Biology Reference
In-Depth Information
task of economic theory”: to establish those relations that might be expected to
possess as high a degree of autonomy as possible.
The problem is that it is not possible to identify the reason for the factual
influence of a factor, say x n +1 , being negligible, F n +1 Δ
x n +1
0. We cannot distin-
guish whether its potential influence is very small, F n +1
0, or whether the factual
variation of this factor over the period under consideration was too small,
x 1
Δ
0. We would like only to “throw away” factors whose influence was not observed
because their potential influence was negligible to start with. At the same time, we
want to retain factors whose influence was not observed because they varied so little
that their potential influence was veiled.
The variation of x n +1 is determined by other relationships within the system. In
some cases a virtually dormant factor may become active because of changes in the
economic structure elsewhere. However, deciding whether a factor should be
accounted for in the relationship under investigation should not depend on such
changes. The relationship should be autonomous with respect to structural changes
elsewhere.
How autonomous an equation is depends on our knowledge of the potential
influence of each factor, F i , which will inform us about the formal properties of the
function F (see Haavelmo 1944 , p. 24). Both Frisch and Haavelmo were pessimistic
about whether it was possible to acquire knowledge about the autonomy of an
equation through passive observation alone (therefore, they both advocated
experiments in economics). However, the problem is not insurmountable if we
use our knowledge about the time shapes of the phenomenon we want to explain.
Facts about the time shape of y can be fed back to the form F of the relation being
investigated. However, this only works if facts about the time shapes of the
phenomenon are invariant and stable, and not temporary characteristics. As we
discussed earlier, Haavelmo considered this strategy to be “very dangerous” and
therefore abandoned it. Frisch also, mistakenly, believed that the time shapes were
not sufficient to gain knowledge about the complete list of causal factors.
Unlike Haavelmo, Frisch and Tinbergen used (stylized) facts about the time
shape of the business-cycle phenomenon. For Tinbergen, whether integral terms
(hidden or not in the observations) should be included in the business-cycle
mechanism depended on assumptions about the periodicity and amplitude of the
business cycle. Frisch's Propagation and Impulse model ( 1933b ) also used time
shapes to gain knowledge about the business cycle's generating mechanism. How-
ever, his 1938 memorandum shows that the possibilities to identify the full list of
causal factors depend on the connection between assumptions about the mathemat-
ical representation of the business cycle and assumptions about the mathematical
representation of the explaining mechanism.
Haavelmo's design rules for doing econometrics were considered as an alterna-
tive to the experimental method of science (Morgan 1990 , p. 262). However,
although researchers at the Cowles Commission adopted Haavelmo's “blueprint”
for econometrics (Morgan 1990 , p. 251), they scrapped the term “autonomy”
because it was believed that structural relations were autonomous (see Aldrich
1989 ). Only Haavelmo kept the term “autonomy” alive. If one looks at the index of
Search WWH ::




Custom Search