Presentation Speech by Professor Bengt-Christer Ysander of the Royal Swedish Academy of Sciences, December 10, 1989
Translation from the Swedish text
Your Majesties, Your Royal Highnesses, Ladies and Gentlemen,
This year’s laureate, Professor Trygve Haavelmo of the University of Oslo, has been awarded the Alfred Nobel Memorial Prize in Economic Science for his pioneering contributions to the development of econometrics, i.e., methods used to estimate and test quantitative economic relations. Tangible instances of such econometrically estimated relations confront us all in our daily lives. Examples include estimates of the price sensitivity of demand and supply in different markets, the effects of reciprocal tariff exemption in the Common Market countries or forecasts of future macroeconomic fluctuations.
There are three requirements for any such estimate – or indeed for any empirical science. The first prerequisite is theories which reflect reality. The second consists of data or observations which conform to theoretical concepts. The third requirement is a methodology which can be used to quantify and test theoretical relations on the basis of empirical observations. Prior to the 1940s, economic research was impeded by a lack of systematic data and reliable testing methods. As a result, theories could often provide no more than qualitative conclusions and their empirical validity could seldom be tested in a satisfactory way.
Conditions for scientific work in economics underwent radical change during the 1940s. New economic policy initiatives necessitated the development of systematized economic data, particularly in the form of national accounting. The Prize was recently awarded to the English economist Richard Stone for his contributions in this field. It was also during this period that modern econometrics, based on probability theory, was established; economic theories and models could now be quantified and tested more consistently and systematically. And it was Trygve Haavelmo who initiated and provided the guidelines for this important methodological development in economics.
The results of earlier quantitative studies in economics had exposed two fundamental problems associated with the possibility of testing economic theories. First, theories regarding economic relations can never be expected to conform fully with available data. Underlying individual decisions are affected by numerous determinants which we are unable to take into account when formulating our economic relations.
Second, economists can seldom or never carry out controlled experiments in the same way as natural scientists. Available observations of market outcomes – prices, quantities, etc. – reflect the results of a multitude of different behaviorial relations with mutually interacting effects. Thus, an underlying relation – such as a demand function – can never be observed in isolation, but only as conditioned by a number of simultaneous relations and circumstances in the economy. This gives rise to what economists call “interdependence problems,” i.e., difficulties in using observations to identify, estimate and test underlying relations in an unambiguous way.
In his pathbreaking dissertation entitled “The Probability Approach in Econometrics ” and a number of subsequent studies, Trygve Haavelmo was able to show convincingly that these two fundamental problems could be solved if economic theories were formulated in probabilistic terms. Methods used in mathematical statistics could then be applied to derive stringent conclusions about underlying relations from the available random sample of empirical observations. Haavelmo also demonstrated that interdependence problems could be solved using statistical methods. One of his most noteworthy results concerns measurement of the bias which can arise in attempts to make isolated estimates of relations in large interdependent models. Haavelmo proposed a means of avoiding such bias through simultaneous estimation of the entire model structure.
Haavelmo’s research program quickly attracted a number of outstanding economists and statisticians, who gathered at the Cowles Commission in Chicago. This resulted in extraordinarily rapid methodological development, primarily during the 1940s. The foundation of modern econometric methods was thus established.
A testing method is not an end in itself, however. Realistic and testable theories are also required. A large portion of Haavelmo’s later research was devoted to the task of formulating testable dynamic theories. In particular, he made outstanding contributions to the theories of economic development and of investment.
Professor Haavelmo, you have established the foundations and guidelines of modern econometrics. You have shown that – despite difficulties in performing controlled experiments – empirical estimates and tests of economic theories can be derived by statistical methods. Quantitative studies of economic relations, used today as bases for private and public decision-making, are to large extent structured according to the methodological development you initiated.
It is an honor and a pleasure to congratulate you on behalf of the Royal Swedish Academy of Sciences and to ask you to receive, from the hand of His Majesty the King, the 1989 Alfred Nobel Memorial Prize in Economic Sciences.