Models, Methods, and Applications of Econometrics: Essays in Honor of A.R. Bergstrom.

Caudill, Steven B.

A. R. Bergstrom is a prominent econometrician from New Zealand. His major contributions are the development of continuous-time econometric models and finite sample theory. Bergstrom's former student P. C. B. Phillips has assembled a collection of essays from an impressive group of econometricians to honor Bergstrom on the occasion of his sixty-fifth birthday. With the exception of an introductory chapter discussing Bergstrom's career, this volume is a collection of essays dealing with both theoretical and empirical aspects of continuous-time econometric models. The volume contains twenty-four papers organized into five groups. Part I discusses Bergstrom's education and his contributions to econometrics. Part II contains four papers on continuous-time models, and Part III contains five papers on finite-sample theory. Part IV contains seven papers on dynamic econometric models, and the final section, Part V, contains four empirical applications.

The first essay in Part I, written by Phillips, is a brief overview of Bergstrom's education and career. The second essay, entitled, "What is Econometrics?," is written by Bergstrom and was first published in 1966 in The University of Auckland Gazette. The third essay is a reproduction of Bergstrom's interview in Econometric Theory published in 1988. Part I concludes with a list of Bergstrom's publications.

Part II contains four papers on continuous-time models. The first paper, by Chambers, compares forecasts between continuous-time and discrete-time econometric models. The second paper, by Harvey and Stock, illustrates the application of continuous-time econometric modelling to the problem of interpolation (estimation of intermediate values of discretely sampled time series). The third paper, by Robinson, describes frequency-domain approaches to the estimation of parametric and semiparametric continuous-time econometric models. In the final paper, Wymer discusses the estimation of nonlinear continuous-time models.

Part III contains five papers on finite-sample theory. The first paper, written by Hillier and Skeels, presents some finite sample results for LIML, OLS, and TSLS estimators of the coefficients of exogenous variables in structural models. The second paper, by Phillips, presents a general mathematical framework for determining the distribution of any estimator or test statistic. The third essay, by Richmond, deals with multiple comparisons. The fourth paper, by Sargan, presents some alternatives to the Edgeworth approximation. A final paper by Ellison and Satchell is a Monte Carlo study of the properties of tests for the presence of cointegration.

Part IV contains seven papers on dynamic econometric modelling. The first paper, by Bowden and Martin, describes how a principal component decomposition of the spectral-density matrix can be used to identify peaks and troughs in the business cycle. The second paper, by Gregory, Pagan, and Smith, is an effort to synthesize two ideas. One idea, prominent in Bergstrom's work, is that the specification of the objective function faced by economizing agents determines the specification of the estimating equations. The other idea is that the estimation theory and procedures must be adapted for the data. The next paper, by Hannan, discusses parameter reduction in econometric models. The paper by Hansen obtains an efficiency bound for semiparametric estimators in time-series models. Next is a paper by Hendry and Mizon which discusses evaluating econometric models using the encompassing principle. A paper by Maheswaran and Sims uses continuous data to show that there are probability models for price behavior outside the semimartingale class. A final paper by Wickens examines rational-expectations models with integrated variables.

Part V contains four empirical applications. The first paper in the section is written by Agbeyegbe and examines the stochastic behavior of mineral-commodity prices. The second paper, by Gandolfo and Padoan, uses continuous-time models to examine the effects of capital liberalization. The third paper, by Giles and Wyatt, examines economies of scale in electricity distribution in New Zealand. The final paper in the section is written by Hall and Trevor. This paper estimates a continuous-time error-correction model using Australian household-level data on disposable income and aggregated and disaggregated consumer expenditures.

This volume is a serious effort in the area of econometrics, and is not light reading. Most of the essays are well written extensions or illustrations of Bergstrom's contributions to continuous-time econometric modelling or finite-sample theory. In several of the essays the reader is reminded that Bergstrom's contributions have been largely overlooked in North America. In fact, my own informal survey of the most popular econometrics books in my possession failed to yield a reference to Bergstrom. Consequently, I am sympathetic to the view that some have overlooked Bergstrom's important contributions. This oversight could be corrected by reading this volume or Bergstrom's book Continuous Time Econometric Models [1].

Reference

1. Bergstrom, A. Rex. Continuous Time Econometric Modelling. Oxford: Oxford University Press, 1990.

Steven B. Caudill Auburn University