« What I've been reading | Main | Don't burn it »
Johannes Fedderke and the importance of good governance
File him in the category underappreciated economists. Does good governance matter for growth? Could there be a more important question for economists? The standard cross-sectional growth tests do not show much of a robust effect. But Johannes, along with co-authors Robert Klitgaard and Kamil Akramov, has a 150-page paper showing that if you take all the relevant heterogeneities into account yes, Adam Smith and Doug North were right after all.
Or do you prefer simple regressions which meet the eyeball test?
Here is the full paper. Here is Johannes's long paper on South African economic history.
Posted by Tyler Cowen on February 15, 2008 at 11:36 AM in Economics | Permalink
Comments
I prefer the eyeball test.
Posted by: AMW at Feb 15, 2008 1:53:45 PM
I prefer the eyeball test.
Posted by: AMW at Feb 15, 2008 1:54:20 PM
Minor annoyance: in this day and age, I do not understand why people continue to put tables and figures at the end of the paper.
Plus, you'll never know if you have *all* of the relevant heterogeneities, only the ones you can identify.
Posted by: bartman at Feb 15, 2008 2:08:18 PM
And how does a country get/sustain/foster good governance? Isn't governance just the distilled cultural institutions of a national culture?
Good, similar, work in The Central Liberal Truth: How Politics Can Change a Culture and Save It from Itself by Lawrence E. Harrison.
Posted by: The other Eric at Feb 15, 2008 3:03:56 PM
Who is in favor of calling the
"Or do you prefer simple regressions which meet the eyeball test?"
The Al Gore Method?
Posted by: Jay at Feb 15, 2008 3:21:54 PM
Who is in favor of calling the
"Or do you prefer simple regressions which meet the eyeball test?"
The Al Gore Method?
Posted by: Jay at Feb 15, 2008 3:21:56 PM
"Plus, you'll never know if you have *all* of the relevant heterogeneities, only the ones you can identify."
Trivially true.
"in this day and age, I do not understand why people continue to put tables and figures at the end of the paper."
That's what the journals demand.
Posted by: LemmusLemmus at Feb 15, 2008 4:34:32 PM
Yes, one can never be sure that one has identified all the relevant factors for correction. But unless it expresses the result of study with a control group, one can be sure that a "simple regression that meets the eyeball test" has corrected for no relevant factors, which makes it even more suspect.
Posted by: David Wright at Feb 15, 2008 6:53:44 PM
Since good government is a public good (at least in a democracy), how can we trust judgments of what "good" governance is? Who has the incentive to judge and report it accurately? It seems like its easy for politicized groups to cherry-pick pieces of governance that they agree with and claim those are the causes of growth. I've often wondered if Cato's and Heritage's indexes simple correlate freedom with wealth, then define freedom from that. Of course, I don't mean to imply that all such biases are conscious; I think most probably aren't.
Posted by: Grant at Feb 16, 2008 2:52:55 PM
At least this is a + serious attempt to look at institutions that the Cambridge agenda -e.g, Acemoglu, Robinson, Rodrik, Banerjee, etc.
Can some tell me why settler mortality is a good instrument? The authors said: there is a bundle of institutions that matter. They only use one indicator in their regressions. Therefore, the IV is correlated with the error term. Confessed by the authors.
Posted by: Lennon at Feb 17, 2008 3:00:55 PM
