This paper argues that capital requirements can't prevent financial firms from becoming under-capitalized, and should be replaced with transaction-specific reserving of the type applied to insurers.

On Preventing the Next Crisis
February 16, 2009

While policy-makers and their kibitzers, among which I count myself, debate what is needed to cure the current crisis and associated recession, another debate brews in the background. It is about how to fix the system so that it doesn’t happen again.

Background: Causes of the Crisis

 Any coherent proposal for fixing the system is necessarily based on judgments about the causes of the current crisis. While there are many differences in emphasis, I believe that most observers would agree on the essentials: the crisis originated with a bubble in the residential real estate market, followed by its inevitable aftermath of declining home prices, and a subsequent explosion of home mortgage defaults and foreclosures. Because a large portion of the defaulted mortgages were in securities, the market  for which shut down, the asset value declines were several times larger than those that would have occurred if the defaulted mortgages had been held by investors in their original form. The losses were world-wide because foreign investors held enormous amounts of US mortgage-related assets. Financial institutions world-wide did not have the capital to absorb these losses, resulting in the collapse of many, and enormous infusions of capital by governments, plus loans and guarantees, to prevent the collapse of many more.

 The Critical Requirement: Adequate Capital

 This sequence of events could be prevented by blocking the bubble, or  by shoring up the capacity of the financial system to absorb the losses resulting from a bubble’s collapse. In my opinion, the second should have priority. We don’t know where the next bubble will come from, but if the system has enough capital, a crisis can be averted regardless of its source.

Why Private Firms Become Under-Capitalized

 Private financial institutions will never voluntarily carry enough capital to cover the losses that would occur under a disaster scenario.  For one thing, such disasters occur very infrequently, and as the period since the last occurrence gets longer, the natural tendency is to disregard it – to treat it as having a zero probability. In a study of international banking crises, Richard Herring and I called this “disaster myopia”.

Disaster myopia is reinforced by “herding”. Any one firm that elects to play it safe will be less profitable than its peers, making its shareholders unhappy and even opening itself to a possible takeover.

Furthermore, even if those controlling financial firms knew the probability of a severe shock, and the very large losses that would result from it, it is not in their interest to hold the capital needed to meet those losses. Because they don’t know when the shock will occur, playing it safe would mean reduced earnings for the firm and reduced personal income for them for what could be a very long period. Better to realize the higher income as long as possible, because if they stay within the law, it won’t be taken away from them when the firm becomes insolvent.

 Indeed, insolvency may not mean the demise of the firm if many firms are affected at the same time. Government can’t allow them all to fail without allowing the crisis to become a catastrophe. This is clearly borne out by the Government’s actions in the current crisis. “Bail-outs” by Government further validate the premise that it is foolhardy for a financial firm to hold the capital needed to meet the losses associated with a very severe shock.

This appears to lead logically to the conclusion that Government ought to impose capital requirements on financial firms. Capital requirements stipulate the amount of capital firms must have, based largely on the amounts and types of assets and liabilities they have.

 Why Capital Requirements Don’t Work

Unfortunately, capital requirements won’t prevent financial crises. An inherent flaw in capital requirements is that required capital varies by broad asset categories, which allows the regulated firms to replace less risky assets with more risky assets within any given asset class. The shift to sub-prime mortgages during the last bubble, for example, did not increase their required capital.

 In principle, regulators can offset this by making discretionary adjustments in the requirements in response to changing economic conditions. For this to work, however, regulators must have better foresight than those they regulate, which they don’t. Neither should we expect regulators to have the political courage to “remove the punchbowl from the party.”

An increase in capital requirements large enough to burst a bubble would be extremely disruptive, forcing many firms to sell stock at the same time, and/or to substantially reduce their lending. Concerns about such disruptions reinforce disaster myopia and political timidity among regulators.

The proof is in the pudding. Banks and other depositories have been subject to capital requirements since the 1980s but no adjustments in the requirements were made in response to the recent housing bubble.

Would Automating Capital Requirements Work?

Proposals have emerged to rectify these weaknesses of capital requirements by automating the adjustment process. This would require identifying one or more statistical measures to which capital requirements would be tied. When the measures indicated that a bubble was underway, capital requirements would increase automatically, and when the measures indicated that markets were contracting, requirements would decline.

While there are many good indicators of a contracting system that follows a bubble, there are no universal indicators of bubbles themselves. Bubbles can arise anywhere, and they can involve newly fashioned financial instruments that did not exist before. Automating capital requirements is not going to work.

An Alternative to Capital Requirements: Transaction-Based Reserving

An alternative to capital requirements that can work is transaction-based reserving, or TBS. Under TBS, financial firms are regulated as if they were insurance companies that are obliged to contribute to a reserve account in connection with every asset they acquire. The portion of the cash inflows generated by the asset that is allocated to the reserve account depends on the potential future outflows associated with the asset. For example, a life insurance company that sells a policy to a 70-year old will allocate a larger portion of the premiums it receives to a reserve account than the same policy sold to a 30-year old.

Applying TBS to a Depository

As applied to a depository, the required allocation to a contingency reserve would be, say, 50% of the portion of any charge that is risk-based. If a prime mortgage was priced at 6% and zero points, for example, the reserve allocation for a 7% 2 point mortgage might be ½% plus 1 point.

Contingency reserves can’t be touched for a long period, perhaps 15 years, except in an emergency. Of course, income allocated to reserves would not be taxable until it was withdrawn 15 years later.

Advantages of TBS

A great advantage of TBR, relative to capital requirements, is that TBR does not depend on discretionary actions by the regulator to offset the excessive optimism that feeds bubbles. A shift to riskier loans during periods of euphoria automatically generates larger reserve allocations because riskier loans carry higher risk premiums.

Another advantage of TBR is that it applies to every transaction with a risk component, whether it is shown on the firm’s balance sheet or not. The principal responsibility of the regulator is to establish the risk component of every type of transaction. When credit default swaps appeared, for example, the TBR regulator would immediately have realized that the premium was 100% risk-based, and sellers would have been obliged to reserve 50% of their premium income.

 Experience of Private Mortgage Insurance Companies


Private mortgage insurance companies (PMIs) are subject to much the same kind of mortgage default risks as depositories that invest in mortgages. But where depositories have been subject to capital requirements, PMIs have been subject to TBR.  PMIs allocate 50% of their premium income to a contingency reserve for 10 years. These reserves have allowed the PMIs to meet all their obligations in connection with the extraordinary losses suffered by lenders during the current crisis. While their shareholders have taken a beating, PMIs are doing exactly what they were chartered to do: cover losses out of their reserves.  

In retrospect, the major shortcoming of the TBR rules under which they have operated is that the 10-year period is too short. Given the infrequency of major crises, 15 years seems more appropriate.

Sign up to Receive New Articles
Print