In this paper, the authors reexamine the data from B. O'Neill's (1987) experiment involving a repeated, two-person, constant-sum game. They find that there is less evidence in support of the minimax hypothesis than indicated by O'Neill. There is strong evidence of serial correlation in players' choices, with several players displaying statistically significant dependence on the past moves of their opponents. The authors interpret this finding as evidence that the plays themselves rejected minimax play as the appropriate model for their opponents' behavior. They find no evidence that players' behavior approached minimax behavior as players became more experienced. Copyright 1990 by The Econometric Society.

This paper describes a new approach to normative economics, combining the theory of social choice with econometric modeling of aggregate consumer behavior. The author first derives a system of aggregate demand functions by exact aggregation over individual demand functions. He then constructs measures of individual welfare from systems of individual demand functions. Finally, the author incorporates these measures into a social welfare function, introducing ethical assumptions based on horizontal and vertical equity. To illustrate the application of this approach, he considers the U.S. standard of living and its cost over the period 1947-85. Copyright 1990 by The Econometric Society.

Individuals with finite private information independently choose acts and messages. Their utilities may depend on all acts and information, including the center's. Incentive payments are separable and fully transferable. Implementable incentives making specified behavior a Bayesian equilibrium are derived whenever the center's information depends stochastically, however slightly, on all relevant private information, and also whenever individuals' relative valuations of acts, however divergent, are not too dissimilarly affected by different states of nature. Feasibility is resolved whenever the desired strategies reveal the agents' beliefs about the center's information. Key concepts of agent similarity are developed for nonresponsive and budget-balancing cases. Copyright 1990 by The Econometric Society.

This paper considers estimation and hypothesis testing in linear time series when some or all of the variables have (possibly multiple) unit roots. The motivating example is a vector autoregression with some unit roots in the companion matrix, which might include polynomials in time as regressors. Parameters that can be written as coefficients on mean zero, nonintegrated regressors have jointly normal asymptotic distribution, converging at the rate of T(superscript "one-half") In general, the other coefficients (including the coefficient on polynomials in time), and associated t and F test statistics, have nonstandard asymptotic distributions. Copyright 1990 by The Econometric Society.