Print
E-MailAlex Lee is Professor of Law at Northwestern Pritzker School of Law. This post is based on a recent article, forthcoming in the Wake Forest Law Review.
How might a financial regulator, such as the SEC, engage in an empirically informed rulemaking? This question has been an interest of mine since my days of working at the SEC. For example, how exactly would empirically informed rulemaking work in a setting where a regulatory agency seeks to adopt a rule of first impression—a rule for which the agency (as well as the industry) lacks data to support its position?
There are a few possible approaches. One approach is for the agency to reason by way of analogy: the agency can try to argue that its new rule will operate in a similar manner as another known regulation that has been tried and tested. If the agency is lucky, it may even be able to cite an empirical study that documents the effectiveness of this other regulation. Imperfect as it is, even this option is not always available. The fact is that there isn’t always a similar rule out there for each new proposed rule. Another approach is for the agency to rely on a trial regulation: the agency can adopt a version of the rule on an experimental basis, assess the rule’s effectiveness and efficiency after some time, and then adopt a final version of the rule informed by the industry’s compliance experience. This would be considered an ex post approach—in the sense that the agency would gather compliance data from the industry after the rule has been in effect for some time. While promising, this approach has two limitations. First, reliable compliance data may not become available for a long time. Second, it is difficult to use this approach for decisionmaking purposes when the rule’s effects are irreversible.
In my forthcoming article, I suggest a novel ex ante approach for the SEC—a rulemaking mechanism based on empirical information that can be gathered before rule adoption. My proposal is that when the SEC seeks to adopt a rule for which reliable data on point is lacking and there are strong opposing views about the rule’s likely effect, it should consider directing its staff economists from the Division of Economic and Risk Analysis to examine stock market reactions to any intervening events that present surprise elements—to the market—about the rule. In some cases, the agency’s issuance of the notice of proposed rulemaking may itself qualify as such an event—if the agency has otherwise tightly controlled the flow of information to the market. In other cases, such events may arise without any act on the part of the agency—for example, as a result of certain unforeseen acts of Congress. In still other cases, the agency can strategically initiate them through carefully planned rule-related announcements (of previously concealed information) or by means of clarifying the scope of the rule’s application. (The article discusses the SEC’s “proxy access” rule from 2010 as an example of a rule for which this mechanism would have been useful.)
My suggestion is that such an event study should constitute the default starting point of the regulatory dialogues that take place during the comment period. The ensuing discussion can in turn inform (i) the agency’s decision as to whether to adopt the rule, to abandon the rule, or to modify the rule, and (ii) the agency’s final cost-benefit analysis in case it decides to adopt a version of the rule. More specifically, the discussion can begin with a rebuttable presumption that positive (negative) market reactions to announcements that increase the likelihood of rule adoption or the scope of rule’s applicability indicate that the regulation is expected to be net beneficial (costly) to investors at large. If the SEC’s study turns out to present a finding that is adverse to the agency’s position but the agency nonetheless seeks to adopt the rule, the agency should state the grounds for rebutting the presumption afforded by the finding or the grounds for adopting the rule despite the adverse finding. Of relevance, note that a rule that is net beneficial (costly) to investors is not necessarily an efficient (inefficient) one from the perspective of cost-benefit analysis. As I’ve previously discussed elsewhere, there may be genuine conflicts at times. For this reason, it makes sense for the agency to include the study’s finding as merely one factor of consideration, rather than having the finding dictate the agency’s policy decision.
To be sure, event studies are hardly without flaws and even the most carefully conducted studies may prove to be incorrect in terms of their predictions. Nevertheless, a rulemaking mechanism that permits disagreements regarding relevant empirical measures is admittedly a significant improvement over a mechanism in which interested parties, with no data on point, freely disagree over the likely equilibrium state that will materialize upon rule adoption. Put differently, such an event study would provide one quantifiable empirical data point (which is better than having no empirical data), and at a minimum, the resulting discussion will likely be more constructive and focused.
An astute reader at this point may raise a potential concern. Wouldn’t the agency’s reliance on the market’s reaction influence the way the market may react to these announcements? Perhaps speculators who would normally trade on a rule proposal might refrain from trading altogether out of the concern that there is a high degree of uncertainty as to whether the rule will get adopted. Alternatively, one might worry about the possibility of market manipulation: an investor who holds a short position in a firm and would thus benefit from a decrease in the firm’s stock price, upon learning that the proposed regulation would be beneficial for the firm, may be motivated to sell a large quantity of the firm’s stocks in the hope that the rule will be abandoned. Such transactions can send mixed signals to the SEC.
A relevant theoretical inquiry, therefore, is whether such a rulemaking mechanism can remain effective even if the investing market were to anticipate the SEC’s potential reliance on such data. Fortunately, this question has been studied, in various settings, by a number of recent finance articles that analyze the “feedback effects” of financial markets. In a companion paper, I build a model to analyze this specific question and derive findings that are broadly consistent with other feedback-effects models: aggregate stock market reactions can remain informative for the regulator as long as the regulator retains some degree of independence and commits to relying on the market’s reaction only partially. The narrow conclusion of the model can be stated as follows: If, in the process of agency rulemaking, an event arises that can provide useful information about the market’s expectation of the rule’s effect—under the assumption that the market does not expect the regulator to rely on its reactions—then the same event can still be expected to provide useful information (in most instances) even if the market were to expect the regulator to rely on the aggregate market’s reaction to guide its decisionmaking. One take-away is that while the regulator should be cautious about identifying an event that is clean and informative about the rule’s effect, given such an event, she can dependably rely on the event study’s finding without worrying that her reliance may be compromising the integrity of market prices.
The complete article is available here.
