David F. Hendry 5main aim of an introductory chapter is often to overview the contents of the vol-
ume, but that is manifestly impossible for theHandbook of Econometricsgiven its
wide and deep coverage. In any case, since theHandbookis itself an attempt to
overview Applied Econometrics, such an introduction would be redundant.
Thus, my focus on empirical econometric modeling concerns only one of the
activities, but I will also try to present an interpretation of what “Applied Econo-
metrics” is; what those who apply econometrics may be trying to achieve, and how
they are doing so; what the key problems confronting such applications are; and
how we might hope to resolve at least some of them. Obviously, each aspect is con-
ditional on the previous one: those aiming to calibrate a theory model on a claimed
set of “stylized facts” are aiming for very different objectives from those doing
data modeling, so how they do so, and what their problems are, naturally differ
greatly. This chapter will neither offer a comprehensive coverage, nor will it be an
uncontroversial survey. En route, I will consider why “Applied Econometrics” does
not have the highest credibility within economics, and why its results are often
attacked, as in Summers (1991) among many others (see Juselius, 1993, for a reply).
Evidence from the contents of textbooks revealing the marginal role of “Applied
Econometrics” and “economic statistics” within the discipline has been provided
recently by Qin (2008) and Atkinson (2008) respectively. Since two aspects of our
profession with even lower status than “Applied Econometrics” are data (measure-
ment, collection and preparation), and teaching, I will try and address these as well,
as they are clearly crucial to sustaining and advancing a viable “Applied Economet-
rics” community. Economic forecasting and policy are not addressed explicitly,
being uses of empirical models, and because the criteria for building and selecting
such models differ considerably from those applicable to “modeling for under-
standing” (see, e.g., Hendry and Mizon, 2000; and for complete volumes on fore-
casting, see Clements and Hendry, 2002a, 2005; Elliott, Granger and Timmermann,
2006).
Economists have long been concerned with the status of estimated empirical
models. How a model is formulated, estimated, selected and evaluated all affect
that status, as do data quality and the relation of the empirical model to the ini-
tial subject-matter theory. All aspects have been challenged, with many views still
extant. And even how to judge that status is itself debated. But current challenges
are different from past ones – partly because some of the latter have been success-
fully rebutted. All empirical approaches face serious problems, yet the story is one
of enormous progress across uncharted terrain with many mountains climbed –
but many more to surmount. I will recount some of that story, describe roughly
where we are presently located, and peer dimly into the future. Why “Applied Econo-
metrics Through the Looking-Glass”? Lewis Carroll was the pseudonym for Charles
Dodgson, a mathematician who embodied many insights in the book which is
cited throughout the present chapter: a Looking-Glass is a mirror, and applied
findings in economics can only reflect the underlying reality, so obtaining a robust
and reliable reflection should guide its endeavors.
Following the brief section 1.2 on the meaning of the topic, section 1.3 sum-
marizes some of the history of our fallible discipline. Then section 1.4 proposes a