Letter from James Hamilton, former chair of Department of Economics at UCSD

Two UCSD Professors Share Nobel Prize in Economics
Two economists who spent most of their careers at UCSD shared the 2003 Nobel Prize in Economics. Clive Granger has been on the UCSD faculty since 1974 and Robert Engle since 1975. Both retired from UCSD this summer and are now Emeritus Professors in the UCSD Economics Department. Engle currently spends most of his time at the Stern School of Business at New York University. The pair were recognized for their contributions in econometrics, which refers to statistical analysis of economic data.

It is hard to imagine what empirical research in economics would be like today in the absence of the contributions of Robert Engle and Clive Granger. Engle and Granger have been instrumental in developing economics as an empirical science, cultivating a tradition of describing what’s actually in the data. In the process, they have introduced and championed a series of methodological innovations that now define the standard of how one goes about analyzing and forecasting economic and financial time series.

Granger’s seminal contributions began with the publication in 1964 of Spectral Analysis of Economic Time Series (co-authored with M. Hatanaka). This monograph and Granger’s related early work were instrumental not only in establishing the methods and uses of spectral analysis of economic data, but also in championing the philosophy that one of the chief goals of econometrics should be to describe the properties of the data as faithfully but as simply as possible. Granger’s 1969 article in Econometrica, “Testing for Causality and Feedback,” was equally influential in forcing economists systematically to investigate and characterize the temporal ordering of dynamic relations.

Granger’s 1974 article with Paul Newbold, “Spurious Regressions in Econometrics,” introduced yet another revolution in econometric practice, alerting the profession to the need to model accurately the separate trends in individual economic time series before claiming to draw any inference about the affect of one variable on another.

Robert Engle’s introduction of the ARCH class of models in 1982 opened up a whole new field in econometrics devoted to predicting not only where a variable is headed (the conditional mean) but moreover how much it is likely to deviate from that target (the conditional variance). Such models form the empirical backbone of many of the computer calculations that drive world financial markets today, as analysts rely on them to quantify risk and calculate the value of a variety of financial options and derivatives.

Engle and Granger jointly introduced the concept of cointegration in 1987. This tool enabled economists to describe in a rigorous way the sense in which certain fundamental economic relations might be stable and highly predictable over time, even though our forecast of any individual component of the relation may deteriorate rapidly the farther in advance one tries to anticipate. This innovation prompted a phenomenal effort to characterize which economic relations satisfy this requirement of long-run stability, and has proven to be the single most successful econometric innovation of the last fifty years.

Engle and Granger have made a series of other contributions of lasting impact, including long-memory processes, combination of alternative forecasts, the Lagrange multiplier principle of model diagnostics, band-spectral regression, nonlinear modeling, and forecasting marked point processes, to mention but a few. Their influence on economic methodology has been profound, and their contribution well worth recognition by the international scientific community.