The Nobel Series: A complete econometrician
Trygve Haavelmo’s fundamental contribution was in looking at economic problems in terms of probability, which had a major impact on the development of the discipline of econometrics
The 1989 Alfred Nobel Memorial Prize in Economic Sciences was awarded to Professor Trygve Haavelmo. As the Nobel website tells us, the prize was awarded to him "for his clarification of the probability theory foundations of econometrics and his analyses of simultaneous economic structures".
Haavelmo did his undergraduate studies at the University of Oslo in 1930 and then was Ragnar Frisch's assistant at the Institute of Economics. In 1936, Haavelmo studied statistics at University College, London. After World War II broke out, Haavelmo left Norway and presented his doctoral dissertation, 'The Probability Approach in Econometrics,' at Harvard University in 1941. Although he had two doctorates from the University of Oslo, the work chosen by the Nobel committee was first published in 1944 in an American periodical 'Econometrica'. During the 1940s Haavelmo taught at the University of Chicago (where he was also a visiting professor in the late 1950s) before returning to Norway in 1947. He retired from the University of Oslo faculty in 1979, becoming professor emeritus.
Haavelmo's fundamental contribution was in looking at economic problems in terms of probability. Using the probability theory, conclusions could be drawn about relations among economic variables based on samples. Haavelmo's work had a major impact on the development of econometrics. As the Nobel website tells us:
Haavelmo's doctoral thesis had a swift and pathbreaking influence on the development of econometrics. His probability theory research programme attracted several outstanding economists – among them were subsequent Nobel laureates such as Koopmans and Klein. This gave rise to extraordinarily rapid methodological development, primarily during the 1940s. The foundation of modern econometric methods had thus been established.
In this article, we will review the main works of Haavelmo and how they have had an impact on public policy.
The most cited works of Haavelmo are perhaps 'The Probability Approach in Econometrics' published in 'Econometrica', 1944 and 'Multiplier Effects of a Balanced Budget', published in the same year. As mentioned above, Haavelmo was one of the earliest economists to use probability theory to draw statistical inferences about economic theory and testing such theories. Haavelmo believed that it was difficult for economists to 'fully' explain the many decisions taken by millions of economic agents. Decisions would have to encompass a 'stochastic' term (a term that has some probability of occurring and is not certain) which would explain the deviations from the predicted path. As the Nobel website tells us:
According to Haavelmo, it is unreasonable to believe that economists could ever "fully" explain or predict such individual decisions on the basis of necessarily simplified assumptions. Decisions are in fact affected by individual characteristics and numerous temporary conditions which change over time. Therefore, economists' explanations of decisions always have to encompass a stochastic term that summarizes these different kinds of "disturbances". As economic theories, in general, do not refer to individual decisions but are concerned with relations that comprise long sequences of decisions and a multitude of decision-makers, there are frequent opportunities to make relatively simple assumptions about the probability distribution of these aggregate relations.
Haavelmo also worked on the issue of the interdependence of various economic variables, and the problem this causes in specifying and estimating economic relations. To get over the problem, he underlined the importance of choosing those relations or variables that are 'autonomous' or those that are not affected by changes elsewhere in the system.
Haavelmo was also the first to give a mathematical formulation of what is known as the 'identification' problem in econometrics. The identification problem is basically a logical issue that must be solved before estimating an economic model. For example, in a demand and supply model, the equilibrium point belongs to both curves, and many possible curves can be drawn through such a point. For this model to be identified, parameters such as price, quantity demanded and supplied and the slopes of the demand and supply curves have to be specified. It may be noted that Haavelmo did not talk of 'model selection' in the modern sense of choosing one from many possible models, nor of alternative approaches to doing so. Haavelmo used the term 'testing' to include theory-model selection by checking the 'goodness' of the match of a theoretical model to the evidence, as well as a hypothesis and misspecification testing. Haavelmo (1944) was more concerned with measurement and inference methods which follow from the views econometricians hold about the nature of the world represented in their theories and in the data, and what are the right tools (designs of experiments) to uncover economic structures.
Haavelmo also pointed out the simultaneity problems that arise when variables are interdependent. This happens when the explanatory variable is determined at the same time as the explained variable. As the Nobel website tells us, Haavelmo had a fundamental contribution to this problem:
Using a probability theory framework, Haavelmo provided a generally valid formulation and method of measuring this bias in isolated estimates of individual relations in an interdependent system. He also showed that the problem could be avoided by using a method of simultaneous estimates of interdependent models.
Another major contribution of Haavelmo (1944) was formalizing economic forecasting, although based on the assumption that the right model was being used.
As noted above, Haavelmo's most well-known work is his doctoral dissertation entitled 'The Probability Approach in Econometrics' published in 1944 as a supplement to 'Econometrica'. This was extended in two articles in 'Econometrica'—'The Statistical Implications of a System of Simultane-ous Equations' (1943) and 'Statistical Analysis of the Demand for Food' (1947, co-authored by MA Girshick). Haavelmo also wrote a book titled, 'A Study in the Theory of Economic Evolution' (1954).
Haavelmo's work in economterics was in the tradition of Jan Tinbergen and Koopmans. As mentioned above, his contributions in identification and the simultaneity problem helped a generation of econometrics researchers in taking econometric modelling forward.
Towards the end of his career, Haavelmo was dissatisfied with the implementation of his 'Probability Approach in Econometrics' research pro-gramme. This was because the economic theory of the time was not keeping up with the sophistication of the econometric models suggested by him. He had stated as much in his presidential address to the Econometric Society, published as 'Haavelmo' (1958). To quote Haavelmo (1989):
The basis of econometrics, the economic theories that we had been led to believe in by our forefathers, were perhaps not good enough. It is quite obvious that if the theories we build to simulate actual economic life are not sufficiently realistic, that is, if the data we get to work on in practice are not produced the way that economic theories suggest, then it is rather meaningless to confront actual observations with relations that describe something else.
Be that as it may, Haavelmo's work of incorporating probability theory in econometric modelling and then using these models to test economic theory, is still widely accepted and continues to guide applications of econometrics.
The writer is an IAS officer, working as Principal Resident Commissioner, Government of West Bengal.
Views expressed are personal