Law and Economics
February 28 and March 1, 2013
Alexander Dyck, University of Toronto, and Adair Morse and Luigi Zingales, University of Chicago and NBER
Dyck, Morse, and Zingales estimate what percentage of firms engage in fraud and at what economic cost. Their estimates are based on both detected frauds and frauds that they infer are started but are not caught. They take advantage of an exogenous shock to the incentives for fraud detection: Arthur Andersen's demise, which forces companies to change auditors. By assuming that the new auditor will clean house, and examining the change in fraud detection by new auditors, they infer that the probability of a company engaging in a fraud in any given year is 14.5 percent. They validate the magnitude of this estimate using alternative methods. They further estimate that on average corporate fraud costs investors 22 percent of enterprise value in fraud committing firms and 3 percent of enterprise value across all firms.
Edward Glaeser, Harvard University and NBER, and Cass Sunstein, Harvard Law School
Many studies find that presentation of balanced information, offering competing positions, can promote polarization and thus increase preexisting social divisions. Glaeser and Sunstein offer two explanations for this apparently puzzling phenomenon. The first involves what they call asymmetric Bayesianism: the same information can have diametrically opposite effects if those who receive it have opposing antecedent convictions. Recipients whose beliefs are buttressed by the message, or a relevant part, rationally believe that it is true, while recipients whose beliefs are at odds with that message, or a relevant part, rationally believe that the message is false (and may reflect desperation). The second explanation is that the same information can activate radically different memories and associated convictions, thus producing polarized responses to that information, or what we call a memory boomerang. An understanding of these explanations reveals when balanced news will produce unbalanced views. The explanations also account for the potential influence of "surprising validators." Because such validators are credible to the relevant audience, they can reduce the likelihood of asymmetric Bayesianism, thus promoting agreement.
Scott Baker, Washington University School of Law, and Albert Choi, University of Virginia Law School
An infinitely-lived firm sells a product to a sequence of short-term consumers, where the quality of that product depends imperfectly on unobservable effort exerted by the firm in each period. The firm can solve the moral hazard problem by relying on either formal or informal sanctions. By promising to pay damages, the firm induces consumers to sue when product quality is low. Lawsuits also generate information that future consumers can use to unleash informal sanctions: the nature of the information depends on the liability regime chosen by the firm (no-fault versus fault-based damages). The firm makes a trade-off between the litigation costs of formal sanctions and inefficient failures to trade under informal sanctions. Baker and Choi construct a model that demonstrates that even when the firm can solve the moral hazard problem with only formal (informal) sanctions,the firm may deliberately rely on both types of sanctions. They also compare no-fault and fault-based liability regimes (negligence versus strict liability) and extend the analysis to settings where the remedy is chosen by law.
Louis Kaplow, Harvard Law School and NBER
In many settings, there are preliminary or interim decision points at which legal cases may be terminated: for example, motions to dismiss and for summary judgment in U.S. civil litigation, grand jury decisions in criminal cases, and agencies' screening and other exercises of discretion in pursuing investigations. Kaplow analyzes how the decision whether to continue versus terminate should optimally be made when: 1) proceeding to the next stage generates further information but at a cost to both the defendant and the government; and 2) the prospect of going forward, and ultimately imposing sanctions, deters harmful acts and also chills desirable behavior. This subject involves a mechanism design analogue to the standard value of information problem, one that proves to be qualitatively different and notably more complex. Numerous factors enter into the optimal decision rule some expected, some subtle, and some counterintuitive. The optimal rule for initial or intermediate stages is also compared to that for assigning liability at the final stage of adjudication.
Joshua Fischman, Northwestern University School of Law
Scholars have long debated whether, and to what extent, law constrains judicial decisions. Because law cannot be objectively measured, many believe that judicial decisions cannot be empirically evaluated on grounds internal to the practice of law. Fischman demonstrates that empirical analysis of judicial decisions nevertheless can provide objective, albeit limited, conclusions about subjective criteria for evaluating a system of adjudication. He begins by formalizing three criteria: inter-judge inconsistency, legal indeterminacy, and judicial error. He then clarifies what can be learned about these criteria from observational data on single-judge adjudication. The precise level of inconsistency cannot be identified, but it is possible to estimate a range of feasible values. Similarly, rates of indeterminacy and error cannot be estimated in isolation, but it is possible to estimate a curve that identifies feasible combinations of these rates. The methodologies developed in this Article are illustrated using data on immigration adjudication.
Steven Shavell, Harvard Law School and NBER
The central point that Shavell makes here is that legal change often should be attenuated - that is, should be less rigorous than conventionally efficient legal change - in order to reduce risk-bearing burdens. This conclusion rests on two arguments. First, insurance against legal change is largely unavailable (primarily because of the correlated nature of the losses usually caused by legal change). Second, given the unavailability of such insurance coverage, it is desirable in principle for legal change to be less than conventionally efficient when the parties governed by it are risk averse.
Special Session on Corporate Governance
John Matsusaka and Oguzhan Ozbas, University of Southern California
Matsusaka and Ozbas develop a theory of corporate decision making in order to study the benefits and costs of shareholder empowerment. They show how permitting shareholders to propose directors or policies can cause value-maximizing managers to take value-reducing actions in order to accommodate activist investors with non-value-maximizing goals. Their model identifies an important distinction between the right to approve and the right to propose. The right to approve is weak though always beneficial; the right to propose does have an impact but can help as well as hurt shareholders. The authors identify implications for current policy discussions about director elections, proxy access, bylaw amendments, and shareholder voting.
Martijn Cremers, University of Notre Dame, and Allen Ferrell, Harvard Law School
Cremers and Ferrell explore the robustness of the positive association between shareholder rights and abnormal stock returns (using the Fama-French-Cahart four factor model) and potential explanations for that relationship. Using hand-collected shareholder rights data for the period 1978-89, together with existing post-1990 RiskMetrics data, they show that: 1) over the entire 1978-2007 period, the association is generally robust to a variety of controls and to estimating abnormal returns at the portfolio or the firm-level; 2) this association co-varies with merger and acquisition (M&A) waves; 3) while being acquired and making acquisitions are both strongly associated with abnormal stock returns, these effects do not explain the positive association; and 4) once the four-factor model is supplemented with the Cremers, Nair, & John (2009) takeover factor which captures risk associated with time-varying investment opportunities and thus relates to the state of the M&A market the association disappears.
Fabio Braggion, Tilburg University, and Mariassunta Giannetti, Stockholm School of Economics
An intense debate on the use of non-voting shares developed in the United Kingdom in the second half of the twentieth century. Using a unique hand-collected dataset, Braggion and Giannetti ask to what extent variations in firms' fundamentals, vis-a-vis changes in investor preferences toward non-voting shares, explain the relative price of voting and nonvoting shares (that is, the voting premium). They show that negative news coverage of dual class firms is associated with an increase in the voting premium, even if no new material information has been revealed. Furthermore, a higher voting premium and negative news for dual class firms are followed by lower returns for voting shares than for non-voting shares, which suggests a reversion to fundamentals. These results indicate that, during this period, market participants may have started to consider non-voting shares as inferior claims, and that this limited firms' ability to use dual class shares.
Kose John, New York University, and Dalida Kadyrzhanova, University of Maryland
John and Kadyrzhanova document a novel agency cost that arises because managers of potential takeover targets forgo merger opportunities in industry merger waves. They present comprehensive evidence that the entrenchment effect of a classified board varies dynamically over time by industry. While the effect is strongly economically significant in years when industries are undergoing a synergistic merger wave, it is muted in years when synergistic industry M&A activity subsides. In wave industry-years, firms without a classified board are more than three times as likely to receive a takeover bid than firms with a classified board. This difference is even larger for less anticipated waves and for firms that also have a high level of takeover protection, based on the GIM index of Gompers, Ishii, and Metrick (2003). By contrast, the difference in takeover odds is an order of magnitude smaller and not statistically significant in non-wave industry-years. These results are driven by economic, technological, and regulatory shocks that create economic opportunities to merge in the industry. Overall, the evidence here broadens the classical agency view and suggests that the agency cost of classified boards varies significantly over time.