An Insurance Complement to TARP II

Posted by Ricardo Caballero on 18 February 2009


The basic principles underlying Secretary Geithner’s sketch of a financial stabilization plan are the right ones: To stabilize the financial system without nationalization; to jump start frozen financial markets by reducing the perceived tail risk from private investments; and to reduce the risk, especially systemic risk, being held by banks and other leveraged financial institutions. Politics require that a “good deal for the taxpayers” is added to these principles, but the truth is that the best deal for the taxpayers, once one considers the endogenous response of the economy, is anything that works to stabilize the financial system, almost regardless of how many short-term transfers the financial institutions may receive.

It is true that the recent announcements are lacking specific details, and perhaps revealed that the Treasury’s economic team overestimated people’s ability to distill the good news in an abstract message of principles when in panic mode. But there is good news in them, as they reflect a much deeper understanding of the fundamental uncertainty problem ravaging insurance and credit markets than commentators and politicians have. It is time for all of us to focus on facilitating their difficult task and to try to fill some of the gaps.

My preferred (part of a) solution is to provide universal insurance for the assets that are currently clogging the balance sheets of banks and other financial institutions. There are two issues often raised in the context of an insurance arrangement of this kind: How to determine the fair price of insurance in an environment where there are no sensible market prices? And, how to prevent financial institutions from selectively insuring their worst assets without disclosing them as such?

Fortunately, once the policy decision is made that these key systemic institutions will survive no matter what, these problems have a relatively easy solution. The great advantage of dealing with long-lived institutions holding a large number of assets is that there is no need to resolve the thorny issue of the insurance price and the quality of the assets right now. We can wait for the passage of time and a return to normality to determine whether their assets were worse than the representative asset in the corresponding asset class. Concretely, I would suggest that:

  • The price of the insurance should be set at pre-crisis levels for the corresponding asset class. If there is a sense that these assets were over-rated to begin with, then we should adjust the prices accordingly (for example, use AA pre-crisis insurance prices for overly-rated AAA assets).
  • This arrangement should be coupled with tight monitoring of the insured institutions and with retroactive fines a few years down the road to those institutions (and their management) whose assets underperform relative to their asset class.

    A key aspect of this insurance arrangement is that it is aimed mainly at removing the aggregate risk from the balance sheet of financial institutions, although it does so asset-class by asset-class, so it considers the different portfolio composition of different financial institutions. Moreover, since these institutions have a large number of assets within each asset-class, one can rely on the law of large numbers to net out idiosyncratic accidents.

    One potential criticism to this mechanism is that it does not resolve an adverse selection problem faced by potential equity investors: Since banks that choose to lie now will be eventually punished retroactively, then how can investors be reassured that the bank they are choosing to recapitalize is not lying? My reaction to this valid concern is twofold: First, my sense is that at this point in time the main risks are aggregate rather than of this kind. Thus, if enough aggregate uncertainty is removed from their balance sheets, the banks will not need, for the time being, much new capital. Second, the aggregate insurance arrangement need not work in isolation, but it can be implemented very quickly and be used as a stopgap. It can then be followed by the much slower process of getting onto the books of the banks to build more public trust. Moreover, by proceeding in this sequence, the stress tests do not put the burden of further negative aggregate shocks on the banks balance sheets, and hence further reduce the need for nationalizations and other remedies that have brought great distress to equity markets and taxpayers wealth in recent days.

    I view this insurance proposal as complementary to and in the spirit of Secretary Geithner’s two-step proposal. On one hand, the former appears to have two advantages over the latter: First, it is simpler and faster to implement since it is done in one step, instead of having to go through a bad bank and then enticing the private sector to buy these assets. My sense is that the non-banks private sector will require more insurance than the banks themselves, since they will be exposed to an additional layer of adverse selection as they will be buying assets from the banks. Second, it immediately raises the tangible capital of the distressed financial institutions, without further need for capital injections. On the other hand, the two-step approach also has two advantages over the single-step insurance proposal: It addresses, more directly, the illiquidity problem in underlying asset markets, and it provides a channel to inject resources into the harder-to-get-to shadow financial system.

    Clearly, there is no single dominant strategy to deal with the current crisis, but whatever we do, it is central that we focus on the whole system and on the role played by uncertainty. Standard recipes used to deal with isolated problems in a bank or two can easily backfire. Nationalizations and related measures are fine solutions in partial equilibrium, but they are recipes for disaster in an environment where systemic risk and uncertainty are at the core of the problem.


    Ricardo J Caballero,
    Head of the department of economics at MIT

    Editor's note:  This column first appeared on the Real Time Economics Forum, 17 February 2009