Search and the market for ideas

We study a market where innovators, who are good at coming up with ideas, can sell them to entrepreneurs, who might be better at implementing them. The market is decentralized, with random matching and bargaining. Ideas are characterized by five salient features: they are indivisible; partially nonrival; intermediate inputs; subject to informational frictions; and difficult to collateralize. This last feature gives rise to a demand by entrepreneurs for liquidity. We determine which ideas get traded in equilibrium and compare this to the efficient outcome, emphasizing the impact of bargaining and liquidity considerations. Among other applications, we study how outcomes in the idea market affect the labor market.     

Introduction to Search Theory and Applications

This introduces the Symposium on Search Theory and Applications.     

A theory of commerce

The theory of commerce advanced here captures prominent features of retail trade: large employment, congestion effects, anonymous posted prices, and quantity discounts. This theory is built around a directed search model where buyers’ preferences are private information. The analytical solution is easily inserted in a Neoclassical growth framework. In this framework, the parameters of retail trade are calibrated using commercial margins and employment. Welfare properties depend on the sellers’ ability to charge two-tier prices. With two-tier prices, the directed search equilibrium is efficient. Otherwise, it is not. This contrasts with the full information benchmark, where directed search is always efficient.     

Anti-herding and strategic consultation

In this paper I analyze how careerist decision makers aggregate and use information provided by others. I find that decision makers who are motivated by reputation concerns tend to ‘anti-herding’, i.e., they excessively contradict public information such as the prior or others’ recommendations. I also find that some decision makers may deliberately act unilaterally and not consult advisers although advice is costless. Moreover, advisers to the decision maker may not report their information truthfully. Even if the advisers care only about the outcome, they bias their recommendation since they anticipate inefficient anti-herding behavior by the decision maker.     

Wage and price controls in the equilibrium sequential search model

In this paper, we study the effects of wage and price controls on employment, output, and welfare in a simplified version of the Bénabou (J. Econom. Theory 60 (1993) 140) equilibrium sequential search model with bilateral heterogeneity. We show that a price ceiling increases output but the change in welfare depends on three effects: the reduction in aggregate search costs, the increase in surplus due to increased output, and the transfer of production to the least efficient firm. The model is formally identical to a standard equilibrium search model of the labor market so analogous results hold for the minimum wage.     

A Walrasian Rubinstein and Wolinsky model

We provide a full dynamic analysis of a continuous-time variant of Rubinstein and Wolinsky (1985) matching and bargaining model with unbalanced flows of buyers and sellers. The focus is on the price limit as the frictions of search are removed. It is found that a necessary and sufficient condition for the limit price to be Walrasian at all times is the alignment of the initial buyer and seller stocks with the flows.     

Matching with noise and the acceptance curse

This paper explores matching with both search and information frictions. Specifically, everyone observes only a noisy signal of the true type of any potential mate. In this context, matching decisions must incorporate not only information about a partner's attribute conveyed by the noisy signal, but also—as in the winner's curse in auction theory—information about a partner's type contained in his or her acceptance decision.We show that there exists an equilibrium exhibiting a stochastic positive assorting of types, generalizing [Becker, J. Polit. Economy 81 (1973) 813-846]. In equilibrium, selection is adverse: being accepted reduces an agent's estimate of a potential partner's type, a phenomenon that we call the acceptance curse effect.     

Good news and bad news in two-armed bandits

This note provides a characterization of optimal policies in a class of two-armed bandits that allows for correlated arms. This characterization extends previously known results in the literature. In particular, it implies, under a mild continuity assumption, the optimality of cutoff belief strategies for this class of bandits. Cutoff beliefs are not unique, though.     

Optimal guessing: Choice in complex environments

In this paper, we extend the analysis of our earlier work on boundedly rational learning in an i.i.d. setting [Easley and Rustichini, Econometrica 67 (1999) 1157-1184] to complex decision problems. We show that the axioms from our earlier analysis can be applied in this more complex setting, and along with some new axioms, they asymptotically yield expected utility maximization. Perhaps most important is our demonstration of a simple procedure that insures expected payoff maximization no matter what Markov process the underlying process on states follows. We view this result as providing a positive learning result for all worlds in which learning is possible.     

Learning aspiration in repeated games

We study infinitely repeated symmetric 2×2 games played by bounded rational agents who follow a simple rule of thumb: each agent continues to play the same action if the current payoff exceeds the average of the past payoffs, and switches to the other action with a positive probability otherwise. By applying the stochastic approximation technique, we characterize the asymptotic outcomes for all 2×2 games. In the prisoners’ dilemma game, for example, the players cooperate in the limit if and only if the gain from defecting against cooperation is “modest.”     

Reconstructing dynamics from intertemporal economic data

Summary This paper is concerned with the relationship between a continuous dynamical system and the trajectory generated by such a system. The main result provides necessary and sufficient conditions for an infinite data stream to be rationalized as the output of a continuous law of motion. The paper develops concepts of informativeness of a given set of intertemporal data and shows that informativeness is maximal when the data is chaotic. It also demonstrates that with probability one the sample paths from a non-trivial independent and identically distributed stochastic process cannot be rationalized as the output of a continuous deterministic system. Two impossibility results are discussed which show that even with an infinite amount of data the hypothesis that the data has been generated by a non-monotonic function cannot be ruled out. An application concerning the recovery of the excess demand function from a sequence of price observations from the tatonnement process is also given.I would like to thank Professors Bent Christensen and Mukul Majumdar for valuable discussions and an anonymous referee for helpful comments. Financial support from SSHRC and Quebec's Fonds FCAR is gratefully acknowledged.     

The relative contributions of private information sharing and public information releases to information aggregation

We calculate learning rates when agents are informed through public and private observation of other agents' actions. We characterize the evolution of the distribution of posterior beliefs. If the private learning channel is present, convergence of the distribution of beliefs to the perfect-information limit is exponential at a rate equal to the sum of the mean arrival rate of public information and the mean rate at which individual agents are randomly matched with other agents. If, however, there is no private information sharing, then convergence is exponential at a rate strictly lower than the mean arrival rate of public information.     

Income distribution, efficiency and rationing

This paper studies the impacts of income distribution on the efficiency of trade mechanism and rationing of agents who are priced out of trade in a dynamic search model with two-sided asymmetric information. Buyers and sellers have asymmetric information about valuations and incomes respectively. In such a frictional environment, the effects of several variants of changes in income distribution on efficiency of trading mechanism and the population of rationed buyers and sellers are elaborated.     

Recursive preferences,learning and large deviations

We estimate the relative contribution of recursive preferences versus adaptive learning in accounting for the tail thickness of price–dividends/rents ratios. We find that both of these sources of volatility account for volatility in liquid (stocks) but not illiquid (housing) assets.     

Conditional preferences and updating

This paper axiomatizes updating rules for preferences that are not necessarily in the expected utility class. Two sets of results are presented. The first is the axiomatization of conditional preferences. The second consists of the axiomatization of three updating rules: the traditional Bayes rule, the Dempster-Shafer rule, and the generalized Bayes rule. The last rule can be regarded as the updating rule for the multi-prior expected utility (Gilboa and Schmeidler, J. Math. Econom. 18 (1989) 141). Operationally, it is equivalent to updating each prior by the traditional Bayes rule.     

From posteriors to priors via cycles

We present new necessary and sufficient conditions for checking if a set of players' posteriors may come from a common prior. A simple diagrammatic device calculates the join and meet of players' knowledge partitions. Each cycle in the diagram has a corresponding cycle equation. Posteriors are consistent with a common prior if and only if all cycle equations are satisfied. We prove that in games of two players, where the join partition has only singletons, a common prior exists if each player's distribution of beliefs over the elements of her opponent's partition is independent of her own private information.     

Inflation, price dispersion, and market structure

In this paper, we use a novel data set containing prices from bazaars, convenience stores, and supermarkets in Istanbul to re-examine the relationship between price dispersion and inflation. Although existing evidence is mixed, we find positive and significant relationships between dispersion, on the one hand, and lagged dispersion and unexpected product-specific inflation on the other. We also find evidence that dispersion is initially decreasing in anticipated aggregate inflation but is eventually increasing. Finally, average price duration and dispersion are lowest in the bazaar. This is intuitive, since menu and search costs should be minimal in that market structure.     

Dynamic price competition

We consider the model of price competition for a single buyer among many sellers in a dynamic environment. The surplus from each trade is allowed to depend on the path of previous purchases, and as a result, the model captures phenomena such as learning by doing and habit formation in consumption.We characterize Markovian equilibria for finite and infinite horizon versions of the model and show that the stationary infinite horizon version of the model possesses an efficient equilibrium where all the sellers receive an equilibrium payoff equal to their marginal contribution to the social welfare.     

Do Consumers Learn from Their Own Experiences?

It is natural to think that a household may learn from its own experiences and subsequently increase savings. This paper tests empirically the hypothesis that Japanese households learn from their experiences of large expenditure and increase their targets for precautionary savings after such experiences. The results imply that households raise their targets for precautionary savings by 4–5 percent of annual income in response to such experiences. Moreover, data are consistent with the argument that targets for savings affect actual savings. Assuming this holds, the results in this paper suggest that consumers may increase their actual savings following large expenditure.     

Search by committee

We consider the problem of sequential search when the decision to stop is made by a committee and show that a unique symmetric stationary equilibrium exists given a log concave distribution of rewards. We compare search by committee to the corresponding single-agent problem and show that committee members are less picky and more conservative than the single agent. We show how (i) increasing committee size holding the plurality fraction constant and (ii) increasing the plurality rule affect the equilibrium acceptance threshold and expected search duration. Finally, we show that unanimity is optimal if committee members are sufficiently patient.