Identification versus Importance
26 November 2008 at 11:17 am Peter G. Klein 9 comments
| Peter Klein |
At a recent workshop the subject of econometric identification came up. Identification is of course the major issue of our day among mainstream empirical economists. Some have described the dissertation process as the “search for a good instrument.” Instrumental-variables estimators have their critics, of course, but these critics are in the minority.
One of the workshop participants, a regular attendee at NBER events, summarized the consensus view among the elites of the profession with the following diagram:
A research problem can be important, and it can be well identified. The ideal problem is one in quadrant B, both important and identified. However, a problem in quadrant C is much more likely to be published in a top journal than a problem in quadrant A.
What does this say about the economics profession?
Entry filed under: - Klein -, Methods/Methodology/Theory of Science.
1.
MomentarilyAnonymous | 26 November 2008 at 12:53 pm
Obviously there are two ways: a–>b) to properly identify important issues or c–>b) to ‘problematize’ identified issues. If I was forced to choose, I would prefer pushing A to B because that’s a good way to make research in economics relevant to society at large. For example, I think a good deal of the behavioral economics literature represents a shift from A to B.
2.
david | 26 November 2008 at 2:12 pm
What does this say about the economics profession?
That we have trouble identifying important issues?
3.
Stuart Buck | 26 November 2008 at 5:07 pm
Brings to mind the old joke about the drunk and the streetlight.
4.
Stuart Buck | 26 November 2008 at 5:12 pm
What about Michael Murray’s conclusion: “The barriers to Archimedes moving the earth with a lever were more daunting
than the challenges facing instrumental variables estimation, but the comparison is apt. The perils of invalid and weak instruments open all instrumental variable estimates to skepticism.”
5.
Jason | 22 December 2008 at 10:20 pm
That economics is on the right track?
Anybody can come up with stuff tackling the ‘big’ issues if they’re lax in their identification. You don’t want to call that research properly done though.
For other social sciences, change “identified” to “having good evidence” and you’ll want C types more than A types too.
6.
David Hoopes | 26 April 2011 at 11:17 am
Saw a paper presented at Wharton a “few” years ago with very big identification problems. The results were useless. It was published, as is, in ASQ a few years later. At least economists are trying.
7. Randomized Controlled Trials and Economic Questions | 28 December 2018 at 3:13 am
[…] design may come at the expense of the importance of the underlying economic question (a point I have also made). RCT enthusiasts have been called randomistas, not always as a […]
8. Randomized Controlled Trials and Economic Questions – USSA News | The Tea Party's Front Page | 28 December 2018 at 6:32 pm
[…] design may come at the expense of the importance of the underlying economic question (a point I have also made). RCT enthusiasts have been called randomistas, not always as a […]
9.
Pruebas controladas aleatorizadas y preguntas económicas, por Mises Hispano. – Minarquia 2 | 18 January 2019 at 1:01 am
[…] puede producirse a costa de la importancia de la cuestión económica subyacente (algo que yo también he dicho). A los entusiastas de las PCA se les ha llamado randomistas, lo que no siempre es un […]