David F. Hendry 11theory already existed, so complicated issues of model choice could be avoided
by imposing valid restrictions derived from correct economic theories: on discov-
ering that such theory was not available, many turned to help develop it (see,
e.g., Qin, 2008; Bjerkholt, 2007 (Bjerkholt, 2005, is a useful precursor). Koop-
mans, Hurwicz and Arrow all made major contributions to economic theory, and
to quote Bjerkholt (2007): “Haavelmo stated later on various occasions that eco-
nomic theory needed further development for the econometric methods to become
fully applicable” (also see Moene and Rødseth, 1991). Indeed, to quote Haavelmo
(1989) himself:
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.He reiterated that view in his presidential address to the Econometric Society
(published as Haavelmo, 1958):
What I believe to be true, however, is this: The training in the technical skills
of econometrics can represent a powerful tool for imaginative speculation upon
the basic phenomena of economic life; and, furthermore, it would be fruitful to
bring the requirements of an econometric “shape” of the models to bear upon
the formulation of fundamental economic hypotheses from the very beginning.Once model choice cannot be avoided, methodology becomes a salient issue,
and it would seem every conceivable methodology has at least one advocate. Pagan
(1987) considered what he viewed as the three main econometric methodologies,
relating mine to Leamer (1978) and Sims (1980), yet the ubiquitous “theory-based”
approach was not mentioned, albeit that there are really many variants thereof.
1.3.4 The decline of data concerns
“You’re travelling the wrong way.” (Train guard to Alice in Lewis Carroll,
1899)At about the same time thata prioritheory-based econometrics became dominant,
data measurement and quality issues were also relegated as a central component of
empirical publications. Early on, data series were often published in their entirety,
with careful caveats about accuracy, but later, at best, were recorded in appendices,
or not at all. For example, Clark (1932) allowed the first famous estimate of the
size of the Keynesian “multiplier.” Although computerized databases have recently
started to compensate for the absence of the printed record, electronic data are
sometimes revised in situ, making it difficult for later investigators to duplicate
previously-published findings. In a detailed study of a number of cases, Atkinson