Just a Little Off the Top

I'll admit to being more than a little shocked at the findings of this new NBER study on corruption in America. A couple of Harvard economists did a cross-county comparison and found, not terribly surprisingly, that corruption is lower in areas with higher education and income (where, presumably, people have the luxury of keeping closer tabs on their elected officials). The shocker was the absence of any independent correlation between corruption and the size and scope of government, which amounts to saying that you don't get more corruption when the opportunities to be corrupt (more rules for which to peddle exemptions, more revenue to skim) increase. Haven't had a chance to read the whole paper yet, but it's an interesting and highly counterintuitive finding if it holds up.

Comments

[Kevin Carson]

Correlating corruption to the size of government depends on how you define "corruption." For most people, corruption refers less to the general goals of the system than to how they're implemented.

The purpose of the present system is largely corporatist: to subsidize private profit and externalize cost and risk on the taxpayers. But that purpose is sold to the general public in terms of the "public welfare" or "common good."

And since the corporatist system is administered by the New Class, one of its central values is "professionalism." To the average person, "corruption" means deviating from this ethos of professionalism under the influence of good ol' boy politics. The system becomes corrupt only when the people at the helm of the state stop acting as an executive committee of the ruling class as a whole, and start engaging in crony capitalism on a more personal level.

[Jacob T. Levy]

The authors are very cautious about that finding for methodological reasons.

One of their measures of government size seems silly to me: number of state legislators per capita. (By that measure sales-and-income-taxless New Hampshire has a vastly larger government than any other state.) The tax measures I don't especially expect to see a correlation with. The regulatory measures are the interesting and important ones-- though it's hard to measure the real key variable, which is *discretionary* regulation. Still, I'm surprised by the lack of correlation even with the gross regulation variable, even given the authors' methodological cautionary notes.

[Grant Gould]

The entire purpose of the regulatory state is to make corruption unnecessary, by allowing notionally lawful mechanisms like regulatory capture to take the place of unlawful mechanisms like kickbacks. It should hardly be surprising that it succeeded -- indeed, larger government should show less corruption because it offers a wider range of legal methods for government to pick and choose winners and losers.

Nothing is as corrupt as a midsize-town Zoning Board of Appeals. Nothing is so fastidiously legal as the same board in a large-city. But the difference does not extend to different behaviours: They do the same thing, but the latter has the resources and power to find ways to shoehorn kickbacks and graft into a legal framework.

[joe]

Kevin, Grant, forgive me, but you're both completely insane.

Only on a libertarian site would people argue that there is no difference between requiring an applicant to put in a traffic light, and requiring him to give you an envelope full of twenties.

[db]

Only on a libertarian site would people argue that there is no difference between requiring an applicant to put in a traffic light, and requiring him to give you an envelope full of twenties.

Maybe, but usually the application process (including compliance with all implied requirements) costs a lot more than an "envelope of twenties".

[joe]

Maybe maybe not, db. My point is, calling both situations "corruption" obscures more than it reveals about what's going on.

[ldy]

How are they measuring corruption?

Are they measuring reported corruption, and if so how does that correlate with actual corruption?

Or are they measuring unreported corruption, in which case how the hell do they do that?

[Mo]

Jacob,
What does number of state legislators per capita have to do with taxes? It's just the number of legislators per person.

[Jacob T. Levy]

Mo:
My point is that number of legislators per capita doesn't proxy for size of government in any way that's possibly theoretically relevant to the claim that larger governments have more rents and therefore attract more corruption. Taxes and regulations do, as does their general economic freedom measure. The number of state legislators per capita seems like an irrelevance. It's a "size of government" measure only on a very simple-mindedly literal interpretation of "size of government," as if one used the square-footage of the governor's mansion.

ldy:
They use federal corruption convictions, which is a pretty good idea. Unreported convictions are unreported; state convictions or prosecutions are endogenous (corrupt judicial systems are less likely to prosecute for corruption). When measuring variation across the states, using a measure that isn't on its face tied up with state-level corruption seems like the way to go. Note that it doesn't matter if the federal judicial system is also systmetically corrupt (though it's not, or no tnearly at the level of some states), as long as its corruption is generalized and uncorrelated with state-level corruption. Variation in federal corruption convictions across states is probably a very good proxy for variation in corruption across states.

We also don't need to know what proportion of corruption is detected; we only need to know that there's no reason to expect other-than-random state-to-state variation in that proportion.

[Jamie S]

Federal corruption convictions as a proxy for "Corruption." Since the feds can't be everywhere at once and selectively prosecute only the most egregious offenders to "send a message" to the rest, I would not expect the above proxy to work very well. The feds are only going to prosecute to the extent that the administration prompts them to. In this sense, lower levels of prosecutions may actually hide greater levels of corruption, since it's easier to get away with under those administrations.

To those of you who think it's crazy to equate lawful graft with unlawful corruption, you reveal an inability to separate morality from legality. A lawmaker saying something is legal doesn't mean it's not corrupt. This is not necessarily a libertarian viewpoint. The Dems say the the war is corrupt but legal, while the religious right says the same about abortion.

[Jacob T. Levy]

Jamie S: You seem to be talking about variation in levels of federal prosecution over time. But that's not at stake here. The paper is about variations across jurisdictions (states/localities). Unless we have some reason to think that the U.S. Attorneys are less likely to press cases in more corrupt states than in less corrupt ones, your point (even if true) doesn't affect this paper.

If it's the case that, for example, U.S. Attorneys under Bush are unwilling to bring prosecutions in Florida or Texas, and as a result corruption flourishes in those jurisdictions, then Jamie's right and the paper is invalid from the word Go. But I rather doubt that.

[Kevin Carson]

joe,

I challenge you to show where I said there was "no difference." My intent was precisely the opposite: to show how the two styles of "corruption" differed, and point out why the public saw one as corrupt but not the other.

And what I had in mind was not so much the government requiring a developer to put in a traffic light, as a national corporation getting the feds to fund most of its R&D expenses.

[Agammamon]

". . .Only on a libertarian site would people argue that there is no difference between requiring an applicant to put in a traffic light, and requiring him to give you an envelope full of twenties. . "

that may be Joe, but I consider the latter to be more honest.