John R. Conlon

Two New Conditions Supporting the First-Order Approach to Multisignal Principal-Agent Problems

This paper presents simple new multisignal generalizations of the two classic methods used to justify the first-order approach to moral hazard principal-agent problems, and compares these two approaches with each other. The paper first discusses limitations of previous generalizations. Then a state-space formulation is used to obtain a new multisignal generalization of the Jewitt (1988) conditions. Next, using the Mirrlees formulation, new multisignal generalizations of the convexity of the distribution function condition (CDFC) approach of Rogerson (1985) and Sinclair-Desgagne (1994) are obtained. Vector calculus methods are used to derive easy-to-check local conditions for our generalization of the CDFC. Finally, we argue that the Jewitt conditions may generalize more flexibly than the CDFC to the multisignal case. This is because, with many signals, the principal can become very well informed about the agents action and, even in the one-signal case, the CDFC must fail when the signal becomes very accurate. Copyright 2009 The Econometric Society.

The economics of natural language interfaces: natural language processing technology as a scarce resource

This paper discusses appropriate application areas for natural language interfaces (NLIs) to databases. This requires comparing NLIs with competing approaches, including other user-friendly interfaces, and training of users with less user-friendly interfaces. Also, since NLI technology is still limited, users may need to learn how to use NLIs themselves. This suggests that NLI popularity may snowball at some point, as users become familiar with NLIs. We use a simple prototype NLI to illustrate when NLIs can achieve flexibility unattainable by simpler interfaces. Currently existing commercial NLIs and application-specific customization are also discussed.

Should Central Banks Burst Bubbles? Some Microeconomic Issues

Anti‐bubble policy is examined in a finite‐horizon ‘greater fool’ bubble model, with rational agents, asymmetric information and short‐sales constraints. This permits the use of standard tools of welfare economics to analyse bubble policies. Policy is modelled as deflating overpriced assets by revealing information about this overpricing. If the central bank is following such a policy, then the market interprets inaction as an implicit endorsement of asset prices, which raises these prices. Also, the central bank can deflate overpriced assets even if it has no informational advantage over any investors about this overpricing. However, unless it has such an informational advantage, a bubble‐bursting rule may only make things worse.

Policy Reform and the Free Rider Problem

We investigate policy reform in a modelwith both lobbying, which involves afree-rider problem, and ordinary rentseeking, which does not. These activitiesinvolve similar skills, so a reform whichreduces rents shifts labor into lobbying.Also, because of the free-rider problem,the marginal return to the industry fromlobbying may greatly exceed an individualfirms return to lobbying. Thus, the shiftinto lobbying caused by rent reduction maylead to large increases in transfers to thelobbying industry. Under somecircumstances, a reform which reducesavailable rents increases total rents plustransfers to the industry.

Hope springs eternal: learning and the stability of cooperation in short horizon repeated games

Abstract This paper considers learning rates in finitely repeated prisoners’ dilemmas. If players think their opponents might be relatively cooperative (e.g., tit-for-tat or grim types), they will cooperate in finitely repeated prisoners’ dilemmas (see Kreps et al., J. Econom. Theory 27 (1982) 245). However, if there are actually no cooperative types, players will eventually learn this and cooperation will break down. This paper shows that this learning is extremely slow, so it will take an extremely long time for cooperation to break down. Thus, suppose the world is either “good” or “bad.” The probability of a grim type is δ>0 if the world is good, and zero if the world is bad. Successive generations pair up to play finitely repeated prisoners’ dilemmas. Players observe play in previous generations and use Bayes’ rule to update their prior, π, that the world is good. We show that, if the world is really bad, then π falls O(δ 2 log (1/δ)) per generation on average. Thus, if δ is small, there is less cooperation if the world is good, but cooperation may become more stable. For a representative 19 period repeated prisoners’ dilemma, beliefs fall one percentage point on average after a thousand generations. To derive these learning rates, we must refine existing results on the sensitivity of repeated games to Kreps et al. (1982) type perturbations. Specifically, we show cooperation is possible in perturbed prisoners’ dilemmas repeated O(log(1/δ)) times. This improves significantly on the O(1/δ) results in previous work. The paper thus provides two new reasons why cooperation tends to be stable, even in short horizon repeated games.

Optimal use of an information retrieval system

This article examines a simple model of an information retrieval (IR) technology, and uses this model to examine the optimal use of such a technology. In particular, we show that, if we understand optimal behavior, then we can (a) use observed behavior to shed light on the users objective function, and (b) analyze how a users behavior changes as her environment changes, for various specifications of the users objective function. We also show how user satisfaction varies with the underlying stochastic behavior of the retrieval process.

Profit, Supply, and Factor Demand Functions: Comment

In Lau and Yotopoulos (1972) a new method of estimating production relations was developed, building on Lau and Yotopoulos (1971). This method, bringing together estimates of profit and factor demand functions, was further developed in Yotopoulos and Lau to incorporate a test of rationality. Their method has since been applied by Yotopoulos, Lau, and Lin; Sidhu; Sidhu and Baanante (1979, 1981); Jamison and Lau; and Junankar (1980a, b, 1982). Because the assumption of rationality is so fundamental to economics, and because few tests of this hypothesis exist, the Lau-Yotopoulos method is very important. However, these studies rest on advanced economic theory and sophisticated econometrics, so it is easy to be intimidated by them. Sophisticated estimators, however, are not necessarily free from er-

Are People Sometimes Too Honest? Increasing, Decreasing, and Negative Returns to Honesty

We show that sender honesty can hurt receivers in simple signaling games. The receiver faces a trade-off between its ability to work with senders and the quality of information it can get and use from them. Our example also contradicts recent work suggesting that returns to honesty should be increasing. Positive, increasing returns are restored in our model if the receiver can precommit.

Robustness of well-designed retrieval performance measures under optimal user behavior

This article considers a simple model of an information retrieval (IR) technology, and uses this model to examine performance measures for IR systems. In particular, we show that, if a user is employing an IR system optimally, then reasonable performance measures will tend to be relatively robust, or insensitive, to small variations in assumptions about the users behavior. A simple empirical illustration of this approach is reported, though more work will be needed to determine exactly how robust performance measures will be in practical applications.

Separability and separability flexibility of the CRESH function

The CRESH functional form has many desirable properties. It is globally regular for a wide range of parameter values, it includes the CES as a special case, and it has stable patterns of relative substitutability. However, while the CRESH form is defined to be implicitly additively separable, theordinary separability structure of the CRESH form has never been investigated. In this article, that deficiency is corrected by describing the separability structure of CRESH functions. In particular, it is shown that the CRESH form is separability flexible in the important three input case. This and its other properties may make the CRESH function the form of choice among existing three input production functions.