Behavioral law and economics has arisen to international prominence; between Cass Sunstein’s appointment to head the Office of Information and Regulatory Affairs the United Kingdom’s appointment of a “nudge” bureau, behavioralism has enjoyed a meteoric impact on policymakers. Thus far, behavioral economists have almost exclusively focused on the myriad foibles or purported cognitive errors which hamper consumer decision-making. These traits include “optimism bias,” the tendency for an individual to underestimate the likelihood of negative results from their behavior, and hyperbolic discounting, where individuals reveal time-inconsistent preferences (often by over-valuing immediate consumption, at least as measured against some third party’s valuation).
Several recent proposals seek to import these observations wholesale into the antitrust context. Advocates of “behaviorally-informed” antitrust consistently note the potential presence of one or more individual errors to conclude the necessity of greater interventionism in antitrust, including a Leegin repealer, aggressive Section 2 enforcement, or utilizing Section 5 to “gap-fill” (or intrude upon) Supreme Court precedent hostile to usage of the antitrust laws as a general code of business regulation. Even conceding the most charitable set of assumptions to the behavioralist enterprise – that behavioral errors are consistent and can be documented within individuals; that these errors persist in markets and are not learned away over time; that these errors aggregate to the firm level; and, necessarily, that these errors are somehow not simultaneously visited upon putative regulators – this attempt is misguided. In a recent article with Professor Wright, we argue that there is not yet a behaviorally-informed theory of economic behavior relevant to antitrust. Behavioralist antitrust proposals suffer from a fatal theoretical failure.
Antitrust-relevant models of firm behavior necessarily depend on examining strategic behavior between incumbent and entrant firms. Timely and sufficient entry can dissipate monopoly profits, destabilize cartels, and ameliorate the anticompetitive effects of mergers. Any potential attribution of behavioralist biases to firm behavior must therefore similarly account for the presence of behavioralist biases among both incumbents and entrants. The typical application of behavioral economics to antitrust is a straightforward enterprise: behavioralists assume – without proving – that incumbent firms are irrational, that this irrationality necessarily leads to more predation than rational choice models would predict, and entrants, potential or actual, do not compete away rents from this predation. I summarize these three assumptions as the “Naïve Model” of competition.
The behavioralist enterprise within antitrust depends on this Naïve Model. There are only four possible distributions of biases between potential incumbents and entrants: either neither incumbents nor entrants suffer behavioral biases; only incumbents suffer behavioral biases (the Naïve Model); or both entrants and incumbents suffer behavioral biases. The first possibility is essentially the status quo for antitrust enforcement: neither incumbent firms nor entrant firms suffer behavioralist biases in a regulatory-relevant way. This defeats the behavioralist enterprise. The last example, in contrast, leads to a behaviorally-inspired nullification of effects. Irrationally optimistic – and therefore aggressive – incumbents predate more than rational choice models would predict are met with irrationally optimistic incumbents who enter markets more often than rational choice models would predict, thereby competing away monopoly profits. Irrationally risk-averse entrant firms that do not deter monopolization or inefficient mergers as robustly as rational choice models predict do not need to do so; they are offset by irrationally risk-averse incumbents, who are necessarily more hesitant to engage in monopolistic conduct. This offsetting behavior holds for the entire panoply of proposed behavioralist biases.
Behavioralist theories potentially demonstrate relevance when one assumes a heterogeneous, rather than homogeneous, distribution of biases across firms. Yet even here, behavioralists predicate their theories on a conceptually counterintuitive heterogeneous distribution. One could posit a world in which entrant firms suffered myriad behavioralist biases, while incumbent firms behaved more or less rationally. This heterogeneity would lead to either too much or too little predation, depending on the bias assumed. On a macrocosmic level, forging an antitrust-relevant theory of behavioral economics would require offsetting these competing vectors in various product and geographic markets. This model would at least map onto some intuitive model of dynamic learning effects; incumbent firms enjoyed their rationality by virtue of extensive forays into various irrational behaviors, while irrational incumbents learned the market through trial-and-error. (Of course, that this behavior does not suggest “irrationality” in any meaningful sense is a matter for another post; we must even charitably construe “irrationality” broadly enough here to encompass mere ignorance.) Problematically, the overwhelming majority of behavioralist biases – cited in the consumer protection and credit literature – focus on excessively aggressive present-tense behavior. Indeed, as Avishalom Tor correctly notes, behaviorally-motivated entrants should behave more aggressively than expected, serving as the “cannon fodder” of innovation. If anything, this behaviorally-inspired model suggests that rational choice models of antitrust over-deter behaviorally-afflicted entrant firms. The only combination by which the behavioralist proposition leads to the behavioralist prescription – that is, additional antitrust intervention – is the Naïve Model: an assumption that incumbent firms behave irrationally while entrant firms act rationally. This naturally requires some inherent “unlearning” effect, as all incumbent firms of course began their existences as entrants, and thereby must acquire such a behavioral defect sometime in media res.
The critical failure with the behavioralist narrative is one of under-theorization. The behavioralist literature does not yet articulate a theory predicting the necessary or sufficient conditions for which antitrust scholars and regulators may observe a given behavioral bias within firms. More critically, the behavioralists have yet to explain why or when a given behavioral bias may be distributed across firms within a market so as to justify the Naïve Model. This is, however, merely one failure of behavioral law and economics due to under-theorizing. Behavioralists have yet to propose even theoretically supportable reasons to presume that individual biases aggregate to firms, that these biases persist maintain despite learning incentives within the marketplace, and that regulators do not suffer these biases as well. Behavioral economics finds itself a victim of its own rapid success: having ascended to prominence within the academy in the last decade, the Obama administration articulates behavioralism as the intellectual lodestar of its regulatory agenda. Without a more rigorous examination of its theoretical underpinnings, however, behavioral law and economics threatens to act as, as one behavioralist advocate put it, “simply liberalism masquerading as economic thinking.” Until behavioralists satisfy these conceptual lacunae, behavioral economics must tentatively remain irrelevant to antitrust.