The AIG Saga: A Brief Primer

The
awarding of $165 million in bonuses to AIG executives has dominated the
news in the last week. There has been widespread outrage over the idea
that taxpayers' dollars are being used to reward the people who
effectively bankrupted AIG and cost the government more than $160
billion in bailout funds to meet the company's obligations. This primer
addresses some of the issues raised by both the bonuses and the much
larger sum going toward the AIG bailout.

The Bonuses: What Did They Know and When Did They Know It?

One
of the silliest distractions in the AIG saga has been the various
accounts of when AIG told Treasury Secretary Geithner of the bonuses
and when Geithner passed the information along to President Obama. This
discussion is silly because Geithner almost certainly knew of the
bonuses ever since the initial takeover on September 15th. He just
didn't think they were important.

Geithner was the chair of the New York Fed at the time of the original
takeover. In that capacity, he was the person directly overseeing the
takeover. As the chairman of the New York Fed, Mr. Geithner was
undoubtedly familiar with the Wall Street culture and knew that
financial firms paid out large bonuses each year to their most-valued
employees. Since he did not issue any directives to AIG telling them
not to pay bonuses, it was reasonable to expect that AIG would do so,
just like it always did.

In other words, Geithner had every reason to believe that AIG would
continue to pay out bonuses even after it was bailed out by the
government, because he did not tell it stop paying bonuses. He may not
have considered this issue important until the last week. And, he may
not have known the exact size and the structure of the bonuses, but for
all practical purposes he has known for six months that AIG would be
issuing million dollar bonuses to certain employees, in spite of the
fact that it was dependent on massive infusions of government money to
stay alive.

Does the Government Have to Pay the Money?

It
is not easy to find legal ways to avoid paying for work that was
already done. It is possible that the government could make it
difficult for the bonuses to be collected by breaking off AIG's
Financial Products division (the one responsible for bankrupting the
company) and then letting this company go bankrupt.1

However, this route has two major problems. First, a main purpose of
the bailout was precisely that the government wanted to honor the
obligations of the Financial Products division, ostensibly to maintain
the stability of the financial system. If this division went bankrupt,
then it could pose risks to the stability of the financial system. The
second problem is that the bonuses have already been paid. Any action
would now require taking back money that was already paid out. This is
considerably more difficult than preventing money from being paid in
the first place.

A second path that is currently being pursued by Congress is to tax
back the bonuses with a tax that is designed explicitly to apply to
bonuses given to workers for companies that are being bailed out by the
government. This sort of measure is a rather blunt instrument to
address the problem. The resulting compensation system is certainly
less than perfect (the bill passed by the House would tax back 90
percent of the bonuses received by highly paid executives), but it
could hardly be worse than the compensation structure currently in
place.

A third possibility is to insist that the private shareholders pay for
the bonuses. Private shareholders still own 20 percent of the company.
The market capitalization is approximately $2.6 billion. This means
that the 20 percent stake ($520 million) owned by private shareholders
can easily cover the $165 million in bonuses.

Under this arrangement, the government would tell AIG to sell enough
new shares to cover the $165 million cost of the bonus. Since the money
is supposed to come from the private shareholders 20 percent stake, for
every share that AIG sells to the public, 4 shares will be awarded to
the government. This keeps the government stake at 80 percent.

The current share price is about 95 cents. If it fell to 60 cents as a
result of the newly issued shares, the company would have to sell 275
million shares to the public and issue another 1.1 billion shares to
the government. This would leave the government's stake unaffected,
while cutting the value of the current private shareholders' stake by
roughly one-third. This route would leave the executives with their
bonuses, but they would come at the expense of the private
shareholders, not the taxpayers.

AIG Bailout Issues

Thus
far $170 billion has been spent on the AIG bailout, more than 1000
times as much as is at stake with the bonuses. For the first time last
weekend, the Treasury Department released information about how this
bailout money was used. It reported that much of the AIG money went to
large U.S. banks, most notably Goldman Sachs, Bank of America, and
Merrill Lynch, in addition to several large foreign banks, including
the French bank Societe Generale and Deutsche Bank. Most of these
payments were in connection with their holdings of credit default swaps
(CDS) issued by AIG.

There are at least three obvious issues that arise with these payouts:

1. Did the banks hold the underlying assets, or just the CDS?
2. Did the government have to buy back the underlying assets from the banks, or could it have waited to see what happened?
3. Could the support for the banks have been done directly, including some quid pro quo, without having AIG as an intermediary?

These three issues are outlined below.

CDS: Insurance or Gambling?

In
2007 the outstanding nominal value of all credit default swaps was
close to $75 trillion. This was approximately five times as large as
the outstanding value of insurable bonds. This meant that there was an
average of five CDSs issued for every insurable bond, which implies
that at least 80 percent of CDSs were not owned by institutions that
actually owned the bond being insured.

The Treasury and Fed have not released the rules they applied in
dealing with AIG's CDS. They may have only honored CDS where the
institution held the bond being insured or they may have honored all of
AIG's CDSs, regardless of whether or not the bank held the bond being
insured.

This makes a big difference in terms of the purpose of the bailout. If
a bank had bought a CDS to protect itself against losses on a mortgage
backed security, and the CDS was not honored, then it would be an
unexpected blow to its balance sheet. On the other hand, if the bank
was just gambling that a bond that it did not hold would go bad by
buying a CDS issued against it, it is difficult to see how a failure to
honor the CDS would impose a serious hardship.

There may be legal issues that would prevent a non-bankrupt AIG from
choosing which CDSs it chooses to honor, but that fact may have
implications for the wisdom of rescuing AIG, as opposed to just
directly supporting the counterparties, where it is considered
appropriate.

Did the Government Pay Off the Bets Before the Race Was Over?

The
government, through AIG, paid an additional $30 billion to
counterparties because it paid off CDSs at their notional value rather
than their market value. In principle, AIG would have owed the notional
value of the CDSs if the underlying bond had defaulted. In these cases,
the bond had not defaulted. In effect, the government acted as though
AIG had already lost its bet, at a time when it was still possible that
the underlying bonds would not go bad.

It is important to keep in mind that CDSs are typically relatively
short-lived assets. Many provide insurance for only three years and
most insure bonds for five years or less. Most of AIG's CDSs were
issued before 2007. This means that by late 2008, they would have
already been two years old or older. In this context, it might have
been reasonable to take a chance to see whether the CDS would actually
have to be paid. In any case, there was no obvious reason to pay above
the market value for the CDS. This seems like a straight gift to the
banks.

Should the Government Have Gotten Something in Return for Giving Tens of Billions to the Banks?

When
the government lent hundreds of billions of dollars to the banks
through TARP, it got preferred shares of stock in return, in addition
to placing conditions on the banks' conduct. By contrast, the
government received absolutely nothing for the tens of billions of
dollars that it passed on to the banks through AIG. It may have been
desirable to ensure that AIG's defaults did not lead to the collapse of
the major banks that were its counterparties, but this could have been
accomplished by directly giving these banks capital through TARP or
some equivalent mechanism. There is no obvious reason why it was
necessary to give the money through AIG without getting anything in
return.

It is worth noting that if the government had instead lent the AIG
money to the banks through TARP, and under similar conditions, it would
own an even larger share of these banks. Obviously the banks prefer
that the money instead pass through AIG without conditions, but there
is no reason that the taxpayers should prefer this route.

It is also worth noting that several of the recipients of AIG money
were foreign banks. While the public has an interest in the stability
of the world economy, which means preventing major foreign banks from
going bankrupt, there is no obvious reason that American taxpayers
should be forced to bail out foreign banks of wealthy countries. It is
possible that there is some quid pro quo under which foreign
governments are bailing out U.S. banks on losses suffered in their
countries, but there has been no public acknowledgement of such an
arrangement.

There is a possible alternative explanation. The government may have
made these payments in order to preserve the international reputation
of the U.S. financial industry. If that is the case, then this is a
rather expensive subsidy to the financial industry. To date there has
been no explanation as to the reason for making these payments.