The primary purpose of this paper is to compare the predictive accuracy of four models: (1) Sargent's classical macroeconometric model, (2) Sim's six-equation unconstrained vector autoregression model, (3) a "naive" eighth-order autoregressive model, and (4) my model. A recent method that I have proposed for estimating the predictive accuracy of a model, which takes account of the four main sources of uncertainty of a forecast, is used for the comparisons. The results indicate that Sargent's and Sims's models are the same as or less accurate than the naive model, depending on the variable, and that my model is more accurate for real GNP, the GNP deflator, and the unemployment rate and less accurate for the money supply and the wage rate than the naive model. A secondary purpose of the paper is to point out some econometric mistakes that Sargent made in his empirical work and to propose an alternative technique that can be used to estimate a rational expectations model like his.

In this paper a model is developed that explains the allocation of an individual's time among work and two types of leisure activities: time spent with spouse, and time spent with paramour. Data from two recent magazine surveys are available that can be used to test the predictions of the model regarding the determinants of time spent with paramour. The results of estimating the equation explaining time spent with paramour, by the Tobit estimator, are generally supportive of the model, although more evidence is needed before any definitive conclusions can be drawn. The model can also be applied to the allocation of time among other types of leisure activities.

The housing and mortgage markets have long been considered to be markets that may not always be in equilibrium, and many econometric models of the housing and mortgage markets have tried in one way or another to account for disequilibrium effects. In this paper a critique of previous models of the housing

[Was the Depression forecastable? After the Crash, how long should it have taken contemporary forecasters to realize how severe the downturn was going to be? Data assembled by the Harvard and Yale forecasters--together with modern historical data--are subjected to statistical analysis. Neither contemporary forecasters nor modern times-series analysts could have forecast the large declines in output following the Crash.]

This paper is concerned with the problem of estimating demand and supply schedules in disequilibrium markets. The results of Fair and Jaffee are expanded in three ways. (1) Their directional method I is modified to yield consistent estimates. (2) A maximum likelihood alternative to their quantitative method is proposed. (3) The price equation is generalized to be a multivariate, stochastic function, and a method is proposed for estimating demand and supply schedules in this case.

[This paper is concerned with the econometric problems associated with estimating supply and demand schedules in disequilibrium markets. The general problem is that in the absence of an equilibrium condition the ex ante demand and supply quantities cannot in general be equated to the observed quatity traded in the market. Four methods of estimation, differing primarily in their use of information on price-setting behavior, are developed in this paper. The first method is a generalization of an earlier meothd developed by R. Quandt and is based upon the maximization of a likelihood function. The method does not require any specific assumption about price-setting behavior, and it allows the sample separation (into demand and supply regimes) to be estimated along with the coefficient estimates. The second and third methods use the change in price as a qualitative proxy in determining the sample separation. The fouth method uses the change in price as a quantitative proxy for the amount of excess demand (supply) in the market. In the final section of the paper the four methods are used to estimate a a model of the housing and mortgage market in an effort to gauge the potential usefulness of each of the methods.]

The informational content of different forecasts can be compared by regressing the actual change in a variable to be forecasted on forecasts of the change. We use the procedure in Fair and Shiller (1987) to examine the informational content of three sets of ex ant. forecasts: the American Statistical Association and National Bureau of Economic Research Survey (ASA), Data Resources Incorporated (DRI), and Wharton Economic Forecasting Associates (UEFA). We compare these forecasts to each other and to quasi ex ante forecasts generated from a vector autoregressive model, an autoregressive components model, and a large-scale structural model (the Fair model).

This paper extends the classical test for structural change in linear regression models (see Chow (1960)) to a wide variety of nonlinear models, estimated by a variety of different procedures. Wald, Lagrange multiplier-like, and likelihood ratio-like test statistics are introduced. The results allow for heterogeneity and temporal dependence of the observations. In the process of developing the above tests, the paper also provides a compact presentation of general unifying results for estimation and testing in nonlinear parametric econometric models.

The information contained in one model's forecast compared to that in another can be assessed from a regression of actual values on predicted values from the two models. The authors do this for forecasts of real GNP growth rates for different pairs of models. The models include a structural model (the Fair model), various versions of the vector autoregressive model, and various versions of a model the authors call the "autoregressive components" model. The authors' procedure requires that forecasts make no use of future information and they have been careful to try to insure this, including using the version of the Fair model that existed in 1976, the beginning of their test period. Copyright 1990 by American Economic Association.