Toxic Assets: Bad Bank or Guarantees? Some Intertemporal Trade-Offs

At this point, it seems that the government response to the financial crisis (aside from monetary policy) consists of two main elements.  On the one side, a fiscal stimulus package; on the other hand, a financial rescue program.

Regarding the fiscal side, given the lack of financing to the private sector jointly with a contracting economy, an expansionary fiscal policy seems to be the right thing to do.  Although we certainly believe the private sector—not the government—needs to drive the economy, the current situation is clearly not “business as usual.”  This is not a garden variety recession; it is a growing “effective demand failure.”  These rare circumstances argue for a clear role for government intervention.  After all, the only entity that can tax people (in the future, once we return to more normal times) is the government.  Moreover, it is part of the government’s implicit mandate to optimally redistribute resources across sectors and generations; the more so under the current circumstances.  The U.S. has built its reputation by paying back its obligations during good times. Thus, this is the right time for a fiscal stimulus—and a big one. However, our main focus here is on the financial leg of the program.  The current discussion seems to have converged to two main strategies to improve banks’ balance sheets.  One strategy suggests the government guaranteeing the so-called toxic assets as a mechanism to improve lending.  The alternative would be for the government to buy these toxic assets altogether (probably at prices above the current ones) and hold to them over a long time period.

A bad bank scheme should include protecting taxpayers against huge losses due to buying toxic assets—this may include actions such as diluting existing shareholders, temporary nationalization, and antitrust restrictions that break up the recapitalized institutions, actions that will almost certainly upend the status quo.  But the end result is that the government purchases toxic assets from financial institutions, cleansing banks’ balance sheet, and then allocate those assets to the “bad bank.”  Further recapitalization may still be necessary (hence required protection for taxpayers), but the assets are now purged and the healing process can accelerate.

The government will eventually sell the toxic assets into the market in the future.  The difference between the cost of the recapitalization and the selling of removed toxic assets is the ultimate cost to the taxpayer.  It is likely that the overall social cost is significant, with some estimates on the order of a $1 trillion or higher.  Over time, the latter is worth the money for the taxpayer.  Otherwise, the macroeconomic—and consequently social—costs of not doing it will be much bigger: the economy will move into depression mode, many more financial institutions will go bankrupt, investment, consumption, and employment will collapse, and of course, tax revenues will decrease.  In such a situation even the U.S. will find it difficult to borrow so as to smooth the effects of the deep and protracted slump in which the economy will fall—i.e. the government will end up being unable of helping its people.

We favor the bad bank.  From an intertemporal perspective, guarantees are relatively cheap upfront, whereas the bad bank will be more expensive.  No question about that.  If the private sector participates in buying toxic assets[1] it will certainly reduce any risk premium on U.S. debt, thus reducing taxpayers’ direct exposure. However, regardless of whether the private sector helps shouldering the problem or not, one key problem these days refers to wealth misperceptions and a BIG coordination failure.  In this sense, guarantees (or a combination of guarantees with a “small” bad bank component) might be read by the market as a lukewarm response.  If so, it will imply throwing money for bad—and the situation could actually worsen as expectations might coordinate in the wrong equilibrium.  The bad bank, on the contrary, will work as a strong signaling device, coordinating expectations.  Indeed, creating conditions for sustained coordination of expectations is a key role for the government, regardless of the existence of crisis.  Arguably, the lack of a coordinating influence allowed financial firms to pursue excessively risky growth strategies.

Additionally, the bad bank, because of it expectations coordination device will likely imply a high short run cost with a low long run cost—once the economy is back on trend, a big share of the toxic assets will be repaid.  (Sweden is a great example: close to 90% of the value was recovered; the U.S will probably achieve a lower rate of success, but, again, it is still worth it to prevent the costs associated with persistent output gaps.)  For if not, the low short run costs of the guarantee scheme might face the high long run costs of not being strong enough to truly jumpstart the economy.  In the latter case, the long-run effect on the U.S. (and world economy) will be much gloomier…

One additional comment is the following.  Is this a liquidity or a sustainability problem?  We really don’t know for sure.  If it is just liquidity the above two-leg scheme should work.  In a broad sense, we could argue that in the Swedish case 90% was liquidity while 10% was a real wealth problem.  Why is this important?  If we believe that the share of real wealth trouble is much bigger we should add a third leg to the program: a debt forgiveness scheme for truly insolvent debtors. Otherwise, who will be able to borrow anyway?

Bottom line: in our view, a guarantee scheme seems to depict short run low costs with long run high costs, whereas the bad bank looks as having the opposite intertemporal trade-offs.  And since the government is the sole player with the ability to tax your future income—once the economy returns to trend growth—it thus seems the only one able to correctly coordinate intertemporal expectations.  (And if the government loses money in smoothing the recovery, it is still maximizing the economy’s lifetime welfare; isn’t this its true objective?)  This is a very dynamic path; let’s hope the best possible strategy is implemented.


[1] As has been recently suggested.

2 Responses to "Toxic Assets: Bad Bank or Guarantees? Some Intertemporal Trade-Offs"

  1. charlie   February 10, 2009 at 4:00 pm

    I agree with you. The government has already committed over $3T, but not spent, to this crisis if you include all of the half-cocked programs they have set up. I just do not understand why the government does not combine all these programs and buy the toxic assets? This is the best thing that they can do. They can hold them for as long as needed, restructure loans as they wish and push this crisis into the 8th inning. In addition it will instill confidence into the market, which is what is needed. Without confidence no private money will enter the market. These fools need to stop playing politics.

  2. David Roberts   February 10, 2009 at 4:09 pm

    Thanks for posting a cogent argument for the bad bank approach. It is unfortunate that this discussion, even among economists, has adopted a vocabulary of value-laden and mis-leading terms: “toxic assets,” “bad bank,” “bail-out,” “moral hazard,” etc. Each of these phrases carries the connotation that any solution to this crisis results in a negative sum outcome for the public and a positive net return for the private owners/managers. This connotation obscures the reality that we need to choose the best social outcome among a range of ugly choices.