حماقت های تحت کنترل : بهره وری اطلاعاتی تحت انتظارات تطبیقی ​​و گرایش های تاییدی

We study the informational efficiency of a market with a single traded asset. The price initially differs from the fundamental value, about which the agents have noisy private information (which is, on average, correct). A fraction of traders revise their price expectations in each period. The price at which the asset is traded is public information. The agents’ expectations have an adaptive component and a social-interactions component with confirmatory bias. We show that, taken separately, each of the deviations from rationality worsens the informational efficiency of the market. However, when the two biases are combined, the degree of informational inefficiency of the market (measured as the deviation of the long-run market price from the fundamental value of the asset) can be non-monotonic both in the weight of the adaptive component and in the degree of confirmatory bias. For some ranges of parameters, two biases tend to mitigate each other’s effect, thus increasing informational efficiency.

“Investing in speculative assets is a social activity. Investors spend a substantial part of their leisure time discussing investments, reading about investments, or gossiping about others’ successes or failures in investing. It is thus plausible that investors’ behavior (and hence prices of speculative assets) would be influenced by social movements” Shiller (1984) In most economic interactions, individuals possess only partial information about the value of exchanged objects. For instance, when a firm “goes public”, i.e. launches an initial public offering of its shares, none of the participants in financial market has complete information concerning the future value of the profit stream that the firm would generate. The fundamental question, going back to Hayek (1945), is then: To which extent the market can serve as the aggregator of this dispersed information? In other words, when is the financial market informationally efficient, meaning that the market price converges over time to the value that would obtain if all market participants had full information about the fundamental value of the asset exchanged? Most studies that address this question build on the assumption that individual market participants are fully rational. Under full rationality, the seminal results on the informational efficiency of centralized markets were established by Grossman, 1976 and Wilson, 1977 and Milgrom (1981), whereas for decentralized markets they have been proved by Wolinsky, 1990 and Blouin and Serrano, 2001 and Duffie and Manso (2007). However, research in experimental economics and behavioral finance indicates that traders do not behave in the way consistent with the full-rationality assumption. For instance, Rabin and Schrag (1999) discuss the evidence that individuals suffer from the so-called confirmatory (or confirmation) bias: they tend to discard the new information that substantially differs from their priors. One model that captures this kind of deviation from full rationality is proposed by Brock and Durlauf (2001). They introduce a setup in which individual utility exhibits social interaction effects: the individuals desire to conform to the behavior of the social groups to which they belong. One requirement of this approach is that each agent observes the behavior of a large number of other individuals. As noted by Shiller (1984), individuals actually update their prior information in (mainly bilateral) discussions with others. Therefore conforming to some average ‘social’ behavior, or information, is unlikely to occur in an environment consisting of bilateral interactions. Along a different dimension, Haruvy et al. (2007) find that traders’ expectations are adaptive, i.e. they give more importance to the past realized price of the asset than the fully-rational agent would. This constitutes a deviation from full rationality because (under full rationality) past prices cannot serve as predictors of future prices. Understanding whether (and under which conditions) the financial markets are informationally efficient when agents do not behave fully rationally remains an open question. From the policy perspective, it is important to understand if asset prices bubbles derive from incomplete information (and therefore increasing information flows would solve the problem) or from the irrationality of agents (in which case a different policy approach should be designed). In this paper, we study the informational efficiency of a market with a single traded asset, in which agents can have both aforementioned forms of deviation from full rationality. The price, which is public information, initially differs from the fundamental value, about which the agents have noisy private information (on average, correct). A fraction of traders revise their price expectations in each period giving some weight to the past prices and also exchanging opinions about future prices in a social interaction with another agent. Integrating new information from social interaction is subject to a certain degree of confirmatory bias. We show that, taken separately, each of the deviations from rationality worsens the information efficiency of the market. However, when the two biases are combined, the degree of informational inefficiency of the market (measured as the deviation of the long-run market price from the fundamental value of the asset) can be non-monotonic both in the weight of the adaptive component and in the degree of confirmatory bias. In other words for some range of parameters, the two biases tend to mitigate each other’s effect, thus increasing informational efficiency. The paper is structured as follows: Section 2 presents the setup of the model. Section 3 derives analytical results for each bias taken separately. In Section 4, we present the simulation results when two biases are combined. Finally, Section 5 discusses the implications of our results and suggests some future avenues for research.

This paper has studied the informational efficiency of an agent-based financial market with a single traded asset. The price initially differs from the fundamental value, about which the agents have noisy private information (which is, on average, correct). A fraction of traders revise their price expectations in each period. The price at which the asset is traded is public information. The agents’ expectations have an adaptive component (i.e. the past price influences their future price expectations to some extent) and a social-interactions component with confirmatory bias (i.e. agents exchange information with their peers and tend to discard the information that differs too much from their priors). We find that the degree of informational inefficiency of the market (measured as the deviation of the long-run market price from the fundamental value of the asset) can be non-monotonic both in the weight of the adaptive component and in the degree of confirmatory bias. For some ranges of parameters, two biases tend to mitigate each other’s effect, thus increasing informational efficiency. Our findings complement the well-known results in the theory of markets showing that allocative efficiency can be obtained even under substantial deviation from individual rationality of agents ( Gode and Sunder, 1993 and Gode and Sunder, 1997). We show that deviations from individual rationality, under certain conditions, can also facilitate informational efficiency of markets. The key condition for this property is that the various behavioral biases that agents possess should mutually dampen their effects on the price dynamics. Given the potential importance of this insight for financial economics, the natural extension of this work is to test its predictions experimentally. This would require to construct experimental financial markets with human traders, similar to the setting of Haruvy et al. (2007), with the additional feature of allowing agents to share their information (in some restricted form). The outcomes of interest in such an experiment would be both the evolution of market price of the asset and the elicited price expectations of traders.