In the past few years, several important financial regulations have been struck down by the D.C. Circuit Court of Appeals because the regulatory agency failed to prove that the benefits of those regulations exceeded the costs. There is no current explicit legal requirement for financial agencies to conduct cost-benefit analyses, but given vagaries in the underlying statutes, the Court has felt that it has the authority to insist on a greater degree of economic rigor than agencies often display. In a parallel development, Senator Shelby has introduced a bill that would explicitly require financial agencies to perform cost-benefit analyses. If the bill is enacted, we will see even greater bloodshed in the courts.
Until these recent judicial decisions, financial regulators did not believe that they were required to perform cost-benefit analysis on regulation. This is not to say that they made decisions by flipping a coin. They relied on a mix of economic theory, published empirical studies, and intuition when deciding how to regulate. But they lacked a methodology for systematically evaluating all the costs and benefits of a regulation, based on reasonable estimates derived from statistical analysis of empirical data. The adverse judicial decisions have set them scrambling to find a rigorous methodology. Unfortunately, no such methodology exists.
This state of affairs contrasts with the practices of agencies in other parts of government. Since 1981, most of the non-financial agencies—agencies like EPA and NHTSA—have used cost-benefit analysis. They have been required to do so by a series of executive orders, as well as some statutory mandates. Over the years, economists have helped them develop a relatively rigorous protocol. However, because most of the financial agencies are so-called independent agencies, they did not regard themselves as bound by the executive order, and so did not use cost-benefit analysis.
In our paper, Benefit-Cost Analysis for Financial Regulation, we sketch out what a financial cost-benefit analysis might look like. Obviously, all financial regulations impose costs on regulated parties, who either must pay more to transact or must refrain from certain transactions. It is straightforward to quantity these costs.
However, quantifying benefits is more challenging. We identify three main benefits. First, many financial regulations are designed to reduce the risk of a systemic crisis. A systemic crisis predictably leads to an economic downturn, and historical data suggest that the cost is 1 to 20 percent of GDP. A greater challenge is to estimate the effect of a particular regulation on the probability of a crisis. But a first step would be to require agencies to use a uniform cost of a statistical crisis and state clearly their estimates about the extent to which the regulation will reduce the risk of a crisis.
Second, financial transactions can produce both positive and negative informational externalities. A well-known phenomenon is that people may overinvest in financial information (illustrated by investment in high-speed trading) and may underinvest in financial information (because the price-effect of their behavior is public). Thus, regulations that (for example) improve market liquidity can both lead to financial overinvestment and underinvestment. We provide a simply framework for enabling agencies to use empirical data in order to estimate the net effect of their regulations.
Third, financial transactions can be used both to spread risk (insurance) and concentrate risk (gambling). Risk-spreading is socially valuable, while gambling is socially costly, either because it represents transfers along with transaction costs, or it contributes to systemic risk. We show how agencies can estimate whether a regulation will have a greater effect on gambling or insurance, and point the way toward incorporating these estimates into a cost-benefit analysis.
The full paper is available for download here.