The Treasuiys bailout plan was lui attempt to impnwe bank baliince
sheets and tliereby spur bank lending. The justification offered was that,
as of early September 2008, major huiks were facing imminent tidlure
iK-cause ¿leir mortgage-backed assets had declintxl rapidly in value.
No one disputes that several banks were in danger of failing, but
this doe.s not justify a bailout. Failure is an essential aspect of capitalism.
It provides information about good and bad investments, and it
releases resoijrces from bad projects to more productive ones. As
noted earlier, housing prices ¿ma housing constructton were too high
at tlie end of 2005. This condition implied a deterioration in bank
balance sheets iind a retrenchment in the banking sector, so some
amount of failure was both inevitable and appropriate.
Thus, an economic case for the bailout needed to show that failure
by some banks would hann the economy beyond what was
unavoidable due to the fiill in housing prices. The usual argument is
that failure by one bank forces other banks to fail, generating a credit
freeze. That outcome is possible, but it does not mean the
Treasury’s bailout plan wiis tlie right policy.
To see why, note first that allowing banks to fail does not mean the
government plays no role. Federal deposit insurance would prevent
losses by insured depositors, thus limiting the incentive for bank
runs. Federal courts and regulatory agencies (such as tlie FDIC)
would supervise bankmptc)’ proceedings for failed institutions.
Under bankruptcy, moreover, the activities of failing banks do not
necessarily disappear. Some continue during bankruptcy, and some
resume after sale of a failed institution or its assets to a hetildüer
bank. In other cases, merger in advance of failure avoids bankruptcy
entirely Private shareholders and bondholders take the losses
required to make these mergers and sales attractive to the acquiring
parties. Taxpayer funds go only to insured depositijrs (see Fama
2009, Zingales 2008).
C^onsider, therefore, how bailout compares to bankruptcy from
three perspectives: the impact on the distribution of vvealdi, the
impact on economic efficiency, and the impact t)n the lengtli and
depth ofthe fiiiiuicial crisis.
From a distributional perspective, bailout is unambignonsly perverse;
it transfers resources from tlie genei al taxpayer to well-off economic
actors who profited from risky investments. This is not a
criticism of risk-taking; that is appropriate so long as those benefiting
in good times bear tlie costs in bad times. This is exactly what occurs
under the bankruptcy approach.
From an economic efficiency perspective, bailout is again problematic.
Mere consideration of a bidlout distracts attention from the
fact that government was tlie single most important cause of tlie crisis.
Relatedly, bailout creates a moral hazard, thereby generating excessive
risk-taking in the fviture. Bailout.s often adopt goals that are not
economically sensible, such as propping up lK)using pnces, limiting
mortgage defaults, or preventing the failure of insolvent institutions.
More broadly, a bailout encourages perverse actions by institutions
that lire eligible for the money, such as acíjuiring toxic a-ssets that the
Treasury might buy or taking huge risks with Treasury capital injections.
The Treasurv’ bailout of 2008 also initiated a government ownersi
lip stake in the financial sector. This means that, going forward,
political forces are likely to influence decision m akin g in the extension
of credit and the iilloeation of capital. Government, for example,
might push biuiks to iiid borrowers with poor credit histories, to subsidize
politically connected industries, or to lend in the districts of
powerful legislators. Government pressure is difficult for banks to
resist, since government can threaten to withdraw its ownership
stake or promise further injections whenever it wants to modify Í>ank
behavior. Further, bailing out banks sets a precedent for bailing out
other industries. Thus, the long-run implications of bailout are
unambiguously bad.
Bailout is superior to bankruptcy, therefore, only if allowing
bank failures would cause or exacerbate a credit crunch. Neither
theory nor evidence, however, makes a compelling case for sueh
an effect. As a theoretical matter, failure by a bank means that it
cannot extend credit, but this means a profit opportunity exists for
someone else. As an empirical matter, it is difficult to estabhsh
whether panics cause credit freezes or underKinjî adverse shocks
to the economy cause both reduced lending and panics. Ben
Bemanke’s famous paper on the Great Depression (Bemanke
1983) suffers exactly this problem; it shows that bank failures and
output losses are correlated, but it does not pin down the direction
of causation.
This is not to deny that credit freezes occur and cause harm, nor
to assert that credit markets would have been healthy under the
bankniptcy approach. Rather, the claim is that overinvestment in
housing and the excessive level of housing prices that existed in
the United States meant that an unwinding was necessary to make
the economy healthy This restnicturing implied reduced residential
investment, declines in housing prices, plus shrinkage and consolidation
of the banking sector. All of this would plausibly have
2010
10
Jan
