Knowledge that Transforms

The predictive content of the quantity-quality model of fertility and the empirical information required for verification under a minimal set of restrictions on the utility function is described. It is demonstrated that commodity-independent compensated price effects must be known to infer the existence of the unobservable interdependent shadow prices of the model with a relatively weak structure improsed on preference orderings. A method of using multiple birth events to substitute for these exogenous prices is proposed and applied to household data from India.

Previous equilibrium "business cycle" models are extended by the incorporation of an economy-wide capital market.One aspect of this extension is that the relative price that appears in commodity supply and demand functions becomes an anticipated real rate of return on earning assets, rather than a ratio of actual to expected prices.From the standpoint of expectation formation, the key aspect of the extended model is that observation of the economy-wide nominal interest rate conveys current global information to individuals.With respect to the effect of money supply shocks on output, the model yields results that are similar to those generated in simpler models.A new result concerns the behavior of the anticipated real rate of return on earning assets.Because this variable is the pertinent relative price for commodity supply and demand decisions, it turns out to be unambiguous that positive money surprises raise the anticipated real rate of return.In fact, this response provides the essential channel in this equilibrium model by which a money shock can raise the supply of commodities and thereby increase output.However, it is possible through a sort of "liquidity" effect that positive money surprises can depress the economy-wide nominal interest rate.

Given a preference basis for the model, a household's general equivalence scale is seen to be a cost of living index relative to the reference household type defined at constant prices. A maximum likelihood estimation procedure which can be applied both to cross-section data and pooled time-series/cross-sections is derived. The lack of identification of the model is established theoretically and checked empirically on British family expenditure survey data. With prior information, e.g. a nutrition based food scale, identification can be reached but ultimately the model is rendered implausible because of its zero substitution implication and the empirical results bear this out.

This paper presents necessary and sufficient conditions for the expected value of consumer surplus to correctly represent a consumer's preferences. A theorem characterizing utility functions which represent preferences over conditional probabilities is used to derive this. An application to price stabilization policy is presented.

Data sets which contain jointly endogenous discrete and continuous variables often occur in practice. This paper presents a model of the economic and stochastic processes generating such data as well as methods of estimation. A maximum likelihood estimator is examined and found to exhibit the usual optimality but it is computationally burdensome. A simpler estimator, the QREG, which is a simple weighted average of separate probit (or logit) and regression estimates is suggested as an attractive alternative. The QREG is also found to be optimal but only when a certain covariance restriction is found to hold. Thus a test of the restriction based on the joint distribution of separate probit and regression estimates is proposed.

[A number of estimators of parameters in nonlinear models have been proposed in the econometric literature. Various specialized methods have been developed to demonstrate the consistency of the suggested estimators. The first part of this paper presents a general scheme of the consistency proof. This method can be used to prove consistency of the large class of estimators of parameters in nonlinear regression models and nonlinear simultaneous equation models. The second part of this paper utilizes this general method to demonstrate the consistency of maximum likelihood estimators of nonlinear regression models with autocorrelated errors.]

FLWNO