Roger K. Conway

Applying Mixed Estimation in Econometric Research

The validity of the Theil-Goldberger mixed estimation procedure for incorporating prior information into the estimation of econometric models has recently been questioned in the literature. In this article, a set of accepted statistical principles are used both to rationalize past and future applications of mixed estimation through the introduction of a modified mixed estimator and to provide an explicit rationale for the abandonment of the Theil-Goldberger procedure. A previous application of mixed estimation in this Journal is reexamined to illustrate the issues involved.

Is the Phillips curve stable? A time-varying parameter approach

Abstract We derive two empirical Phillips curve models based on Robert Gordons reduced-form specification of conventional wage and price equations of a more complete structural model of the U.S. economy. One is a stochastic-coefficients model and the other is a conventional fixed-coefficients model. We used a stochastic-coefficients empirical model to investigate the volatility of the Phillips curve relationship hypothesized by many economists during the 1970s. The visual evidence of the time-varying parameter plots suggests there has been variation in the short-run Phillips curve. Comparative forecasting shows that the stochastic-coefficients model dominates the fixed-coefficients model, further supporting the hypothesis of volatility.

A forecast evaluation of capital investment in agriculture

Abstract A stochastic coefficients model developed by Swamy and Tinsley is used to forecast agricultural investment. In two sets of out-of-sample forecasts, one for 5 years, the other for 10 years, the Swamy-Tinsley stochastic coefficients model outperforms competing fixed and stochastic coefficients empirical models of agricultural investment for a wide array of risk functions. The Swamy-Tinsley stochastic coefficients investment model forecasts continued declines in net investment for farm machinery, with greater declines toward the end of the forecast period. The Swamy-Tinsley method produced better predictions than both stochastic and fixed-coefficients competitors.

Forecasting livestock prices: Fixed and stochastic coefficients estimation comparisons†

Agricultural commodity analysts have systematically overpredicted livestock prices during the 1980s by using econometric forecasting models that do not account for changing economic conditions. This article compares the out-of-sample forecast performance of the Swamy-Tinsley stochastic coefficients model with ordinary least squares, Cochrane-Orcutt, and maximum likelihood procedures that estimate red meat and chicken prices. The ability of a stochastic coefficients model to adapt quickly to changing economic conditions helps make it almost uniformly superior to a fixed coefficients model in forecasting the quarterly retail price for beef and chicken. The Cochrane-Orcutt and maximum-likelihood procedures appear to forecast pork prices better.

On the admissibility of restricted least squares estimators

This article examines the question of admissibility of the restricted least squares (RLS) estimator applied to the linear model , where . Straightforward proofs based on generalizations of the results of Cohen (1966) demonstrate that RLS estimators are admissible under predictive squared error risk iff the dimension of the β parameter space minus the number of linearly independent restrictions on β is less than three. A Stein-type estimator that dominates the inadmissible RLS estimator is presented. Some extensions of the results are indicated.

Estimating the Structure of Agricultural Investment: A Stochastic- Coefficients Approach

Two approaches, a flexibie-accelerator model and a stochastic-coefficients alternative, are used to estimate the structure of aggregate agricultural investment. Structural estimates of the adjustment rates for each model differ. The stochastic-coefficients model, however, performs better in out-of-sample forecasting.

Technology Policy and Agriculture: Discussion

The papers address two important forces affecting agricultural research and development: (a) how technology trade and innovation fit into an increasingly integrated world and agricultural economy and (b) the changing roles of public and private funding. Evenson adds an empirical foundation to the issue of technology flows among countries, a major contribution despite the difficulties of using patent data as an indicator of inventiveness. Three broad points of interpretation are worth emphasizing. First, considerable caution should be used in interpreting a decline in U.S. patenting in the 1980s as a decline in inventiveness. Poor protection of intellectual property rights abroad and increasing internationalization may have provided a growing incentive for firms to use trade secrets even though property rights were protected in the major developed countries considered by Evenson. Comanor and Scherer, for example, found a secular decline in the propensity to patent. Additionally, changing quality of patents, relative propensity of large versus small firms or of firms in different countries to