Welcome back to the Conglomerate Junior Scholars Workshop. Today's paper is Alexander "Sasha" Volokh's Privatization and the Law and Economics of Political Advocacy. Sasha is a Visiting Assistant Professor at Georgetown University Law Center where he teaches Law and Economics, Regulation Law and Economics, Environmental Economics and the Law, and Delegation and Privatization. To do justice to a paper written by an economist, I had to call upon two distinguished economists who are experts in the field of law and economics, Tom Ulen and Paul Rubin. In addition, Brian Galle, former workshop participant, has provided valuable feedback.
We invite readers to comment on the paper (and the comments) in the comments section of this post. In the interest of running this workshop like a physical world conference, no anonymous commenters, please.
The abstract for the paper is here:
A common argument against privatization is that private providers will self-interestedly lobby to increase the size of their market. In this Article, I evaluate this argument, using, as a case study, the argument against prison privatization based on the possibility that the private prison industry will distort the criminal law by advocating for incarceration.
I conclude that there is at present no particular reason to credit this argument. Even without privatization, government agents already lobby for changes in substantive law—in the prison context, for example, public corrections officer unions are active advocates of pro-incarceration policy. Against this background, adding the “extra voice” of the private sector will not necessarily increase either the amount of industry-increasing advocacy or its effectiveness. In fact, privatization may well reduce the industry’s political power: Because advocacy is a “public good” for the industry, as the number of independent actors increases, the dominant actor’s advocacy decreases (since it no longer captures the full benefit of its advocacy) and the other actors free-ride off the dominant actor’s contribution. Under some plausible assumptions, therefore, privatization may actually decrease advocacy, and under different plausible assumptions, the net effect of privatization on advocacy is ambiguous.
The argument that privatization distorts policy by encouraging lobbying is thus unconvincing without a fuller explanation of the mechanics of advocacy.
TrackBack URL for this entry:
Links to weblogs that reference Conglomerate Junior Scholars Workshop: Alexander "Sasha" Volokh's Privatization and the Law and Economics of Political Advocacy:
1. Posted by David on July 9, 2007 @ 11:21 | Permalink
I liked this article, which - full disclosure, - I saw at a slightly earlier stage. In a way, Sasha's account seemed to me to turn on a Coasean style analysis of transaction costs. The question being: will privatized public institutions lead to suboptimal pressures on the political process? And the answer seeming to turn on the likelihood that already vested interests catered to by those institutions - like public employee associations - would provide those pressures anyway. It all turns on who has the lower transaction costs in delivering the suboptimal pressure ... not to toss around ten dollar economic theory lightly.
I wonder if the influence of interest groups in setting public policy - such as incarceration policy - is susceptible to the same analysis.
2. Posted by Sasha Volokh on July 9, 2007 @ 15:46 | Permalink
Thanks to Tom Ulen, Paul Rubin, and Brian Galle for their insightful comments. I’ll try to keep my responses short, and I’ll use a separate comment form for each commenter.
In response to Tom Ulen:
1. The existence of K Street is indeed somewhat of a challenge to any rational-choice theorist who takes Mancur Olson seriously. Fortunately, there are various ways of explaining K Street within the theory.
First, as Olson himself writes, the trade groups may be giving out selective incentives; firms or individual join trade groups for unrelated reasons (insurance services, social networking, etc.), and the group leverages its membership, through dues, into political influence.
Second, as I write in Section IV.B (pp. 39–42), there might be tacit cooperation (what we economists call “collusion”), where contributions to the trade association are enforced by the threat of no longer cooperating — possibly a threat to no longer collude in the underlying product market, i.e., to no longer collude in auctions for private prison contracts. The more interaction there is among the different actors, the greater the possibility for collusion, and so the greater the threat to stop colluding if someone doesn’t make the requisite political donations. This is why I speculate that private-sector cooperation is more likely than cooperation between the private and public sectors, which is why I run my examples with two actors in a 90–10 split instead of with several actors in, say, a 90–5–2–2–1 split.
Third, various trade associations exist but are underfunded, don’t do much, and seem to play no significant role. In the prison context, the Association of Private Correctional and Treatment Organizations (APTCO) seems to fit that bill. (See pp. 30–31 & 41.)
2. About free-riding assumptions, I agree that the experimental games literature, which shows lab subjects violating standard rational-choice free-riding assumptions, is a potential challenge to the standard theory. The title of one of the papers says it all: Gerald Marwell & Ruth E. Ames, “Economists Free-Ride, Does Anyone Else: Experiments on the Provision of Public Goods,” 15 J. Pub. Econ. 295 (1981). (This paper reports the results of lab experiments in which a group of economists was the only group to consistently free-ride.) See generally the sources cited in the second half of note 220, pp. 50–51.
I don’t make much of this literature, aside from citing it, because I find the free-riding hypothesis consistent with what we see in the prison industry; but if we did observe an industry with an unexplainable absence of free-riding, these sort of experimental/behavioral explanations may have some force.
3. Posted by Sasha Volokh on July 9, 2007 @ 15:47 | Permalink
In response to Paul Rubin:
1. I agree that most (not all) lobbying for increased imprisonment and sentencing would occur at the state, not federal, level. This is just because the states still do most of the prosecutions — there are about 7 times more state inmates than federal ones, though this ratio has been dropping, with a 2.5% annual increase for state prisoners over the last 10 years compared to 7.2% for federal prisoners. In any event, this means that both the federal share and the state share matter.
This is why, on p. 21, I report both state and federal shares, and on both a stock and flow basis: “The private sector has a smaller share of the industry. Of the 1.5 million prisoners under the jurisdiction of federal or state adult correctional authorities in 2004, 7% were held in private facilities; this includes 14% of federal prisoners and 6% of state prisoners. Among the 34 states with at least some privatization, the median private percentage was 8–9%. If we are interested in the private share of marginal prisoners—i.e., how likely a prisoner is to go to a private prison if he is convicted today —the private share becomes larger, mainly because private firms have absorbed much of the recent growth in federal incarceration. A reasonable estimate of the private share of marginal prisoners over the period 2000–2005 yields 6% for state systems, 54% for the federal system, and 22% overall.”
The suggestion that I propose some ways to test the hypotheses using state data is well-taken, though I don’t think the current data is good enough to actually run such an analysis.
2 & 3. On the technical version of the paper, available on SSRN, we’ll have to agree to disagree on whether “a little more of the flavor of the mathematical version would have somewhat improved” the current paper. When this paper had more math in it, the eyes of the non-techies — my target audience — glazed over.
This also explains the length: I’ve tried to bend over backwards to make the simple mathematical point accessible to non-techies. The result is a paper of approximately 25,000 words, which is, funnily enough, considered short by law review standards!, though this probably reflects badly on law review standards more than it reflects well on me. Well, this is why we have multiple versions.
4. Posted by Sasha Volokh on July 9, 2007 @ 16:13 | Permalink
In response to Brian Galle:
1. The question what motivates unions still bedevils labor economists, and public-sector unions, doubly so. The “rent-maximization” hypothesis — that unions maximize “union rents times the size of the industry” — is common enough in the labor literature, which I cite at length in note 68, p. 22, and I don’t want to be reinventing the wheel on union motivations.
But I have a better defense than mere deference to labor economists. I think rent maximization is also defensible, at least approximately. If unions are a pure democracy, and if we can model wage determination (through collective bargaining) as a one-dimensional variable, then the union acts as though it had the preferences of the median member. That member wants a higher wage, but he’s also aware that he could become unemployed if the wage goes above the market wage. So, making a ton of assumptions along the way, his utility can be modeled as something like “union rent times the probability that he stays employed.” (Most importantly, we’ve assumed that he’s risk neutral, or else the “union rent” should really be “the difference in utilities between union and market wages.” But see John Pencavel, “Wages and Employment Under Trade Unionism: Microeconomic Models and Macroeconomic Applications,” 87 Scand. J. Econ. 197 (1985), arguing that the rent maximization approach is appropriate if the union redistributes income from employed to unemployed workers so as to equalize incomes.)
Now this median-worker approach doesn’t get us directly to rent maximization. But the probability that the median worker stays employed does at least bear some relation to the size of the industry. (Though, as you say, “prison crowding may make their jobs more secure, and increase their overtime pay, but there are limits on those benefits, and they are balanced to an extent by the likelihood that crowding increases their personal risk.”)
Finally, we can drop the assumption that the union is a perfect representative for its workers, and introduce agency costs — in particular, the tendency of self-aggrandizing union officials to seek union membership growth even when this wouldn’t be in the best interests of union members. As I write on pp. 36–37: “A larger sector may mean a more powerful union and therefore potentially higher wages, benefits, or job security down the road (and perhaps—to introduce agency costs for a moment—perks for union officials).” See, e.g., Stewart J. Schwab, “Union Raids, Union Democracy, and the Market for Union Control,” 1992 U. Ill. L. Rev. 367, 380–81.
To conclude this point, I agree that showing the precise point for prisons is substantially more complicated than just comparing total union rents to total profits. This is why I only characterize my calculations as “rough estimates” (p. 20), and conclude: “my analysis of what motivated public-sector unions, while common in the labor economics literature, is highly simplified; in assuming that private prison firms were profit-maximizing, I suppressed any analysis of agency costs within the firm; and my back-of-the-envelope estimate of the benefit of incarceration to the different sectors was just that—an estimate. . . . So my specific conclusions here are tentative” (p. 52). This is just meant to be a jumping-off point for more detailed studies.
2. Same goes for how exactly political influence is used, once acquired, and whether it’s dissipated or has a cumulative effect. The same goes again for my assumption that public- and private-sector jobs are comparable enough for the rough validity of the back-of-the-envelope calculation; it’s well-known that union workers earn rents from being unionized, but the magnitudes are of course subject to dispute. My main goal is to show how the casual charge that privatization worsens political advocacy can easily be shown to be untrue under certain assumptions. The point is not to prove that the charge is false because those assumptions are true, but just to argue that, since the assumptions might be true, whether the charge is true or false requires more detailed inspection. I hope that future work (by myself or others) will pick up that challenge.
3. In light of this, the suggestion that I overreach at pp. 51 & 53 is well-taken, and I’ll moderate that language in a later draft.