Welcome to Econ Talk, part of the Library of Economics and Liberty. I'm your host Russ Roberts

of George Mason University and Stanford University's Hoover Institution. Our website is econtalk.org

where you can subscribe, find other episodes, comment on this podcast, and find links to

other information related to today's conversation. Our email address is mailadicontalk.org. We'd

love to hear from you. Today is May 3, 2010, and my guest is Ed Lemer, the Chan C.J. Medbury

Chair and Management at UCLA. Ed, welcome back to Econ Talk. It's great to be back. Ed, the topic

for today's podcast is the state of econometrics, the application of statistical techniques to

economic questions. A few weeks ago, Tim Hartford wrote a piece in the Financial Times, referencing a

famous article that you wrote back in 1983 called Let's Take the Con out of econometrics. Hartford

argued we'd finally succeeded in solving at least one crucial problem. It did take 27 years, but we

finally removed the con, and we've got more honesty, and in particular he was focused on the identification

problem, and what he claimed is he was referring to other work by Angerist and Pishky to

econometrisians who argue that by the use of so-called natural experiments and modern techniques, we've

been able to get a much better assessment of relationships in economic data. First, I want to ask you

to talk about your piece in 1983, and what was the con that you were saying is in econometrics that

we ought to be aware of? Well, the con is that depending upon what model you select, you happen to

use for analyzing the data set, you can get dramatically different estimates and dramatically

different conclusions. An economist had not spent enough effort alerting their customers to

that sensitivity, and that's the con. You're pretending that the data sets are providing more

clear information than they possibly can, because the econometric method requires you to make a

complete commitment to assumptions that you have at best a half-hearted commitment to. So I was

arguing, well, what we need to do is develop tools that individuals can use, research can use,

...