Stress tests and the decoupling of Main and Wall Street

Ricardo Caballero 20 April 2009



As the date for the release of the stress tests on the largest 19 US banks approaches, anxiety levels and strategic positioning are on the rise. One of the main concerns (from a macroeconomic point of view) is whether the stress tests’ parameters are extreme enough given the current scenario.  Opponents of the US Administration’s CAP/PPIP program (i.e. the Capital Assistance Programme and the Public-Private Investment Program) have seized on this concern over the level of severity of the tests – which has some validity – to torpedo the credibility of the entire program. However, I argue here that a small amendment to the CAP would render this debate mute and serve to further stabilize the financial system.

Eliminate feedback between economic bad news and the financial system

While I remain broadly supportive of the financial stabilization program put in place by the Treasury, the Fed, and the FDIC, I have argued before that the CAP component leaves too much of the macroeconomic risk in the banks’ balance sheets. This remains problematic because it creates a reinforcing feedback-loop between bad news in the real economy and bad news in the financial system. This perverse loop brought the economy down to its knees during the last quarter of 2008 and first quarter of 2009, and has the potential to abruptly destabilize the incipient recovery in financial markets. This feedback-loop needs to be severed as soon as possible.

A small but necessary amendment to the CAP: Insurance, not capital

For this reason, the results from the stress tests should not be used to determine how much more capital a bank may need in an extreme scenario. Instead, it should be used to determine how much insurance the bank needs to buy from the government to protect itself against such scenarios. That is, a bank should be required to have as much capital as needed for the central scenario. If aggregate conditions are worse than expected, the government should cover the shortage of capital without equity compensation. If conditions are better than expected, the government should be paid a fee that compensates it for the insurance it provided. The government should charge fairly for this insurance, using the same probabilities it assigns to the different scenarios used in the stress tests. A weak bank needs to contract for more insurance than a strong bank.

Insurance as a substitute for capital hoarding/de-leveraging

The immediate impact of such an apparently small modification would be to encourage banks to lend because the insurance policy would replace the need for massive self-insurance induced hoarding.

It makes no sense that the aggregate risk remains on the most leveraged institutions of the system – which is precisely what got us into this mess. Instead, the way to solve this mismatch is to make fair insurance available to the banks rather than to force them to deleverage at great cost for the entire economy.




Topics:  Financial markets

Tags:  CAP, global crisis debate, stress tests, PPIP

Professor of Economics, MIT


CEPR Policy Research