The Regrettable Loss of Mathematical Molding in Econometrics - Mechanism and Causality in Biology and Economics - page 60

The works of the Jan Tinbergen, discussed in Sect. 2 , show how mathematics
was and could be used for identification purposes. The method Tinbergen employed
to arrive at a causal explanation of the business-cycle phenomenon started with
a priori economic-theoretical considerations about which explanatory variables
should be included. Some of the explanatory variables appeared as differential or
integral terms in the model equations. The equations were chosen to be linear, and
the values of the parameters were found by multiple regression analysis. Statistical
tests of significance were applied to measure the accuracy of these results. And,
moreover, the parameter values found for the causal relations were adjusted to
make sure that the model yields a cyclic movement with characteristics in accor-
dance with those of the actual business cycle. As a result of this latter assessment
(which will be called “tuning”), it appeared that integral terms were not of any
significance and therefore could be neglected. Differentials were approximated by
differences. So, after starting with using mixed differential-difference-integral
equations, Tinbergen ended up with representations of the business-cycle mecha-
nism with only difference equations.
In response to Tinbergen's reports on this method, Frisch ( 1995 ), discussed in
Sect. 3 , showed that the initial close relationship between the specific mathematical
representation of the business cycle and the mathematical representation of its
mechanism was lost in the transformation to difference equations. As a result, it
was no longer possible to identify all relevant causal factors. “Passive observation”
alone is not sufficient to detect them; statistics alone cannot reveal inactive but
potential factors. Without any feedback from the phenomena, we have to rely on
economic theory to provide us with a complete list of factors. A similar critique was
brought forward by John Maynard Keynes. Although unjustly addressed to “Pro-
fessor Tinbergen's Method,” it certainly applies to the later Cowles Commission
approach.
Am I right in thinking that the method of multiple correlation analysis essentially depends
on the economist having furnished, not merely a list of the significant causes, which is
correct so far as it goes, but a complete list? For example, suppose three factors are taken
into account, it is not enough that these should be in fact verœ causœ ; there must be no other
significant factor. If there is a further factor, not taken account of, then the method is not
able to discover the relative quantitative importance of the first three. If so, this means that
the method is only applicable where the economist is able to provide beforehand a correct
and indubitably complete analysis of the significant factors. The method is one neither of
discovery nor of criticism. It is a means of giving quantitative precision to what, in
qualitative terms, we know already as the result of a complete theoretical analysis. (Keynes
1939 , p. 560)
In other words, taking a strong apriorist position means that econometrics
becomes a method not of testing or of discovery, but of measurement.
Haavelmo ( 1944 ), discussed in Sect. 4 , discussed the problem of finding a
complete list of causal factors under the heading of the “problem of autonomy.”
However, the problem of autonomy was broader than this; it also covered
the problem of invariance. This latter issue concerns the identification of the
relationships between the causal factors that remain unaffected by changes