[Watts and Zimmerman [1978] tested their theory regarding management preferences over accounting standards using submissions on the General Price Level Discussion Memorandum. We first replicate the Watts-Zimmerman study using the same sample and then compare these results with identical tests based on a larger sample comprised of Exposure Draft submissions. Several biases are identified in the Watts-Zimmerman test procedures. When corrected, the theory's classificatory power deteriorates significantly on both the DM and ED samples. Alternative measurements of the predictor variables are introduced, as well as additional tests, in an attempt to improve the theory's performance. However, the theory's performance remains unimpressive.]

[The core of a game, which is an abstraction of the core or set of cooperative equilibrium states of an economy, is a fundamental notion of social equilibrium. However, except for games derived from special kinds of economic situations or satisfying restrictive (balancedness) conditions, the core is usually empty. In contrast, this paper shows that, with mild and economically natural assumptions, large games always have non-empty approximate cores. The game-theoretic framework is sufficiently general to cover a wide variety of economic situations.]

A NUMBER OF AUTHORS argue that a Bayesian posterior odds criterion is appropriate for model selection.2 This paper considers how to derive this criterion when there is minimal prior information. We propose minimizing measures of prior information relative to the models in question rather than relative to the parameters of the particular models. In so doing, we obtain an expression for the odds that is invariant to the parameterization of the particular models and overcomes certain well known finite sample limiting problems. We illustrate this procedure using two popular measures of information derived from the well known Shannon [26] measure. By minimizing these measures with the sample size held fixed, we obtain the same model selection criterion that Schwarz [25] derived asymptotically for large sample sizes. This expression has a number of desirable properties and is computationally no more

A latent-variable representation of health is used in a system that determines health and various indicators of health and health-care utilization for Nicaraguan women. Estimates differ somewhat from standard estimates. They imply that women's schooling increases health-care utilization, literate women report less disease incidence, women's childhood backgrounds affect their adult health and health-care utilization, and general resources favor childbirth-related care, but women's full income favors general care. The most significant result is that several important characteristics usually associated with development-women's schooling, household resources, and women's labor force participation-have inverse or no associations with women's health. Health status is hypothesized to be important in developing countries, both as a direct indicator of welfare and because of its possible impact on productivity. Health-care utilization also is of interest, both because of its relation to health status and because it is a particular manifestation of demand and supply for a 1 modem service. Work on the determinants of health status and health-care utilization in developing economies, however, is flawed because true health status is not directly observable. The indicators of health status that have been used in empirical studies-anthropometric measures, days ill, self-reported or clinical disease records, inputs such as nutrients, and health-care utilization indicesare imperfect indicators of underlying health status and generally do not give a congruent representation of that status. Thereby they may not measure adequately health status and their use may lead to biases in estimates and incorrect interpretations. We use a latent variable methodology for the first time to explore health status and health-care utilization in a developing country.' This approach employs systematic relations between various indicators of health status and health status itself and between health status and various related outcomes so that it is possible to estimate the determinants of health status and control for health status in other relations, even though health status is not directly observed. By using this methodology we obtain more satisfactory estimates of the determinants of health status and of health-care utilization in a developing country context than previous estimates using direct but imperfect proxies. This should improve the empirical bases for evaluating health and related policies in the developing world. I. Latent Variable Model of Health Status and Health-Care Utilization In the human capital tradition of analysis of health, individual health status is viewed as determined by individual demand factors given supply prices, environment, age and resources (including genes). Health-care utilization is a derived demand for a service which is used to produce better health. To incorporate the latent variable representation of the unobserved health status, we assume that the demand-supply health status and health-care utilization interactions, as well as the relation of other health indicators to health status, can be represented by the following linear system of relations:

Robert E. Baldwin, R. Spence Hilton, A Technique for Indicating Comparative Costs and Predicting Changes in Trade Ratios, The Review of Economics and Statistics, Vol. 66, No. 1 (Feb., 1984), pp. 105-110

tional attainments suggests the usefulness of special training programs for unskilled immigrants. Further, the persistence of racial differentials underscores the importance of the enforcement of anti-bias employment regulations to protect non-white immigrants. Recent changes in immigration law under the Refugee Act of 1980 will increase uncertainty regarding the composition of future immigration waves. It is likely, however, that the proportion of immigrants for whom specialized training and anti-bias regulation enforcement can hasten the traditional catch-up process which has historically characterized the immigrant economic experience will continually increase. Additional research using more recent data will allow both the efficacy of the recommendations advanced and the persistence of the observed patterns to be tested.