The Keynes-Tinbergen debate
The first two macro-econometric models were constructed by Jan Tinbergen (19031994) (Jolink 2003; see Morgan 1990 for a discussion of both models). Tinbergen’s first model was of the Dutch economy, published in 1936 (Tinbergen 1959).
In the same year Tinbergen was commissioned by the League of Nations to perform statistical tests on business-cycle theories. The results of this later study were published in a two- volume work, Statistical Testing of Business- Cycle Theories (1939). The first contained an explanation of this new method of econometric testing as well as a demonstration of what could be achieved, based on three case studies. The second volume developed a model of the United States, the second macro-econometric model in the history of econometrics.Tinbergen’s first League of Nations study provoked a great deal of controversy. It was circulated in 1938 prior to its publication, and generated lively discussions about the role that econometrics could play in the testing of economic theory. It was John Maynard Keynes’s (1939) critique of Tinbergen’s first volume that sparked off the debate about the role of econometrics and what it might be able to achieve.
According to Keynes, the technique of multiple correlation analysis which had been adopted by Tinbergen was solely a method for measurement. It contributed nothing in terms of either discovery or criticism. The implication was that if the economic theorist does not provide the modeller with a complete set of causal factors, then the measurement of the other causal factors will be biased. Moreover, Keynes argued that some significant factors in any economy are not capable of measurement, or may be interdependent.
Another of Keynes’s concerns was the assumed linearity of the relations between these factors. He also noted that the determination of time-lags and trends was too often based on trial and error, and too little informed by theory.
Finally, there was the problem of invariance: would the relations found also hold for the future?Tinbergen’s response (1940) was technical, rather than methodological. He gave a very detailed description of how he had solved each problem by explaining the techniques he had employed. He assumed that through the adoption of these techniques he had overcome any of the methodological problem listed by Keynes.
Tinbergen’s reply was largely dismissed by Keynes. Taking into account all these concerns, Keynes came to the conclusion that econometrics was not yet a scientific approach:
No one could be more frank, more painstaking, more free from subjective bias or parti pris than Professor Tinbergen. There is no one, therefore, so far as human qualities go, whom it would be safer to trust with black magic. That there is anyone I would trust with it at the present stage or that this brand of statistical alchemy is ripe to become a branch of science, I am not yet persuaded. But Newton, Boyle and Locke all played with Alchemy. So let him continue. (Keynes 1940: 156)