such as the reliance on credit rating
agencies and approval of dispersing risk into the
“shadow banking system,” were at the time viewed as
beneficial. The phenomenon of mortgage securitization
is still viewed as beneficial, with a need to curb its
excesses. However, I would question the rationale for
securitization. Given that the government created and
supported mortgage securitization, without government
support or the distortion of capital regulations
perhaps the market would choose a different,
safer
method of mortgage finance. Perhaps old-fashioned
“originate-to-hold” mortgages would make a comeback
if the regulatory playing field were level.
Given this contrast between hindsight and the realtime
perspective, the government needs to display
some humility in promising to prevent future financial
crises. The history of past regulatory mistakes
suggests that we will not come up with a fool-proof
system going forward. In fact, there is a risk of creating
a financial system even more dependent on centralized
regulation, which could leave it at least as
vulnerable to catastrophic failure.
The prospects for regulatory policy are even more
fraught given the extremely skewed conventional
narrative of the financial crisis. Rather than examine
all of the factors looked at in this paper (which in
itself may not be exhaustive) the conventional narrative
looks only at private-sector excesses and an
alleged absence of regulatory oversight. It is unlikely
that our financial system will benefit from a rush to
create new rules and institutions that is based on a
distorted perspective on how the crisis emerged in
the first place.
Based on my research and the findings of this paper,
perhaps the most useful steps that policy makers
could take to prevent a recurrence of the financial
markets crisis would be to tilt policies away from debt
finance. One way to encourage a more stable housing
market would be to provide less encouragement to
mortgage indebtedness. With larger down payments
and smaller mortgages, there would be less of a selfreinforcing
effect of house price appreciation, speculative
demand, and mortgage credit availability.
Policy makers should also rethink the mortgage
interest deduction and reconsider the role played by
Freddie Mac and Fannie Mae. If, without the GSEs,
mortgage financing reverted to a traditional lending
undertaking by banks, that might prove to be more
sound, particularly if monetary policy keeps inflation
under control. If mortgage interest rates are a
bit higher with traditional lending than they could be
with more securitization, that need not be regarded
as a tragedy.
For financial intermediaries in general, a smaller disparity
in the tax treatment of debt and equity might
reduce the incentives for excess leverage. That in
turn might help to moderate excesses. It would also
discourage the sort of debt-laden financial structures
that are conducive to domino effects and bank runs.
The main point of this paper is that in order to get
policy right going forward, the historical narrative
must be accurate. It will not help to airbrush out
of history the role that regulatory policy played in
setting up the crisis. It would be a mistake to creNot
What They Had in Mind: A History of Policies that Produced the Financial Crisis of 2008
44
ate institutions with the presumption that regulators
will correctly diagnose systemic problems, when
the record shows that regulators were subject to the
same cognitive shortcomings as private sector participants.
Unless the United States comes to terms
with the fact that the actions of policy makers and
regulators contribute to financial fragility, it has little
hope of moving in the direction of a less fragile system
for the future.
Mercatus Center at George Mason University
45
Appendix: The Shadow Regulatory Committee
on Barriers to Entry
In 1986, one of the committee’s first statements said:
The Committee recognizes that the legislative
barrier between banking and securities
activities erected by the Glass-Steagall Act is
being eroded in a piecemeal and haphazard
fashion . . . new legislation should be enacted
that is more consistent with both current
market forces and present economic
theory
and fact.35
The committee took the view that there was no natural
boundary between investment banking and commercial
banking. Investment banks were providing
money market funds with checking privileges. They
were underwriting commercial paper, which substitutes
for bank loans. For their part, banks could buy
and sell mortgage securities or municipal bonds. In
the view of the committee, the attempts to maintain
legislative barriers to entry in financial services were
crude and counterproductive.
In a statement in 1994, the committee wrote:
In recent decades, bank holding companies
have been induced to try to expand into
an increasingly wide array of previously
precluded activities, including issuance of
securities and insurance products. At the
same time, nonfinancial and nonbank financial
firms have developed subsidiaries and
affiliates whose products closely substitute
for bank loans and deposits.
. . . As long as supervisors strive to force recapitalization
before net worth can go to zero,
the risks to taxpayers from banks affiliating
with firms engaged in nontraditional banking
or commercial activities are not qualitatively
different from traditional activities provided
they can be adequately monitored.
. . . Banking organizations now operate
nationwide and have diverse product lines.
Market power associated with this expansion
is constrained by nonbank competitors.36
The barriers to entry in financial services had initially
been enacted out of fear of concentrated power in
financial markets. By the 1990s, it was difficult to
see concentration of power as a significant threat.
Instead, what economists saw was an environment
with many firms offering financial services. If anything,
barriers to entry were restricting competition,
not protecting it. Moreover, the formal restrictions
seemed increasingly arbitrary in view of all of the
innovative and competitive activity that was allowing
firms to get around the restrictions.
The committee was frustrated at the absence of legislative
action on this issue.
Again this year, despite considerable
efforts almost to the last day of the session,
Congress failed to pass financial reform
legislation. This has happened so often in
recent years that it calls into question the
ability of Congress to change national policy
in this area, and leads many observers to
believe that it is better to rely on actions by
regulators than to bother with legislation.
. . . In the Committee’s view, a primary cause
of the failure this year as in years past—was
the fallacious notion that banks must be
separated from the rest of the commercial
world. Misplaced allegiance to the so-called
separation of banking and commerce has
made it impossible for Congress to create
35. Statement number 13, November 17, 1986.
36. Statement number 115, December 12, 1994.
Not What They Had in Mind: A History of Policies that Produced the Financial Crisis of 2008
46
the two-way street that would meet the
needs of all the players and best serve the
interests of consumers.37
The parties most interested in this issue were the
institutions themselves, with each sector lobbying to
maneuver for advantage. Insurance companies wanted
to keep out competition from banks, while banks
wanted
to be able to offer insurance through subsidiaries.
Investment banks wanted to compete with
banks for consumers without suffering inroads from
commercial banks in security underwriting and other
traditional investment banking functions. The result of
the interest-group bickering was legislative gridlock.
Sixteen months later, still frustrated, the committee
wrote, “Real banking modernization would require
no more than a single sentence: ‘The Bank Holding
Company Act of 1956 and the Glass-Steagall Act of
1933 are hereby repealed.’”38
To understand the economists’ frustration, keep in
mind all of the innovation that had taken place in
banking and finance between 1960 and 1999. Credit
cards had become widespread. There were interestbearing
checking accounts. There were automated
teller machines. Money market funds were well established.
There was now a national secondary market
in mortgages. Many households had home equity
lines of credit. There were exchange-traded futures
and options in financial instruments. There was electronic
trading of shares of common stock. Consumers
were using the Internet for research and selection of
financial services. It seemed that everything about the
financial services environment had changed since the
1950s—with the exception of legislation.
In 1999, Congress passed the Gramm-Leach-Bliley Act,
which officially ended the Glass-Steagall restrictions.
Although the economists were not happy with the
complexity of the final product, they expressed relief,
“The Congress enacted the Gramm-Leach-Bliley Act
of 1999 (GLBA), which, after almost two decades of
debate, helped bring our financial laws closer to the
realities of the modern financial marketplace.”39
One year later, the committee wrote:
In November, the Gramm-Leach-Bliley Act
reached its second anniversary—enough
time, the Committee believes, to make some
judgments on whether it has resulted
in any
significant improvement in the structure of
the financial services market. Measured
against the balkanized financial services
industry that existed in 1999—with bank
holding companies unable to affiliate with
insurance underwriters, or with securities
firms that were principally engaged in
underwriting and dealing in securities—
there has been some improvement in market
structure. Many bank holding companies
have been able to acquire or establish securities
and insurance activities, and this has
improved competition and enhanced consumer
choice.
However, measured against what the
Committee believes the Gramm-Leach-
Bliley Act should have achieved—the creation
of a two-way street in which insurance companies
and securities firms could acquire or
establish banks, and vice versa—the Act has
been a failure . . . the Act has in fact created a
strong bias in favor of product expansion by
banking organizations and a corresponding
bias against similar expansion by the other
financial services providers.40
37. Statement number 142, December 7, 1997.
38. Statement number 155, April 26, 1999.
39. Statement number 166, December 4, 2000.
40. Statement number 174, December 3, 2001.
About the Mercatus Center at George Mason University
The Mercatus Center at George Mason University is a nonprofit university-based research, education,
and outreach organization that works with scholars, policy experts, and government officials
to bridge academic learning and real-world practice.
Our mission is to generate knowledge and understanding of how institutions affect the freedom to
prosper and hold organizations accountable for their impact on that freedom. The aim of our work
is to enable individuals to live free, prosperous, and peaceful lives.
The ideas presented in this study do not represent an official position of George Mason University.
About the Mercatus Financial Markets Working Group
The Mercatus Financial Markets Working Group combines / academic research on finance and
banking with a deep understanding of the regulatory policy process to offer productive ideas to
address the serious problems in financial markets.
By focusing on identifying the appropriate role for regulation and markets, members of the
Mercatus Financial Markets Working Group provide policy options that encourage:
• innovation within a vibrant financial services industry;
• greater government and private sector transparency;
• well-functioning and dynamic capital markets; and
• reduction of unnecessary systemic risk.
The Mercatus Financial Markets Working Group is made up of academic experts and policy practitioners
working collaboratively to provide a university-based resource to policy makers working
to improve the policy response to the economic crisis.
Policy Regarding Independence of Research
The Mercatus Center is committed to the highest standards of academic quality and to ensuring
that our work stands up to rigorous peer review. Mercatus scholars independently pursue a
research agenda and educational activities that advance our mission. Mercatus does not engage
in research or educational activities directed or influenced in any way by financial supporters. To
view our full policy regarding independent research, visit our Web site at www.mercatus.org.
3301 North Fairfax Drive, Suite 450
Arlington, Virginia 22201