Research Database

This paper surveys cooperative game theory when players have incomplete or asymmetric information, especially when the TU and NTU games are derived from economic models. First some results relating balanced games and markets are summarized, including theorems guaranteeing that the core is nonempty. Then the basic pure exchange economy is extended to include asymmetric information. The possibilities for such models to generate cooperative games are examined. Here the core is emphasized as a solution, and criteria are given for its nonemptiness. Finally, an alternative approach is explored based on Harsanyi’s formulation of games with incomplete information.

We study a one-sector growth model which is standard except for the presence of an externality in the production function. The set of competitive equilibria is large. It includes constant equilibria, sunspot equilibria, cyclical and chaotic equilibria, and equilibria with deterministic or stochastic regime switching. The efficient allocation is characterized by constant employment and a constant growth rate. We identify an income tax-subsidy schedule that supports the efficient allocation as the unique equilibrium outcome. That schedule has two properties: (i) it specifies the tax rate to be an increasing function of aggregate employment, and (ii) earnings are subsidized when aggregate employment is at its efficient level. The first feature eliminates inefficient, fluctuating equilibria, while the second induces agents to internalize the externality.

In this paper we develop a model in which a country faces a balance of payments crisis if constraints on its international borrowing bind. We use the model to describe the dynamics of the trade balance, capital account, and balance of payments of a country that borrows to finance consumption following sweeping macroeconomic and structural reforms and then hits constraints on its international borrowing. We compare the predictions of this theoretical example with events in Mexico from 1987 through 1995.

We highlight an example of considerable bias in officially published input-output data (factor-income shares) by an LDC (Turkey), which many researchers use without question. We make use of an intertemporal general equilibrium model of trade and production to evaluate the dynamic gains for Turkey from currently debated trade policy options and compare the predictions using conservatively adjusted, rather than official, data on factor shares. We show that the predicted welfare gains are not only of a different order of magnitude, but in some cases, of a different sign, hence, suggesting contradictory policy recommendations.

A traditional explanation for why sovereign governments repay debts is that they want to keep good reputations so they can easily borrow more. Bulow and Rogoff show that this argument is invalid under two conditions: (i) there is a single debt relationship, and (ii) regardless of their past actions, governments can earn the (possibly state-contingent) market rate of return by saving abroad. Bulow and Rogoff conjecture that, even under condition (ii), in more general reputation models with multiple relationships and spillover across them, reputation may support debt. This paper shows what is needed for this conjecture to be true.

Most empirical studies of the static CAPM assume that betas remain constant over time and that the return on the value-weighted portfolio of all stocks is a proxy for the return on aggregate wealth. The general consensus is that the static CAPM is unable to explain satisfactorily the cross-section of average returns on stocks. We assume that the CAPM holds in a conditional sense, i.e., betas and the market risk premium vary over time. We include the return on human capital when measuring the return on aggregate wealth. Our specification performs well in explaining the cross-section of average returns.

We provide a brief review of the techniques that are based on the Generalized Method of Moments (GMM) and used for evaluating capital asset pricing models. We first develop the CAPM and multi-beta models and discuss the classical two-stage regression method originally used to evaluate them. We then describe the pricing kernel representation of a generic asset pricing model; this representation facilitates use of the GMM in a natural way for evaluating the conditional and unconditional versions of most asset pricing models. We also discuss diagnostic methods that provide additional insights.

Recent literature suggests that historical accidents can trap economies in inefficient equilibria. In a prototype model in the literature, there are two locations, the productive South and the unproductive North. By accident of history, the industry starts in the North. Because of agglomeration economies, the industry may reside in the North forever—an inefficient outcome. This paper modifies the standard model by assuming there is a continuum of locations between the North and the South. Productivity gradually increases as one moves South. There is a unique long-run equilibrium in this economy where all agents locate at the most productive locations.

This paper examines the sets of feasible allocations in a large class of economic environments in which commitment is impossible (the standard definition of feasibility is adapted to take account of the lack of commitment). The environments feature either memory or money. Memory is defined as knowledge on the part of an agent of the full histories of all agents with whom he has had direct or indirect contact in the past. Money is defined as an object that does not enter preferences or production and is available in fixed supply. The main proposition proves that any allocation that is feasible in an environment with money is also feasible in the same environment with memory. Depending on the environment, the converse may or may not be true. Hence, from a technological point of view, money is equivalent to a primitive form of memory.

This paper surveys implementation theory when players have incomplete or asymmetric information, especially in economic environments. After the basic problem is introduced, the theory of implementation is summarized. Some coalitional considerations for implementation problems are discussed. For economies with asymmetric information, cooperative games based on incentive compatibility constraints or Bayesian incentive compatible mechanisms are derived and examined.

The Economics of QWERTY suggests that historical accidents can trap economies in inefficient equilibria. This paper suggests that such accidents do not have the force that proponents claim. The paper presents a mechanism that may unravel a locational advantage caused by an historical accident. In the model, there are agglomeration benefits from concentrating industry in a particular location because it enables a large variety of local suppliers to emerge. Firms differ by the extent to which they purchase from local suppliers. Low-tier firms purchase little; high-tier firms purchase more. When the industry migrates, the lowest-tier products move first.

We analyze a model in which there is socially inefficient competition among people. In this model, self-enforcing social norms can potentially control the inefficient competition. However, the inefficient behavior often cannot be suppressed in equilibrium among those with the lowest income due to the ineffectiveness of sanctions against those in the society with the least to lose. We demonstrate that in such cases, it may be possible for society to be divided into distinct classes, with inefficient behavior suppressed in the upper classes but not in the lower.

We examine whether the Mortensen-Pissarides matching model can account for the business cycle facts on employment, job creation, and job destruction. A novel feature of our analysis is its emphasis on the reduced-form implications of the matching model. Our main finding is that the model can account for the business cycle facts, but only if the average duration of a nonemployment spell is relatively high—about nine months or longer.

We provide new evidence that models of the monetary transmission mechanism should be consistent with at least the following facts. After a contractionary monetary policy shock, the aggregate price level responds very little, aggregate output falls, interest rates initially rise, real wages decline by a modest amount, and profits fall. We compare the ability of sticky price and limited participation models with frictionless labor markets to account for these facts. The key failing of the sticky price model lies in its counterfactual implications for profits. The limited participation model can account for all the above facts, but only if one is willing to assume a high labor supply elasticity (2 percent) and a high markup (40 percent). The shortcomings of both models reflect the absence of labor market frictions, such as wage contracts or factor hoarding, which dampen movements in the marginal cost of production after a monetary policy shock.

This paper provides new evidence that state policies play a role in the location of industry. The paper classifies a state as pro-business or anti-business depending upon whether or not the state has a right-to-work law. The paper finds that, on average, there is a large abrupt increase in manufacturing activity when crossing a state border from an anti-business state into a pro-business state.

This paper explores the extent to which the Mexican government's inability to roll over its debt during December 1994 and January 1995 can be modeled as a self-fulfilling debt crisis. In the model there is a crucial interval of debt for which the government, although it finds it optimal to repay old debt if it can sell new debt, finds it optimal to default if it cannot sell new debt. If government debt is in this interval, which we call the crisis zone, then we can construct equilibria in which a crisis can occur stochastically, depending on the realization of a sunspot variable. The size of this zone depends on the average length of maturity of government debt. Our analysis suggests that for a country, like Mexico, with a very short maturity structure of debt, the crisis zone is large and includes levels of debt as low as that in Mexico before the crisis.

In this paper we propose a general equilibrium model where heterogeneous agents specialize either in legitimate market activities or in criminal activities and majority rule determines the share of income redistributed and the expenditures devoted to the apprehension of criminals. We calibrate our model to the U.S. economy in 1990, and we conduct simulation exercises to evaluate the effectiveness of expenditures on police protection and income redistribution at reducing crime. We find that while expenditures on police protection reduce crime, it is possible for the crime rate to increase with redistribution. We also show that economies that adopt relatively more generous redistribution policies may have either higher or lower crime rates than economies with relatively less generous redistribution policies, depending on the characteristics of their wage distribution and on the efficiency of their apprehension technology.

Many unions in the United States have for several years engaged in what is known as pattern bargaining—a union determines a sequence for negotiations with firms within an industry where the agreement with the first firm becomes the take-it-or-leave-it offer by the union for all subsequent negotiations. In this paper, we show that pattern bargaining is preferred by a union to both simultaneous industrywide negotiations and sequential negotiations without a pattern. In recent years, unions have increasingly moved away from patterns that equalized wage rates across firms when these patterns did not equalize interfirm labor costs. Allowing for interfirm productivity differentials within an industry, we show that for small interfirm productivity differentials, the union most prefers a pattern in wages, but for a sufficiently wide differential, the union prefers a pattern in labor costs.

This paper considers model worlds in which there is a continuum of individuals who form finite-sized associations to undertake joint activities. We show how, through a suitable choice of commodity space, restrictions on the composition of feasible groups can be incorporated into the specification of the consumption and production sets of the economy. We also show that if there are a finite number of types, then the classical results from the competitive analysis of convex finite-agent economies can be reinterpreted to apply.

We consider the nature of optimal cyclical monetary policy in three different stochastic models with various shocks. The first is a pure liquidity effect model, the second is a cost of changing prices model, and the third is an optimal seinorage model. In each case we solve for the optimal monetary policy and describe how money growth and interest rates respond to shocks under the optimal policy. The shocks we consider are money demand shocks, productivity shocks, and government consumption shocks. All of the models have the feature that the Friedman rule of setting the nominal interest rate to zero is not optimal. Optimal policies are always time inconsistent even though lump sum taxation is allowed. At least in some instances we find that optimal policy dictates responses of money growth and interest rates which run counter to conventional wisdom.

In aggregate unadjusted data, measured Solow residuals exhibit large seasonal variations. Total Factor Productivity grows rapidly in the fourth quarter at an annual rate of 16 percent and regresses sharply in the first quarter at an annual rate of –24 percent. This paper considers two potential explanations for the measured seasonal variation in the Solow residual: labor hoarding and increasing returns to scale. Using a specification that allows for no exogenous seasonal variation in technology and a single seasonal demand shift in the fourth quarter, we ask the following question: How much of the total seasonal variation in the measured Solow residual can be explained by Christmas? The answer to this question is surprising. With increasing returns and time varying labor effort, Christmas is sufficient to explain the seasonal variation in the Solow residual, consumption, average productivity, and output in all four quarters. Our analysis of seasonally unadjusted data uncovers important roles for labor hoarding and increasing returns which are difficult to identify in adjusted data.

This paper develops a two-country model in which trade is central to the process by which technology diffuses from the innovating country (North) to the backward country (South). Innovation in North leads to the introduction of higher-quality equipment goods that South can import only after some resources have been spent to adapt those equipment goods to the local conditions of South. Barriers to trade and policies that increase the cost of adapting equipment goods to the local environment decrease the rate of technology adoption, leading to a lower steady state relative income level in South. The model is calibrated to quantify this negative impact of barriers to trade and technology adoption on relative income levels and explore some additional implications of the model.

A random matching model with money is used to study the nominal yield on small denomination, bearer, safe, discount securities issued by the government. There is always one steady state with matured securities circulating at par and, for some parameters, another with them circulating at a discount. In the former, a necessary and sufficient condition for a positive nominal yield on not-yet-matured securities is exogenous discriminatory treatment of them by the government. In the latter, the post-maturity discount on securities induces a deeper pre-maturity discount even without such discriminatory treatment.

Until the mid-19th century, shortages of currency were sometimes serious problems. One common response was to prohibit the export of coins. We use a random matching model with indivisible money to explain a shortage and to judge the desirability of a prohibition on the export of coins. The model, although extreme in many regards, represents better than earlier models a demand for outside money and the problems that arise when that money is indivisible. It can also rationalize a prohibition on the export of coins.

To illustrate the use of social accounting matrices (SAMs) in applied general equilibrium (GE) modeling, we use an aggregated SAM for the Spanish economy to calibrate a simple applied GE model. The idea is to construct artificial people—households, government, and a foreign sector—who make the same transactions in the equilibrium of the model economy as do their counterparts in the data. This calibration procedure can be augmented, or partially substituted for, by statistical estimation of key parameters. We show the usefulness of such a model by presenting the results of a comparative exercise that mimics the policy changes that took place in Spain during its 1986 integration into the European Community. Sub-sequent data shows the model results to be remarkably accurate, especially if we account for other major shocks affected the Spanish economy in 1986.

This paper surveys recently developed methods for Bayesian inference and their use in economic time series models. It begins by reviewing aspects of Bayesian inference essential to understanding the implications of the Bayesian paradigm for time series analysis. It next describes the use of posterior simulators to solve otherwise intractable analytical problems. The theory and the computational advances are brought together in setting forth a practical framework for decision-making and forecasting. These developments are illustrated in the context of the vector autoregressions, stochastic volatility models, and models of changing regimes.

We don’t really know why the U.S. banking industry is consolidating rapidly, as it has been doing for the last decade or so. After a large number of studies on the topic including this one, what seems clear is that consolidation is not producing significant efficiencies — at least not on average. Other dimensions of the consolidation trend — such as its heavy concentration in large banks — are just as poorly understood. Given this ignorance as to the “why” of consolidation, it is extremely risky to predict its future effects. Based on past experience and data, at least, we can conclude the following.

Constant returns to scale is a central construct of neoclassical theory. Previous studies argued that one must adopt a specification of the production function with substantial unobserved service variation to reconcile constant returns with the data. Some economists have argued that this finding has not resolved the size of returns to scale, since factor service variation is unobserved, and there is no generally accepted theory to guide specification of this alternative framework. In this paper we show that the stochastic version of the neoclassical growth model delivers an orthogonality condition which can be used to estimate returns to scale. Rather than the standard finding of increasing returns, we show that standard theory and conventional measures of output and inputs yield estimates of constant returns to scale at the aggregate level. Our estimates also suggest that factor service variation is not an important determinant of output fluctuations.

We produce a theoretical framework that helps explain the co-evolution of the real and financial sectors of an economy in the growth process, as described by Gurley and Shaw. According to them, self-financed capital investment first gives way to debt finance and later to the emergence of equity as an additional instrument for raising funds externally. As the economy develops further, the aggregate ratio of debt to equity will generally fall. We analyze that portion of their account concerning the evolution of equity markets. We show that in an important sense, debt equity are complementary sources for the financing of capital investments.

A current U.S. policy is to introduce a new style of currency that is harder to counterfeit, but not immediately to withdraw from circulation all of the old-style currency. This policy is analyzed in a random-matching model of money, and its potential to decrease counterfeiting in the long run is shown. For various parameters of the model, three types of equilibria are found to occur. In only one does counterfeiting continue at its initial high level. In the other two, both genuine and counterfeit old-style money go out of circulation—immediately in one and gradually in the other. There are objectives and expectations that can reasonably be imputed to policymakers, under which the policy that they have chosen can make sense.

Using an existing random matching model of money, I show that a once-for-all change in the quantity of money has short-run effects that are predominantly real and long-run effects that are in the direction of being predominantly nominal provided (i) the quantity of money is random and (ii) people learn about what happened to it only with a lag. The change in the quantity of money comes about through a random process of discovery that does not permit anyone to deduce the aggregate amount discovered when the change actually occurs.

We consider an environment in which risk-neutral firms must obtain external finance. They have access to two kinds of linear, stochastic investment opportunities. For one, return realizations are costlessly observed by all agents. For the other, return realizations are costlessly observed only by the investing firm; however, they can be (privately) observed by outsiders who bear a fixed verification cost. Thus, the second investment opportunity is subject to a standard costly state verification (CSV) problem of the type considered by Townsend (1979), Gale and Hellwig (1985), or Williamson (1986, 1987).

We examine the optimal allocations of investment between the two kinds of projects, as well as the optimal contract used to finance it. We show that the optimal contractual outcome can be supported by having firms issue appropriate (and determinate) quantities of debt and equity securities to outside investors.

The optimal debt-equity ratio necessarily depends (in part) on the firm’s asset structure. Investments in projects subject to CSV problems are associated (in a sense to be made precise) with the use of debt—as might be expected from the existing CSV literature. Investments in projects with publicly observable returns are associated with the use of external equity.

We examine in detail the relationship between the optimal asset and liability structure of the firm. We also describe conditions under which an increase in the cost of state verification shifts the composition of investment towards projects with observable returns, and reduces the optimal debt-equity ratio. Interestingly, the optimal debt-equity ratio is also shown to depend on factors that are irrelevant to asset allocations.

Finally, a large part of the interest in CSV environments has been due to the fact that they may result in equilibrium credit rationing. Our analysis has strong implications for the possibility of equilibrium credit rationing in more general CSV models.

In a finite-trader version of the Diamond-Dybvig (1983) model, the symmetric, ex-ante efficient allocation is implementable by a direct mechanism (i.e., each trader announces the type of his own ex-post preference) in which truthful revelation is the strictly dominant strategy for each trader. When the model is modified by formalizing the sequential-service constraint (cf. Wallace, 1988), the truth-telling equilibrium implements the symmetric, ex-ante efficient allocation with respect to iterated elimination of strictly dominated strategies.

This paper is motivated by a variety of empirical observations on the comovements of currency velocity, inflation, and the relative size of the “credit services” sector. By the credit services sector we mean the part of banking and credit sector which provides alternative means of transactions to using currency as well as other services which help people economize on currency. We incorporate the credit services sector into a monetary growth model. Our model makes two specific and new contributions. The first is to show that direct quantitative evidence on the welfare cost of low inflation using measures of the relative size of an appropriately defined credit services sector for the U.S.—essentially the cost incurred by banks and credit unions in providing demand deposit and credit card services—is consistent with the welfare cost measured using an estimated money demand curve following the classic analysis of Bailey (1956) and the more recent analysis of Lucas (1993). Both of these measures amount to about 0.5 percent of GNP. The second contribution is in providing welfare cost of inflation estimates over a range of inflation rates which have some new features. We find that the total welfare cost of inflation remains bounded at about 5 percent of consumption.

In this paper we analyze the rate of convergence to a balanced path in a class of endogenous growth models with physical and human capital. We show that such rate depends locally on the technological parameters of the model, but does not depend on those parameters related to preferences. This result stands in sharp contrast with that of the one-sector neoclassical growth model, where both preferences and technologies determine the speed of convergence to a steady-state growth path.

This article studies the extent to which governments produce goods for the market (that is, the extent of public enterprise production). It concludes that the current literature dramatically understates the role of public enterprises in many low-productivity countries. The current literature focuses on the total value of goods produced by public enterprises. This article focuses on the types of goods they produce. While the total value of goods produced by public enterprises (as a share of total output) differs a bit across countries, the types of goods they produce differ much more dramatically. In many low-productivity countries, the government produces a large share of the country's manufactured goods. In nearly all high-productivity countries, the government stays out of the manufacturing sector altogether. Therefore—and because the manufacturing sector plays a special role in economies—this article concludes that public enterprises play a very large role in many low-productivity countries.

This paper develops a Limited-Dependent Rational Expectations (LD-RE) model where the bounds can be fixed for an extended period, but are subject to occasional jumps. In this case, the behavior of the endogenous variable is affected by the agent's expectations about both the occurrence and the size of the jump. The RE solution for the one-sided and two-sided band are derived and shown to encompass the cases of perfectly predictable and stochastically varying bounds examined by earlier literature. We demonstrate that the solution for the one-sided band exists and is unique when the coefficient of the expectational variable is less than one. In the case of a two-sided band, the RE solution exists for all the parameter values and is unique if the coefficient of the expectational variable is less than or equal to one. These results hold even when the jump probability is stochastically varying and the error terms are conditionally heteroscedastic. As an illustration, we estimate a model of exchange rate determination in a target zone using data for the Franc/Mark exchange rate. Empirical results provide support for the non-linear model with time-varying realignment probability and indicate that the agents correctly anticipated most of the observed changes in the central parity.

This article investigates the business cycle implications of the planning phase of business investment projects. Time to plan is built into a Kydland-Prescott time-to-build model, which assumes that investment projects take four periods to complete. In the Kydland-Prescott time-to-build model, resources for these projects flow uniformly across the four periods; in the time-to-plan model, few resources are used in the first period. The investigation determines that incorporating time to plan in this way improves the model's ability to account for three key features of U.S. business cycles: their persistence, or the fact that when output growth is above (or below) average, it tends to remain high (or low) for a few quarters; the fact that productivity leads hours worked over the business cycle; and the fact that business investment in structures and business investment in equipment lag output over the cycle.

We provide a simple explanation for the observation that the variance of job destruction is greater than the variance of job creation: job creation is costlier at the margin than job destruction. As Caballero [2] has argued, asymmetric employment adjustment costs at the establishment level need not imply asymmetric volatility of aggregate job flows. We construct an equilibrium model in which (S,s)-type employment policies respond endogenously to aggregate shocks. The microeconomic asymmetries in the model can dampen the response of total job creation to an aggregate shock and cause it to be less volatile than total job destruction. This is so even though aggregate shocks are symmetrically distributed.

Using a calibrated growth model, the dynamic effects of NAFTA on Mexican development are studied. Two scenarios are analyzed. In the first, NAFTA is assumed to stimulate inflows of physical capital into Mexico. These inflows reduce the interest rate and raise the wage rates for both skilled and unskilled labor. The skilled wage rises more sharply, however, increasing the skill premium and rapidly accelerating the accumulation of human capital. In the second scenario, NAFTA is assumed to have the effect of fully integrating Mexico with the U.S. and Canada. Integration also reduces the interest rate and raises both wage rates in Mexico, but in this case the skill premium falls and human capital accumulation speeds up only a little. The welfare gains are large in both cases.

Financial planners typically advise people to shift investments away from stocks and toward bonds as they age. The planners commonly justify this advice in three ways. They argue that stocks are less risky over a young person’s long investment horizon, that stocks are often necessary for young people to meet large financial obligations (like college tuition for their children), and that younger people have more years of labor income ahead with which to recover from the potential losses associated with stock ownership. This article uses economic reasoning to evaluate these three different justifications. It finds that the first two arguments do not make economic sense. The last argument is valid—but only for people with labor income that is relatively uncorrelated with stock returns. If a person’s labor income is highly correlated with stock returns, then that investor is better off shifting investments toward stocks over time.

This note shows that basic theorems of dynamic programming hold when the return function is homogeneous of degree theta <= 1.

This study overturns the conclusion of a 1990 study by David Humphrey and Allen Berger, which found that check float is responsible for the popularity of checks despite their high resource cost compared to electronic payment instruments. The new study examines recent data on the costs of checks and automated clearinghouse (ACH) payments. It finds that the value of check float has decreased significantly since the 1990 study and is no longer large enough to make checks more attractive than ACH payments. The study also questions whether the idea that float could be responsible for the persistent use of checks is reasonable given standard assumptions about the behavior of economic agents. The study ends by speculating on why checks are used more than less-costly alternatives and by encouraging policymakers to wait for researchers to adequately answer that question before intervening in the market for payment instruments.

Using U.S. data it is shown that as the stock market goes into a period of high volatility, nondurables consumption is unaffected but durables consumption falls substantially. It is argued that a plausible explanation for this is that consumers face irreversibilities when adjusting their durables stock. They will thus apply Ss-type rules with bandwidths with widths that vary over time as in Hassler (1996). To quantify the aggregate implications of such behavior an aggregated irreversible investment model for consumer durables is estimated on U.S. It is found that a shift to higher risk leads to a simultaneous widening of individual Ss-bands which causes demand to fall substantially. This effect diminishes over time but is substantial also after a year.

A current U.S. policy is to introduce a new style of currency that is harder to counterfeit, but not immediately to withdraw from circulation all of the old-style currency. This policy is analyzed in a random matching model of money, and its potential to decrease counterfeiting in the long run is shown. For various parameters of the model, three types of equilibria are found to occur. In only one does counterfeiting continue at its initial high level. In the other two, both genuine and counterfeit old-style money go out of circulation—immediately in one and gradually in the other. There are objectives and expectations that can reasonably be imputed to policymakers, under which the policy that they have chosen can make sense.

This article describes a new way to use monthly data to improve the national forecasts of quarterly economic models. This new method combines the forecasts of a monthly model with those of a quarterly model using weights that maximize forecasting accuracy. While none of the method's steps is new, it is the first method to include all of them. It is also the first method to be shown to improve quarterly model forecasts in a statistically significant way. And it is the first systematic forecasting method to be shown, statistically, to forecast as well as the popular survey of major economic forecasters published in the Blue Chip Economic Indicators newsletter. The method was designed for use with the quarterly model maintained in the Research Department of the Minneapolis Federal Reserve Bank, but can be tailored to fit other models. The Minneapolis Fed model is a Bayesian-restricted vector autoregression model.

A version of the Diamond-Dybvig model of banking is used to evaluate the narrow banking proposal, the idea that banks should be required to back demand deposits entirely by safe short-term assets. It is shown that the mere existence of an amount of safe short-term assets outside the banking system that exceeds banking system liabilities does not make the proposal either innocuous or desirable. In fact, despite such existence, using narrow banking to cope with banking system illiquidity eliminates the role of the banking system.

This article documents a delay in the public release of Mexican international reserve data in the months before Mexico's debt crisis at the end of 1994. The article establishes that in that year investors did not know the level of Mexican reserves before October; yet this lack of information did not seem to reduce investor confidence in the Mexican economy. The article does not establish whether the delay in releasing reserve data was due to logistical problems or to a government strategy. The possibility that the delay was strategic is evaluated by developing an economic model that captures some of the principal constraints facing the Mexican government in 1994 and that makes explicit the conflicting objectives of the government and investors. The model shows that in such an environment with private information, strategic delay can occur in equilibrium if investors are uncertain about the cause of the delay.