Abstract
How accurately do capital markets price assets? The Efficient Market Hypothesis(EMH) posits that market prices represent the best possible estimate of true value, but advocates of behavioral finance have argued that they can deviate substantially from true value. Camerer and Fehr (2006) suggest that the resolution of this debate depends on the type of strategy used by better-informed investors, who will correct mispricings if they have an incentive to do the opposite of what other players are doing (substitution strategies), but will accentuate mispricings if they have an incentive to match other players' strategies (complementary strategies). EMH assumes that strategies in financial markets are substitutes because better-informed investors focused on objective value will buy when others undervalue an asset. Keynes(1936) suggested that strategies in financial markets are complements because better-informed investors focused on resale value will "pre-imitate" others by buying when they believe others plan to buy. I propose that a key factor determining whether investors substitute or complement is whether they construe the nature of their task as determining an objective value or anticipating subjective opinion. In four experiments, I offer evidence that: (1) individuals are more likely to match strategies when the object of judgment is the opinion of others than when it is an objective value (2) the value an individual assigns to a risky asset is influenced by the opinions of others (3) in a market environment, individuals are likelier to match strategies when relative performance is made more salient.
This research suggests that individuals in a market setting are predisposed to use complementary strategies because they tend to construe the nature of their task as anticipating subjective opinion, partly because this construal rests on a valid theoretical basis and partly because it is empirically valid when enough participants act as if it were true. I propose that a consequence of this tendency is that coordination games offer a useful model for analyzing market phenomena such as momentum beliefs.