Engle and Granger Receive Nobel Prize in Economic Sciences

On October 8, 2003, the Royal Swedish Academy researchers, but also for analysts on financial mar- of Sciences announced that the Bank of Sweden kets, who use them in asset pricing and in evalu- Prize in Economic Sciences in Memory of Alfred ating portfolio risk. Nobel, 2003, would be shared by ROBERT F. ENGLE “for Most macroeconomic time series follow a sto- methods of analyzing economic time series with chastic trend, so that a temporary disturbance in, time-varying volatility (ARCH)” and CLIVE W. J. say, GDP, has a long-lasting effect. These time se- GRANGER “for methods of analyzing economic time ries are called nonstationary; they differ from sta- series with common trends (cointegration).” The tionary series, which do not grow over time but fluc- two will share the prize amount of 10 million tuate around a given value. Clive Granger Swedish kroner (about US$1.3 million). demonstrated that the statistical methods used Researchers use data in the form of time series, for stationary time series could yield wholly mis- i.e., chronological sequences of observations, when leading results when applied to the analysis of estimating relationships and testing hypotheses nonstationary data. His significant discovery was from economic theory. Such time series show the that specific combinations of nonstationary time development of GDP [gross domestic product], series may exhibit stationarity, thereby allowing for prices, interest rates, stock prices, etc. During the correct statistical inference. Granger called this 1980s, the 2003 laureates devised new statistical phenomenon cointegration. He developed methods methods for dealing with two key properties of that have become invaluable in systems where many economic time series: time-varying volatility short-run dynamics are affected by large random and nonstationarity. disturbances and long-run dynamics are restricted On financial markets, random fluctuations over by economic equilibrium relationships. Examples time—volatility—are particularly significant be- include the relations between wealth and con- cause the value of shares, options, and other fi- sumption, exchange rates and price levels, and nancial instruments depends on their risk. Fluc- short- and long-term interest rates. tuations can vary considerably over time; turbulent Robert F. Engle was born in 1942 in Syracuse, periods with large fluctuations are followed by New York. He received his Ph.D. from Cornell Uni- calmer periods with small fluctuations. Despite versity in 1969 and is currently the Michael such time-varying volatility, in want of a better al- Armellino Professor of Management of Financial ternative, researchers used to work with statistical Services at New York University. methods that presuppose constant volatility. Robert Engle’s discovery was therefore a major break- Clive W. J. Granger was born in 1934 in Swansea, through. He found that the concept of autore- Wales. He received his Ph.D. from the University of gressive conditional heteroskedasticity (ARCH) ac- Nottingham in 1959 and is an emeritus professor curately captures the properties of many time of economics at the University of California, San series and developed methods for statistical mod- Diego. eling of time-varying volatility. His ARCH models —From news announcements of the Royal have become indispensable tools, not only for Swedish Academy of Sciences

JANUARY 2004 NOTICES OF THE AMS 43