Showing papers in "The Review of Economic Studies in 1984"

Financial Intermediation and Delegated Monitoring

Abstract: This paper develops a theory of financial intermediation based on minimizing the cost of monitoring information which is useful for resolving incentive problems between borrowers and lenders. It presents a characterization of the costs of providing incentives for delegated monitoring by a financial intermediary. Diversification within an intermediary serves to reduce these costs, even in a risk neutral economy. The paper presents some more general analysis of the effect of diversification on resolving incentive problems. In the environment assumed in the model, debt contracts with costly bankruptcy are shown to be optimal. The analysis has implications for the portfolio structure and capital structure of intermediaries.

Information Reliability and a Theory of Financial Intermediation

Abstract: This paper is an analysis of when it will be beneficial for agents engaged in the production of information to form coalitions. The model is cast in a financial market framework, thus leading to an identification of conditions sufficient for the existence of financial intermediaries. Intermediation is shown to improve welfare if informational asymmetries are present, and the information generated to rectify these asymmetries is potentially unreliable. The usual appeal to transactions costs to explain intermediation is not needed.

Informative Advertising with Differentiated Products

Abstract: In this paper we study the role of promotional expenditures by sellers in a model of product differentiation. Advertising conveys full and accurate information about the characteristics of products. Heterogeneous consumers, who have no source of information other than advertisements, seek to purchase the products that best fit their needs. Despite the roles played by advertising in improving the matching of products and consumers, and in increasing the elasticity of demand faced by each firm, we find that the market-determined levels of advertising are excessive, given the extent of diversity in the market. We derive a promotional equilibrium based on a specific information transmission technology, paying explicit attention to the structure of consumer information and its impact on firms' demand curves. This allows us to study the effects of changes in the advertising technology. including an increased ability to target messages to specific groups of consumers, on the equilibrium in the product market. We find that decreased advertising costs may reduce profits by increasing the severity of price competition.

Optimal Incentive Schemes with Many Agents

Abstract: The Grossman-Hart principal-agent model of moral hazard is extended to the multiple agent case to explore the use of relative performance in optimal incentive contracting. Under the assumption that the principal chooses incentive schemes to implement agent actions as Nash equilibria, necessary and sufficient conditions are derived for the optimality of independent contracts, of rank-order tournaments, and for attainability of the first-best. In this context the relation of the principal's welfare to the correlation between the underlying randomness in outputs of different agents is also investigated. Finally, some problems with the Nash equilibrium implementation assumption are discussed. In this paper we extend the Grossman-Hart analysis of the principal agent problem to a multiple-agent setting. In this version of the problem, a risk-neutral principal is assumed to observe the output or a performance index for each (risk-averse) agent; the output of any agent depends on a privately observable action chosen by the agent and an exogenous random variable (the realization of which is not observed either by the principal, or by the agent before choosing its action). We especially focus attention on the role of relative performance evaluation in optimal incentive contracts, and the way this depends on properties of the underlying production technology and (the joint distribution of) the random variables affecting the outputs of different agents. In Section 2 of the paper we consider the principal's problem of finding an optimal incentive scheme and an action tuple for agents, subject to the constraint that the latter be implemented as a Nash equilibrium between agents by the chosen incentive scheme (where agents must attain at least their reservation utilities). A characterization of optimal incentive schemes is derived, analogous to the Grossman-Hart characterization for the single agent problem. This characterization is then utilized to find sufficient and (generically) necessary conditions for (a) optimal contracts for any agent to be independent of the performance of other agents, and (b) optimality of rank-order tournaments (where payments depend solely on ordinal comparisons of output across agents). We then describe a necessary and sufficient condition for the principal to extract maximum advantage from relative performance clauses, i.e. achieve the first-best level of expected utility in the multiagent situation. An implication of this result is that (under fairly weak conditions) perfect correlation in the underlying production uncertainties for different agents enables the principal to attain the first best if and only if sufficiently heavy penalties can be inflicted on agents. We then enquire what happens in intermediate cases of non-perfect correlation. It turns out that while in general the principal's welfare is a continuous function of the correlation coefficient, it is not always nondecreasing in the latter (because the correlation coefficient is not a general measure of association between two random variables, but rather a measure of linear association between them).

Spatial Equilibrium with Entry

Abstract: This paper examines spatial equilibrium in political competition when established parties choose their platforms competitively while rationally anticipating entry of a vote-maximizing third party. The resulting equilibrium is substantially different from the Hotelling "median" equilibrium. Established parties are spatially separated and third parties will generally lose the election. This provides one theoretical explanation for the stability of two-party systems. Namely that non-cooperative behavior between established parties can effectively prevent third parties from winning.

Pareto Inferior Trade

Abstract: The paper shows that between two competitive but risky economies with no insurance markets, free trade may be Pareto inferior to no trade. The model is simple enough to show clearly the role prices play in transferring and sharing risk when there is an incomplete set of markets, but rich enough to exhibit the resulting inefficiencies dramatically. The belief that free trade is Pareto optimal is one of the few tenets of economics which, at least until recently, would have received almost universal assent. The object of this paper is to demonstrate that this belief may not be well founded. We construct a simple model which lacks a complete set of risk markets but which in all other respects satisfies the conventional assumptions of a competitive economy, and show that free trade may be Pareto inferior to no trade. The basic idea behind our model is simple. There are two countries (regions) both of which grow a risky agricultural crop and a safe crop. The output in the two regions is perfectly negatively correlated. (The model can easily be extended to cases where the correlation is zero or even positive, so long as the correlation is not perfect.) In the absence of trade, price rises whenever outpurt falls. If demand functions have unitary price elasticity the price variations provide perfect income insurance for the farmer. With free trade the variations in the output of the risky crop offset each other and stabilize the price, which no longer varies to offset output variations. Consequently, the revenue from the risky crop now varies and the risk faced by the farmers is increased. This induces farmers to shift production away from the risky crop, raising its average price. Since consumers have unit price elasticity and thus spend a constant amount on both crops, the mean income of the farmers remains constant with the opening of trade while its riskiness increases. Consequently, farmers welfare necessarily decreases, as shown in Figure 1. Whereas before trade was opened, consumers bore all the risk, with free trade they bear none, and, other things being equal, this would make them better off. However, the increased riskiness of the risky crop induces farmers to shift their production to the safe crop, and the consequent rise in the average price of the risky crop can make consumers worse off. Near autarky, the risk benefit dominates this allocation effect, as shown in Figure 1, but near free trade the opposite is the case. If the change in supplies and prices is sufficiently large (which it will be if producers are sufficiently risk averse), and if the consumer risk benefits are sufficiently small (which they will be if consumers are not very risk averse), then consumers will be made worse off by opening trade. Since producers are necessarily worse off (in this model), it follows that free trade is Pareto inferior to autarky. The reconciliation of our results with the standard theorems of Welfare Economics in which free trade is Pareto efficient is straightforward-the conventional argument

Flexibility and Uncertainty

Abstract: The preserving of flexibility when faced with uncertainty is a neglected aspect of behaviour under risk. Yet it is an important factor in decisions to hold liquid assets or delay irreversible investment. This paper formalizes the notion of flexibility in a sequential decision context, and relates its value to the amount of information an agent expects to receive. A rudimentary money demand model is developed embodying these ideas, and the history of flexibility as an economic concept is traced. Choices are frequently made between alternatives that imply different degrees of future commitment-between a short-term investment that leaves future options open, for example, and a long-term one that, by its very nature, forecloses those options. The relative attractiveness of the two depends on their probability distributions of payoffs over time. A basic consideration in this choice is the recognition that beliefs about the risks governing these payoffs may change. Current doubts may be partially resolved in the near future. This prospect decreases the attractiveness of the longer term commitment, in that one is able to respond less fully to new information, and, even if it does not directly affect the risks associated with shorter term choices, enhances their appeal. This paper formalizes the above remarks by establishing connections between the following two (partial) orderings: The first is an ordering based on variability of beliefs. One set of beliefs is more variable than another if more final risk is resolved at an intermediate stage. The more one expects to learn by an intermediate period, relative to what one knows today, the more variation one is anticipating in beliefs about the final outcome. The second is an ordering of current actions, or positions, based on flexibility. One position is more flexible than another if it leaves available a larger set of future positions at any given level of cost. These two orderings are incorporated in a simple sequential decision model to suggest the following behavioural principle: The more variable are a decision-maker's beliefs, the more flexible is the position he will choose. This principle potentially applies whenever (i) there will be opportunities to act after further information is received, and (ii) current actions influence either the attractiveness or availability of different future actions.

The Timing of Sales

Abstract: This paper presents a model of intertemporal price discrimination. A fixed number of sellers produce a homogeneous good. Consumers with different preferences enter the market in each period and leave when they make a purchase. The sellers typically vary their prices over time, charging a high price in most periods, but occasionally cutting the price to sell to a large group of customers with a low reservation price. In some equilibria, all stores lower their price at the same time and to the same level.

The Identifiability of the Proportional Hazard Model

Abstract: This paper presents new identifiability conditions for the Cox proportional hazard model for duration data when unobserved person specific variables are present. We compare our conditions with those presented by Elbers and Ridder. We also present identifiability conditions for a rich class of parametric hazard models without regressor variables.

Uncertainty in the Theory of Renewable Resource Markets

Abstract: The natural growth rate of most renewable resource stocks is in part stochastic. This paper examines the implications of such ecological uncertainty for competitive equilibrium in a market with property rights. We show that stochastic fluctuations add a risk premium to the rate of return required to keep a unit of stock in situ, and we examine the effects of fluctuations on resource rent. Examples are used to show that extraction can increase, decrease, or be left unchanged as the variance of the fluctuations increases, depending on the extent of market "self-correction". Regulatory implications are also discussed. Renewable resource economics has traditionally been concerned with the study of dynamically optimal harvesting policies given a deterministic function for the natural growth of the resource stock. Issues have included the existence and characteristics of steady-state equilibria for the optimally managed resource, the need for and design of regulatory policies to prevent over-exploitation, and conditions under which (as a social optimum or otherwise) the resource will be exploited to extinction.1 Much of this work has been based on the assumption of a fixed and exogenous price for the harvested resource (typically resulting in "bang-bang" solutions for the harvesting policy). However some recent papers make price endogenous, and thereby describe how the extraction rate, and the rate of return and asset value of the resource behave in a competitive market with property rights.2

A Comparison of Posted Offer and Double Auction Pricing Institutions

Abstract: This paper presents an experimental study of a computerized "posted-offer" pricing mechanism that captures many of the basic institutional features of retail exchange in the U.S. Posted-offer market performance is evaluated relative to "double-auction" market performance using two supply and demand designs. Subject experience with the trading mechanism is explicitly considered as an experimental treatment variable. The market data suggest that prices tend to be higher and efficiency lower under posted-offer pricing relative to double auction. However, the institutional effect appears to interact with other design conditions. When feasible, the predictive power of competitive, Nash, and limit-price theoretic equilibria are empirically evaluated.

Profiles of fertility, labour supply and wages of married women: a complete life-cycle model.

Abstract: Of all the major economic and demographic events in the postwar United States period, the trends in fertility and female labour supply stand out as among the most dramatic. Fertility rates, after their Depression lows, rose unexpectedly after World War II far more than was anticipated as an adjustment to the War, and continued to rise until the late 1950's and early 1960's, when they again took an unexpected turn downward all the way to today's historical lows. Female labour-force-participation rates, continuing their historical climb, grew throughout the postwar period at an increasing pace, by 4 percentage points during the 1950's, by 5 percentage points during the 1960's, and by 8 percentage points during the 1970's. These two trends are of course closely related, for most married women have traditionally spent at least some minimal time out of the labour force while caring for their own children. Thus, for example, the slower growth rate of labour-force-participation rates in the 1950's is probably connected with the higher fertility rates in that period. In this paper a complete model of female labour supply and fertility choice is constructed and estimated. The model is more complete than previous models in several respects. First, labour supply and fertility decisions are modelled as completely joint in the same sense as the consumption of two goods is joint. Second, both are modelled as life-cycle decisions, that is, as decisions regarding sequences of labour-supply and fertility events rather than levels of the same. Hence the timing of both is important. Third, the life-cycle path of wage rates is made endogenous to the model-given a wage function showing the effect of work experience on future wages, the path of wages is determined automatically by the chosen path of labour supply, which in turn is related to fertility decisions that require some period of non-work. Econometrically, the three profiles for labour supply, fertility, and wages are estimated with a full-information maximum likelihood technique that solves several econometric problems, such as the problem of heterogeneity of tastes; the selectivity problem of missing wage rates for non-workers; and the problem of simultaneous-equations bias. The value of constructing a complete model is best seen by briefly reviewing past work in this area. The modern literature in the area is based upon the work of Becker

Bargaining with Incomplete Information: An Infinite-Horizon Model with Two-Sided Uncertainty

Abstract: The resolution of any bargaining conflict depends crucially on the relative urgency of the agents to reach agreement and the information each agent has about the others' preferences. This paper explores, within the context of an infinite-horizon bargaining model with two-sided uncertainty, how timing and information affect the rational behaviour of agents when commitment is not possible. Since the bargainers are uncertain about whether trade is desirable, they must communicate some of their private information before an agreement can be reached. This need for learning, due to incomplete information about preferences, results in bargaining inefficiencies: trade often occurs after costly delay. Thus, the model provides an explanation for the inefficient bargaining behaviour that appears to occur often in practice.

An Investigation of the Determinants of Manufacturing Employment in the United Kingdom

Abstract: In this paper we present and estimate an adjustment cost model of industry employment which takes explicit account of both expectations and aggregation over different labour types. The resulting model is subject to a large number of tests and is a highly robust representation of the data. Finally forecasts are produced for manufacturing employment up to 1990.

Prices, Product Qualities and Asymmetric Information: The Competitive Case

Abstract: Recent developments in the economics of information emphasize the informational content of prices. We examine the degree to which prices convey information on product quality to uninformed agents. Under perfect competition, we show that a rational expectations equilibrium may not exist. When an equilibrium does exist, the information on quality conveyed by prices depends on the shape of the average cost curves and the relative numbers of informed and uniformed agents.

Optimal Nonuniform Prices

Abstract: We consider optimal nonuniform pricing schedules, where the price depends upon the amount purchased. Such schedules are regularly used by public utilities and other services. Welfare-optimal nonuniform prices are related to the theory of optimal uniform prices developed by Ramsey. We characterize situations in which upward or downward discontinuities in pricing schedules are optimal. Our results are applicable to a number of related problems, including optimal taxation, insurance, and incentives. Economic theory has confined most of its attention to markets with uniform prices, in which the price per unit charged is invariant to the level of a customer's purchases. This focus is warranted when goods are resellable, since any attempt to charge nonuniform prices could be circumvented by reselling amongst customers. But when goods are not resellable, nonuniform prices are not only possible but common. Public utilities have declining block pricing, with lower unit charges for large levels of consumption. Many agencies providing services, including communications, transportation-even education-use two-part pricing, with a fixed entry or retainer charge, and a further charge per unit consumption. Nonuniform price schedules can be expected to be ever more common as services become a larger fraction of total output. The increasing prevalence of nonuniform pricing policies have made them the focus of recent public and economic attention. Oi (1971), Feldstein (1972), and Ng and Weisser (1974) have examined optimal two-part pricing policies. Leland and Meyer (1976) examine block pricing, and show by example that optimal welfare policies may require nonuniform prices if zero profit constraints are binding. In this paper, we consider arbitrary nonuniform pricing policies, using variational techniques. Two-part and block pricing schedules are, of course, special cases of the environment we consider. The work most closely related to our examination is the seminal paper by Spence (1978) and subsequent analysis by Roberts (1979). Spence, however, does not pay explicit attention to the key role played by constraints on optimal pricing schedules. Roberts considers upward price discontinuities which lead to bunching of consumers at a given level of consumption, but not downward price discontinuities, which lead to regions in which no consumer chooses to consume (gaps). Bunching or gaps may appear in the optimal price schedule even when underlying demand and cost functions are smooth and well behaved. The prevalence of discontinuous pricing schedules (e.g. the block pricing structures of public utilities) argues that these questions are of practical as well as theoretical interest.

Worker Allocation, Hierarchies and the Wage Distribution

Abstract: Economists have traditionally viewed the allocation of workers among jobs through the concept of comparative advantage. This paper investigates hierarchical production models which display an absence of comparative advantage, in order to demonstrate that for hierarchical production there is a second allocating factor which plays an important role. Two results are found. First, the allocation of workers among jobs tends to match high ability workers with positions which value ability highly. Second, despite the fact that the models display an absence of comparative advantage, i.e. the standard theoretical explanation for why wage distributions might be skewed, the models' wage distributions are skewed to the right relative to the underlying ability distributions.

Earnings, Unemployment, and the Allocation of Time Over Time

Abstract: Since an individual's labour market history can be described by a Markov Chain in state space, it is possible to calculate the steady states. It is natural to interpret the steady state proportion in unemployment for a group of identical workers as the natural rate of unemployment for this group, i.e. it is the expected proportion that would be employed if the environment were to remain constant. It is shown how the natural rate of unemployment is different for different groups. The model is estimated using data from a large U.S. income maintenance experiment. The predictions of the model, with some exceptions, bear up well to confrontation by data. The empirical findings also accord well with existing evidence on cross section studies of labour supply. We find that labour force participation rates implied by the estimated models are positively related to earnings for all age, sex and race groups, as expected. We demonstrate that the elasticities of the unemployment rates for all groups with respect to expected wage rates are negative, and large in absolute value in the case of adults. While the role of the wage in allocation of time is not novel, the existing neoclassical labour literature uses static labour supply models to interpret participation rate studies

Estimating Distributed Lags in Short Panels with an Application to the Specification of Depreciation Patterns and Capital Stock Constructs

Abstract: This paper considers the problem of estimating distributed lags in short panels. Though the N time series contained in a panel may allow for relatively precise estimates of identified lag coefficients, identification requires restrictions on the contribution of the unobserved pre-sample x's to the current values of y, and the shortness of panels focuses attention on this matter. We investigate two such restrictions. The first constrains the relationship between the presample and insample x's, while the second constrains the lag distribution itself. An example, which investigates empirically how to construct "capital stocks" for the analysis of rates of return, closes the paper.

The Theoretical Limits to Redistribution

Abstract: This is a revised version of the second Review of Economic Studies Lecture presented in April 1983 at the joint meeting of the Association of University Teachers of Economics and the Royal Economic Society held in Oxford. The choice of lecturer is made by a panel whose members are currently Professors Hahn, Mirrlees and Nobay, and the paper was refereed in the usual way. GEM

Aggregate Consumer Behaviour and the Measurement of Inequality

Abstract: This paper presents an approach to inequality measurement based on an econometric model of aggregate consumer behaviour. The novel feature of this model is that systems of individual demand functions can be recovered uniquely from the system of aggregate demand functions. We present methods for evaluating social welfare based on an explicit social welfare function. This social welfare function incorporates measures of individual welfare based on indirect utility functions for all consumer units. We develop indexes of inequality based on actual and potential levels of social welfare.

Probabilistic Social Choice Based on Simple Voting Comparisons

Abstract: A social choice procedure is developed for selecting an alternative from a finite set on the basis of paired-comparison voting. Ballot data are used to construct a lottery on the alternatives that is socially as preferred as every other lottery. The constructed lottery is then used to select a winner. An axiomatization of social preferences among lotteries that justifies the procedure is included. The procedure will always select a consensus majority alternative when one exists, and it will never select an alternative that is Pareto dominated by another alternative.

Product Differentiation with Imperfect Information

Abstract: The paper employs a model of monopolistic competition and product differentiation with consumers who are not well informed about the specification of the offered brands. Welfare analysis of the degree of product differentiation in such a market concludes that the socially desirable product variety is limited due to consumers' imperfect information. Consequently, when the number of consumers is sufficiently large or economies to scale in production are sufficiently weak, the market would offer excessive variety. The literature on product differentiation is motivated by the observation that only some of the potential brands of a differentiated product are produced and sold. It studies issues such as the role of the market structure in the determination of product diversity and the welfare performance of alternative markets in making this product selection. For example, see Lancaster (1979) and Salop (1979) in the spatial models tradition; Spence (1976) and Dixit and Stiglitz (1977) in the diversifying representative consumer tradition. These studies model product selection as an equilibrium outcome of monopolistic competition. The welfare analysis focuses on the tradeoff between consumer benefits which, naturally, increase with variety and production efficiency which, owing to economies to scale, decreases with variety. They find no natural relation between the socially optimal variety and the market outcome: the latter can be smaller or larger than the former, depending on the parameters of the model. Although it is natural to expect that, in some markets for differentiated products, consumers are imperfectly informed about the available brands, all these studies assume that consumers possess perfect information. The present paper assumes that consumers are imperfectly informed and indeed obtains qualitatively different results. First, regardless of how significant are the production economies of scale, the optimal variety is below a predetermined level that depends on the parameters of consumer utility and information. Second, when consumer population is sufficiently large or economies to scale in production are sufficiently weak, the market would offer excessive variety. The explanation is that consumers' imperfect information limits the extent to which they can exploit the variety offered in the market. It is, therefore, a social waste to increase variety to a level that consumers cannot benefit from.

A Price Discrimination Analysis of Monetary Policy

Abstract: Monetary policy is analysed within a model that appeals to legal restrictions on private intermediation to explain the coexistence of currency and interest-bearing default-free bonds. The interaction between such legal restrictions and monetary policy is illustrated in a version of the overlapping generations model. The model shows that legal restrictions and the use of both currency and bonds permit the government to levy a nonlinear inflation tax and that such a tax may be better in terms of the Pareto criterion than a linear inflation tax.

Comparative Dynamics of an Equilibrium Intertemporal Asset Pricing Model

Abstract: This paper uses recursive competitive theory to develop a general equilibrium asset pricing model. In this framework all prices and rates of return are endogenously determined, thus enabling us to analyze the effects of changes in preferences, technological uncertainty, and expectations on the structure of security prices. In particular we focus on how the market risk premium varies with changes in the underlying economic environment, an issue which other asset pricing models have chosen not to address.

Value of an Additional Firm in Monopolistic Competition

Abstract: The paper develops a model of monopolistic competition in which the satisfaction levels consumers get from products are independently and identically distributed. Potential substitution among products then generates demand curves through the mechanism of order statistics. The value of extra variety can be calculated directly. Pareto, lognormal and beta distributions are investigated using numerical integration. Some but not all cases support the argument that the optimal number of firms exceeds the equilibrium number.

Assets, General Equilibrium and the Neutrality of Money

Abstract: When government liabilities (including money) are held in private portfolios only as stores of value and do not provide additional services (such as liquidity), real variables are not affected by changes in the money supply due to the government's trading in real assets (open market operations). This neutrality of monetary policy fails if the government either trades in nominal assets, or it distributes subsidies and levies taxes. The ability of monetary policy to affect real variables is an important issue and has been the subject of numerous theoretical and empirical studies. The most useful theoretical framework within which to address the problem is a model of asset markets where the demand for money or other government liabilities is determined by rational portfolio choice (Tobin (1958)). These assets may, of course, provide additional services (such as transactions and liquidity). We assume here, however, that these services do not exist. The standard argument for the effectiveness of monetary policy goes as follows: Changes in the supply of money affect the price level and hence the distribution of returns to money as an asset. As individuals realign their portfolios and prices adjust to maintain market clearing, the equilibrium holdings of physical assets change. In this case money is not neutral. We shall argue, however, that, as long as changes in the supply of money are effected through the government's trading in real assets, this argument is, in general, not valid. It is well known that the effectiveness of monetary policy depends on the method of injection of new money into the economy. Two procedures are used by the government to effect changes in the supply of money: It either trades in assets, nominal or real (open market operations), or it distributes subsidies and levies taxes. We show here that open market operations in real assets (capital, or indexed bonds) have no effects on real variables in a general equilibrium model with market clearing. The argument, which can also be applied to the management of the public debt, is presented in a most general framework in the next section; it does not depend on the existence of a complete system of contingent securities or on an operative bequest motive. Furthermore, it is compatible with incomplete and differential information. The result is similar to the Modigliani-Miller theorem in corporate finance. That an argument along these lines can be employed to demonstrate the irrelevance of open

Uncertainty, Asymmetric Information and Bilateral Contracts

Abstract: On considere un monopole bilateral avec incertitude et information asymetrique, on donne les conditions necessaires et suffisantes a l'existence de contrats qui soient efficaces et a la fois stimulants

The Manipulability of Resource Allocation Mechanisms

Abstract: The vulnerability to manipulate behaviour of resource allocation mechanisms is evaluated by determining the Nash equilibria of associated manipulation games. Under manipulation, all monotonic correspondences are essentially equivalent to the Walrasian correspondence. For most non-monotonic correspondences of interest, the initial position appears at equilibrium to be efficient.

Mixed-Strategy Equilibrium in a Market with Asymmetric Information

Abstract: In the mid-1970s several authors studied models of markets with asymmetric information in which equilibria do not exist. Although those authors focused on models of insurance and education, it was recognized that similar nonexistence problems arise in a wide class of models with asymmetric information. Recently, Dasgupta and Maskin have demonstrated that for a game-theoretic version of at least one of those models, although no equilibria may exist in pure strategies, equilibria exist in mixed strategies. In the present paper we construct a mixed-strategy equilibrium for one member of the class—Spence's signalling model of education. Qualitative features of the equilibrium are explored.