The Regulator's Calculation Problem

If federal regulators are thought to have better judgment about risk than the bankers themselves (due to the bankers’ presumed moral-hazard problems), then there really is no reason to allow private banking to continue.

— Jeffrey Friedman and Wladimir Kraus1

We tend to think of regulated markets as if they were a far cry from socialism. We think of the United States as having a free-market economy, partially “tamed” by regulation.

A careful reader of Engineering the Financial Crisis, by Jeffrey Friedman and Wladimir Kraus, would reconsider. The book shows that regulators face the socialist calculation problem. As Ludwig von Mises and Friedrich Hayek pointed out during the socialist calculation debate, central planners lack the information that is produced by markets. By over-riding market prices and substituting their own judgment, regulators incur the same loss of information.

The particular regulations that Friedman and Kraus discuss are the risk-based capital rules that U.S. bank regulators implemented in line with the Basel Accords reached in 1986. They show that these rules contributed to the housing bubble and to the fragile financial mortgage financing system behind it.

The Basel agreement was a response to two financial crises of the previous decade. One was the Latin American debt crisis, in which the world’s largest banks had over-extended themselves in lending to that region. The other was the savings and loan crisis, in which thrift institutions in the United States, already reeling from the impact of rising interest rates on their ability to finance their fixed-rate mortgage portfolios, proceeded to go even deeper in the hole pursuing high-yield bonds and commercial loans.

These crises convinced regulators that they needed tools for controlling risk-taking at financial institutions. Moreover, they sought international harmony among rules, so that no one country would be a weak link in the international chain of financial regulation.

Capital Rules and the Financial Crisis

Other things being equal, the probability of insolvency at a bank increases with its use of financial leverage, meaning the ratio of debt (including deposits) to equity. Other things being equal, the probability of insolvency at a bank also increases with its use of operating leverage, meaning the riskiness of its assets. If regulators only limit financial leverage, then banks may offset this through increased operating leverage, and conversely. The Basel regulations were an attempt to influence both types of leverage.

Financial leverage was addressed by putting a floor under the allowable ratio of equity to total assets in banks. This ratio was set at eight percent. Operating leverage was addressed by assigning risk weights to different classes of assets. Ordinary commercial loans were given the maximum weight of one. Assets deemed safe, including gold, cash, and government bonds, were given a weight of zero. Mortgage loans secured by owner-occupied housing were given a weight of 0.5, while a weight of 0.2 was assigned to the securities of government-sponsored enterprises, such as Freddie Mac and Fannie Mae.

For example, if a bank had $100 million each of commercial loans, government bonds, mortgage loans, and Fannie Mae securities, its risk-weighted asset total was

$100 million times 1.0 +

$100 million times 0.0 +

$100 million times 0.5 +

$100 million times 0.2

= $170 million

With a capital requirement of 8 percent, this meant that the bank must have at least $13.6 million in equity, with a maximum of $386.4 million in deposits and other debt liabilities.2

By assigning risk weights in this fashion, the regulators were influencing the rates of return on the different asset classes. For example, suppose that the required return on equity for a bank is 12 percent, but it pays 4 percent interest on its debt. In that case, an asset funded 8 percent by equity and 92 percent by debt will have a blended financing cost of 4.640 percent. An asset with only a 20 percent risk weight would require only 1.6 percent in equity and 98.4 percent debt, for a blended financing cost of 4.128 percent. In this example, the bank can accept a lower interest rate of about one-half of one percent on the asset that falls into the low-risk bucket, simply based on regulatory capital requirements. The larger the difference between debt and equity cost, the greater the extent that banks are steered toward assets deemed low risk by regulators.

In the late 1990s, regulators confronted an issue raised by private asset-backed securities (as opposed to securities issued by government-sponsored enterprises). Wall Street firms were carving these securities into tranches, with the “equity” tranche bearing nearly all of the risk, junior tranches bearing a small share of the risk, and senior tranches bearing almost no risk. If a bank held an equity tranche worth $3 million on a $100 million security, then even 8 percent of $3 million would provide far too little protection. On the other hand, if a bank held a senior tranche, then 8 percent capital would be excessive.

The solution was the Recourse Rule, which was finally adopted in 2001. For equity tranches, and for other situations in which the bank was deemed to have provided “recourse” to another party that had nominally purchased the asset from the bank, the rule required that the 8 percent capital rule be applied to the full notional value of the asset. In the case of an equity tranche worth $3 million on a $100 million security, this meant 8 percent of the full $100 million, or $8 million in capital would be required.

For junior and senior tranches, the Recourse Rule assigned risk weights according to ratings given by the major rating agencies, such as Moody’s or Standard and Poor’s. In particular, AA or AAA rates tranches were given the same 20 percent risk weight as securities issued by Fannie Mae.

The favorable treatment of highly-rated securities unleashed a flood of securitization. As one paper puts it,

… there is an almost threefold increase in new issuance from 2002 to 2007. In the aggregate, securitization worldwide went from $767 billion at the end of 2001 to $1.4 trillion in 2004 to $2.7 trillion at the peak of the “bubble,” in December of 2006. By late October, 2008, the market had effectively collapsed.3

Friedman and Kraus point out that the sudden collapse of the securities market was exacerbated by the interplay between accounting rules, capital regulations, and securities downgrades. When default rates rose on some securities, the prices of those securities fell, which required banks to mark down their asset values, reducing bank equity. Rating agencies downgraded the rating on these securities, which raised capital requirements on the securities. Thus, banks found themselves with less equity and higher capital requirements, which led them to attempt to sell asset-backed securities, driving prices down further.

In hindsight, it is clear that large errors were made by the private sector, including the rating agencies. Most participants placed too much faith in the notion that house prices would not decline nationwide and that the diversification and risk structuring of mortgage securities could create low-risk mortgage-backed securities almost regardless of the nature of the underlying loans.

Friedman and Kraus emphasize that regulators were no more aware of the dangers than were executives in the private sector. They quote Ben Bernanke in 2005 saying,

House prices have risen nearly 25 percent over the past two years. Although speculative activity has increased in some areas, at a national level these price increases largely reflect strong economic fundamentals, including robust growth in jobs and income, low mortgage rates, steady rates of household formation, and factors that limit the expansion of housing supply in some areas.4

Because regulators were as relaxed as private-sector participants regarding the risks building in the mortgage-backed securities market, Friedman and Kraus argue that risk-based capital rules were counterproductive. Rather than reducing systemic risk, the risk weights steered banks toward holding the very securities that proved to be most problematic during the crisis.

Centralizing risk assessment through regulatory risk weights and rating agency designations has several weaknesses. Local knowledge, such as detailed understanding of individual mortgages, is overlooked. At a macro level, regulators’ judgment of housing market prospects were no better than those of leading market participants. Moreover, regulators imposed a uniformity of risk judgment, rather than allowing different assessments to emerge in the market. As Friedman and Kraus put it,

The cognitive limitations of capitalist decision makers are, of course, just as likely to lead to mistakes. But capitalists may simultaneously put into practice heterogeneous, competing interpretations of the world—all of them fallible, all of them ignorant of most aspects of the world, all of them in one sense or another, “ideological.” At least some of these interpretations, if not all of them, are very likely to be mistaken, but some may be less mistaken than others.

Society might therefore be well advised to diversify its asset portfolio by allowing capitalist competition to proceed unhindered, rather than by predicting that a single interpretation is best in advance and then imposing it on all capitalists’ behavior at once.5

Regulation in General

All regulation is subject to the problems that arose with the regulation of risk-based capital at banks. Like socialist planning, regulation substitutes centralized decision-making for decentralized market processes. Whether we are talking about food and drug regulation, rules for health insurance, taxes and subsidies intended to influence energy choices, a minimum wage, or any other sort of economic regulation, the judgment of a policy maker over-rides that of the private sector.

Centralization in any realm creates at least two problems. One is information loss due to distance from local considerations. As Hayek pointed out, individuals have tacit knowledge that they cannot even articulate, so it will not reach the regulator. In the case of mortgage loans, the centralized assessments of the rating agencies and of bank regulators ignored what an individual mortgage underwriter might have known about the local housing market or the characteristics of the borrower.

The second problem is the reduction in diversity in the ecosystem. As Friedman and Kraus emphasize, we live in a world of ignorance. Individuals and organizations form opinions about unknown characteristics of the current environment and about future prospects. When differences of opinion are allowed to play out in a market economic system, the mistakes that some participants make are mitigated to some extent by the less erroneous assessments of other actors. However, a centralized regulator’s mistakes will be imposed on everyone at the same time.

If we turn away from centralized regulation, it does not mean that we must fall back on some sort of brutal system in which unregulated individuals and organizations exploit weaker individuals with impunity. Instead, a competitive regulatory system can emerge. For example, instead of requiring a single regulator’s approval for all medical drugs and devices, we could allow for competing agencies to certify the safety and effectiveness of treatments. Individual consumers could then make choices of remedies based on the certifications those remedies have received from these competing agencies.

In considering regulatory policy, we need to be more aware of the socialist calculation problem. Friedman and Kraus make a convincing case that the dangerous boom in mortgage-backed securities and house prices was exacerbated rather than mitigated by centralized regulations enacted with the intention of reducing bank risk. Once we recognize that regulators can be fallible, we will be inclined to look for alternatives to centralized regulation.