Home > Uncategorized > How to teach econometrics

How to teach econometrics

from Lars Syll

aWhen-I-tell-people-I-study-econometrics-1280x721Professor Swann (2019) seems implicitly to be endorsing the traditional theorem/proof style for teaching econometrics but with a few more theorems to be memorized. This style of teaching prepares students to join the monks in Asymptopia, a small pristine mountain village, where the monks read the tomes, worship the god of Consistency, and pray all day for the coming of the Revelation, when the estimates with an infinite sample will be revealed. Dirty limited real data sets with unknown properties are not allowed in Asymptopia, only hypothetical data with known properties. Not far away in the mountains is the village of Euphoria where celibate priests compose essays regarding human sexuality. Down on the plains is the very large city of Real Data, where applied economists torture dirty data until the data confess, providing the right signs and big t-values. Although Real Data is infinitely far from Asymptopia, these applied econometricians are fond of supporting the “Scientific” character of their work with quotations from the spiritual essays of the Monks of Asymptopia.

Ed Leamer

  1. Frank Salter
    January 12, 2020 at 11:20 am

    The most important methodology which is essential in understanding the limits of econometrics is the quality calculus(de Boer). Without it econometrics has no understanding of its limitations and therefore over-claims massively.

    J. de Boer, “On the History of Quantity Calculus and the International System”, Metrologia vol. 31, no. 6, pp. 405−429.

  2. ghholtham
    January 12, 2020 at 6:12 pm

    Ed Leamer satirises the misuse of econometrics and in particular the practice of data mining by repeated specification searches. It is indeed all too prevalent. There is a bias in academic journals for “positive” results, which no doubt distorts practice.
    What Leamer does not mention is that econometrics has taken the analysis of small, dirty data sets further than any other discipline. Because physical sciences can usually generate their data by controlled experiment they can be satisfied with classical statistical theory and live in Asymptopia. It is in econometrics theory that we get the study of errors in variables, the trade off between bias and efficiency when tackling multicollinearity and the use of informative or uninformative Bayesian priors. Exactly because the data is generally poor the statistical theory has been pushed further to generate different statistical estimators.
    Lars might discriminate more between bad practice in the discipline of econometrics and the discipline itself if he knew a bit more about it. It is one of the few methods we have, when used conscientiously, for discriminating among economic propositions. We need more empiricism, not less.

  3. January 17, 2020 at 2:20 am

    “What Leamer does not mention is that econometrics has taken the analysis of small, dirty data sets further than any other discipline.” — Econometricians have no understanding of causality, and have resisted incorporating advances in causality, without understanding which it is impossible to do a sound regression analysis. See Pearl’s Book of Why: Chapter 2 tells the bizarre story of how the discipline of statistics inflicted causal blindness on itself, with far-reaching effects for all sciences that depend on data.

  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.