Volume-mates with Orin:

Orin recently announced that his article on Four Models of Fourth Amendment Protection is forthcoming in the Stanford Law Review.

I'm delighted to report that I, too, have accepted a publication offer with Stanford — in my case, for my article on Privatization and the Law and Economics of Political Advocacy. (The link in the previous sentence will take you to the very latest version on SSRN — exactly 25,000 words!) Here, again (for those of you who didn't pay attention when I've posted this before), is the abstract:

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.

Those of you who want a technical, economicsy, paper on the topic can check out this one. In the near future, I'll have a series of posts summarizing the argument of this paper.

liberty (mail) (www):
"exactly 25,000 words!"

nice. How long did that take to achieve?
3.28.2007 2:45pm
Sasha Volokh (mail) (www):
It took me a couple of days to get it down from about 30,000 to about 25,000. Then removing the little stuff to get it right down to the threshold isn't all that hard.
3.28.2007 2:48pm
This leads to the interesting question of whether being other pieces appearing in the same issue can improve citation rank. Obviously it doesn't to the extent that people find articles via Westlaw and Lexis, but now people are starting to go to the journal's website to find them when they first come out.

I'll let you and Orin fight about who's article is more likely to raise the readership of the other.
3.28.2007 4:30pm
Viscus (mail) (www):
Based on the abstract, this paper seems to be theoretical rather than empirical. The claim seems to be that it is theoretically possible that privatization might result in less advocacy to increase market size.

While this might be theoretically possible, is it probable? It doesn't strike me as particularly likely.

I don't see any decreased incentive on part of corrections officer unions to lobby for stricter incarceration policies. While one might not label a government-run prison system a free rider, but then label a privately-owned prison system as such, I don't see how the change from public to private changes the incentives of the unions. Before the third party "beneficiary" was a public system, now it is a private system. I don't see how it makes a bit of difference from the unions' perspectives.

The only argument I can think of is that a private system might better able to curtail the power of correction officer unions, lowering compensation levels, and thus decreasing the resources available for lobbying, decreasing the incentive to lobby for increased incarceration and simultaenously leading to a rational shift of resources to lobbying for increased power vis a vis the now privately run prison system rather than lobbying for increased compensation. But then, you will have to make an argument about how private owners will be better able to resist unions and, I think, to make your argument worthwhile, make an argument that suggests that the difference will be large, rather than small.

There have always been "free riders." Even now, there are non-union beneficiaries of public correction officer lobbying for stricter incarceration (the prison system, crime victims, construction companies, construction workers, food services companies, sanitation companies and so on). Changing the identity of one of these free riders (i.e. the prison system, from public to private) probably does not change union incentives, and even if it does, the effect is likely small. Any effect will be outweighed by the increased political power of the new entity (private prison companies can provide campaign contributions, public prison systems cannot) unless the changes in union incentives are large, which there seems to be no reason to believe.

Overall, based on the abstract, I do not find this paper to be promising. There is no hint that you deal with these issues of magnitude with empirical evidence. Alas, I do not have the time or inclination to read a 25,000 word paper (which is probably not that far from 50 single spaced pages in Microsoft Word) based on this abstract. However, I am sure that other interested parties will find the time to read it, since you have managed to get it published in the Stanford Law Review.
3.28.2007 4:34pm
liberty (mail) (www):

There are a few things you have overlooked. One is that privatization will lead to many individual companies which compete and hence the role of unions will differ. If a union is for all private prison employees, it will want more jobs for private prison employees, not more jobs for all prison employees, and so it will do better to lobby for privatization than to lobby for incarceration. Some unions may be just local to a certain company and will simply lobby for increased wages or job security, etc. Some companies will tend to hire people not in any union.

In addition, I think an important factor is that private companies can focus on more areas than just incarceration. If the company also runs rehab clincs and halfway houses and other alternatives to incarceration then an associated union wouldn't lobby for more incarceration because they would do just as well with lower incarceration and more probation, for example.
3.28.2007 5:17pm
Viscus (mail) (www):

I do not think that the existence of privately run prisons is in fact likely to result in major fissures within the corrections unions. I suppose one could make an argument that with some union members bargaining with a public entity and others bargaining with a private entity, that their interests will diverge so dramatically, as to cause an important split within the union. This strikes me as unlikely.

In general, both unions for public and private firms would benefit from increased incarceration and so would have a common interest in lobbying for increased incarceration. Even if structurally privatization lead to an actual organizational split (which seems unlikely) they would be in a position to form a "joint venture," to fund the costs of lobbying for increased incarceration. More likely, the same union would represent both workers who work for private and public prisons and thus all dues from increases in incarceration would flow to the same entity. UAW, for example, did just fine representing workers for three different companies. Finally, it should be noted that most privatization advocates want all prisons to be privatized, rather than maintaining some sort of hybrid system whose sole purpose is to split the unions. If the whole system where privatized, any obstacles theoretically caused by a public/private split would be cured.

Second, that private companies might diversify to other areas doesn't really lessen their incentive to lobby for incarceration. From their perspective, a halfway house or rehab clinic is the perfect place for a criminal to spend time after they have done their prison time. And the more time they spend in the rehab clinic or halfway house, the better.
3.28.2007 5:45pm
Sasha Volokh (mail) (www):
Viscus: Let me cut through all this by pointing out that private prisons, nationwide, are mostly not unionized. There are strong reasons why the public sector is more likely to be unionized than the private sector.
3.28.2007 6:03pm
liberty (mail) (www):

I think the empirical evidence shows that industries which are fully private and unsubsidized - which are more competitive - tend to have lower unionization. For example, recent unsubsidized competition in the airline industry (Jet Blue has no union membership and Southwest low unionization), or compare farming (with heavy subsidies) and public industries with fully private industries like retail.

Part of this is because evil big business will do everything it can to squash unions and hire non-union labor.

You are correct that the first of the above arguments (advocacy for privatization) only works so long as its not completely private -- but when it is, it is quite likely that unionization will have fallen quite a bit, given that private industries tend to have lower unionization. As far as lobbying for longer incarceration and rehab if the company runs both -- it may depend on how much bang for the buck you can get. For example, lobbying for increased sentencing (perhaps split between pushing for prison times, pushing for longer rehab and pushing for both) gets me longer sentences in prison without affecting probation times 1/4 of the time, longer time in probation without affecting prison times 1/4 of the time, longer prison time but reduced probation rehab 1/4 of the time and longer probation but reduced prison time 1/4 of the time, then I have wasted half my money.

On the other hand, maybe spending nothing would have done nearly as well, and spending the money instead on pro-union ads or just on bargaining with the company would have done the workers better.
3.28.2007 6:05pm
Viscus (mail) (www):

Why are some private enterprises able to resist unionization but others not? It obviously has nothing to do with them being private, per se, but rather goes to structural reasons, whether the underlying legal environment, or the nature of the industry. It is difficult to unionize computer programmers. Computer programmers have vastly different skill sets and levels of productivity. They also have a tendency to move from employer to employer quite ruthlessly. On the other hand, corrections officers are quite interchangeable and they tend to stay put with the same employer over a long period of time. The field of corrections is quite conducive to unionization, while the field of computer programming is not.

Your suggestion that unions will not thrive under private sector management is entirely speculative, especially since unions, to the extent that they exert influence now, are likely to be able to extract legislative protections as part of any privatization deal. Of course, there is something to the idea that private companies will expend resources to decrease the influence of the unions. But there is also something to the idea that unions and private companies will cooperate to increase incarceration.

With respect to your made up probabilities, that part of your analysis is lacking. It is not the case that unions will achieve each of these results precisely 25% of the time when they lobby for increased sentencing. And logically, you have forgotten the big result that will be pushed for by such companies. Longer prison sentences and longer rehab time. These results are not mutually exclusive, and could be accomplished by the very same bill. But they especially are not mutually exclusive in the long-run. Private companies running prisons will be an entrenched interest lobbying for increased prison time and increased rehab in legislative session after legislative session. Not only that, but their long-term agenda of increased prison time and increased rehab is favored by politics. When was the last time you heard a politician run on an agenda that included shorter prison terms?

Private companies will probably be as powerful advocates for increased incarceration as corrections officers unions are now. Not only that, corrections officer unions are not going away. They will join such private companies in lobbying for increased incarceration. Free riding exists now. It will not be greatly increased by privatization. Overall, free riding notwithstanding, lobbying for increased incarceration will increase and likely be more effective. More "bipartisan" deals will be possible. Company-backed Republicans will agree with union-backed Democrats that more incarceration and more treatment is necessary.
3.28.2007 6:33pm
Viscus (mail) (www):

In what states are private prisons prevalent? Red states? Did they actually replace existing union-run prisons, or just start running new prisons that never had union corrections officers? If they replaced union-run prisons, how strong were the unions in those states in the first place?? Would similary success in replacing unions be expected in states that have stronger unions??

Also, if additional prisons are going to be privately run, is there really that much of a "free rider" problem with respect to lobbying for increased incarceration, since the beneficiary of increased incarceration will go to one of a few private companies capable of running newly constructed prisons?? Given the scale at which prisons are run and the specialized expertise required to run them, isn't it reasonable to think that a few major players will emerge. And doesn't our experiences with copyright show that when there are a few major players, they can be very effective at lobbying. Free rider problems notwithstanding, we still have the DMCA. That is what I call effective lobbying.

Overall, there is little reason to doubt that private companies will be as effective at lobbying as private sector unions. They may be even more effective, since they will be able to bring multiple strategies and probably more sophisticated and ideologically diverse actors to the lobbying table. Large prison companies are likely to have a large influence on Republicans. Unions will influence Democrats. The result is bipartisanship.

Given our actual experience with very effective lobbying in the face of a theoretical free rider problem (i.e. copyright) I think the suggestion that free rider problems will overcome the increased lobbying power resulting from privatizing prisons is nothing more than wishful thinking. Large actors tend to overcome free rider problems.

Thinking about the nature of increased prison sentences, this is especially true. An increase in prison time for a particular crime creates a large stream of future revenue that may very well far exceed lobbying expenses. It is especially nice that such sentencing increases do not directly affect state budget bills, but rather bring pressure later on, as the demand for prison construction and prison budgets increase.

I suppose one major question is the following. Is the profit increase to a private company from incarcerating an additional prisoner for one year greater or lesser than the increase in union dues that a union enjoys for that prisoner. My initial guess is that company revenue per prisoner will be higher than union revenue. But that union revenue will still exist, especially in states that are not hostile to unions.
3.28.2007 7:01pm
liberty (mail) (www):

You make the case that certain "structural reasons" explain why some companies are unionized and some not, but some industries (like airlines) have changed in this way, arguably because they have become more private: the new cheap airlines are not dependent on government for bail outs and are more competitive. You can see this in other industries too as they have been privatized or have lost subsidies, etc. There is empirical evidence to support the notion that competitive private industries are less unionized than public or subsidized industries.

You then argue that if these private firms lobby for both longer prison time and longer rehab time, they will get it. But this is assuming the result. Yes, I made up the number 25% but its true that they might not get both: even if they argue for both, a deal may be reached which increases one or the other.

You make an interesting point that if private firms and unions agree on a single goal, they may have better luck achieving it. However this depends on the notion that both actually want to lobby for this outcome and that they want it enough not to free ride off the other. If each knows that only part of his money will go toward his goal and much will benefit other companies and also knows that others will be doing the lobbying he desires and also knows that he can make money whether prison sentences or rehab sentences are expanded and hence unless both are expanded it would be a waste to lobby for it, then each might independently determine that he better off free riding.
3.28.2007 7:15pm
Sasha Volokh (mail) (www):
Viscus: One lesson from the labor economics literature is that there's a public-sector union wage premium -- in other words, unionized public-sector workers make more than their non-unionized private-sector counterparts.

Now how much of this is because of being in the public sector, and how much is because of being unionized? Labor economists differ. Some say it's because being in the public sector itself raises wages, because the government is, in part, trying to achieve social objectives through public-sector employment. Others say it's mostly because of unionization, and public-sector workers just have an easier time unionizing, partly because of existing U.S. law and partly because the government, not being subject to the same sorts of competitive pressures as the private sector, doesn't resist unionization as much.

Either way, this supports my point, which is that it's logical to think that a privatized industry is less likely to be unionized. Fortunately, the empirics are on my side, since we do in fact observe starkly different unionization in the public and private sectors, and in particular we also observe this for prisons.

By the way, it's also logical to think that less competitive industries are more likely to be unionized. If companies are making zero economic profits, there's not much leeway for a union to raise labor costs unless labor law makes this happen across the whole industry. But if companies are making profits -- for instance, because they're monopolists or have very few competitors -- then the precise split of those profits between different factors of production is subject to negotiation, and it's not surprising to see unions getting some of that pie. See, for instance, the auto industry.

Anyway, this is why corrections officers unions are strong for public prisons, and quite weak for private prisons (and nonexistent in many states). This is reflected in the wages these guys make. Public corrections officers' wages are much higher than those of their private counterparts -- by, I suggest in my paper, a range of 30-65%, though there's a lot of variation. Basically, private guards are making market wages; public guards aren't. So public guards have a lot to gain by increasing the size of the industry; private guards have less.

Even without the unions, you would think that the prison firms themselves would be lobbying to increase the size of the industry. This is, in fact, what most of the anti-privatization-for-lobbying-reasons people argue -- the unions are the problem in the public sector, but the firms are the problem in the private sector. In my paper, I explain how this is a non-crazy idea, but (1) I haven't seen any actual evidence of this, and (2) in fact it turns out that the private prison companies aren't terribly profitable. When you're making low profits, you have little to lose by expanding the industry. If I found that private prisons had huge profits, I wouldn't be surprised -- it's a fairly concentrated industry, and maybe these guys are colluding with each other to rig the private prison contract auctions -- but in fact, I find that they don't have huge profits.

So the idea that public-sector actors have the incentive to lobby a lot, and actually do, and that private-sector actors don't have as much of an incentive to lobby, and actually don't, is both theoretically and empirically defensible.

Though the paper is 50 pages long, I recommend you skim it and then let me know what you think.
3.28.2007 7:22pm
Viscus (mail) (www):
I suppose this should be turned into one of Aesops Fable's: never judge a paper by its abstract.

The abstract just made me think that the paper was theoretical, rather than empirical. But if your paper does in fact have good empirical evidence in it, it is worth reading.
3.28.2007 7:58pm
Brian K (mail):

It seems to me that there are two potential causes of why private companies are less likely to be unionized that you're glossing over. (1) Private companies are better able to resist unionization of their employees. (2) Private companies are able to bust an already established union.

(1) You've made a good case that is true esp. since I already believe this. But this situation doesn't really apply since prison employees are already unionized. I agree with Viscus when he made the point that during the privatization process unions are likely to get concessions that make it easier for them to maintain thier membership in the private sector.

(2) This seems to me to be the applicable situation since many (most?) prisons are already unionized. Given the troubles that airline and car companies are having with getting union concessions and/or breaking the union, I can't imagine how it can be expected that prison companies are going to have an easier time, if they are able to do it all. This will be especially harder with the concessions unions bring about during the privatization process.

In the regions where prisons aren't unionized, (1) holds true and in the regions that are unionized (2) holds true. Either way, I don't see much of a decrease in union power or in lobbying efforts. To be fair, i don't see an increase either. It's possible that 50 years from now unions will have lost most of their power and membership through a gradual process, but i think its equally likely when predicting that far out that there might be a resurgence in unionization and their power.


I disagree with your analysis of the free rider problem. I think it is likely that both the private companies and the unions will lobby for harsher sentences. They both have a common goal in increasing the number of people incarcerated and both stand to directly gain from it. This is in contrast to the other industries you mentioned, which would only indirectly profit from an increase in the number of prisoners. For example, I would not expect a construction company to lobby for increased sentences because they derive so little revenue from building prisons. Their lobbying money would be better spent trying to decrease environmental protection so that they can build new developments on highly desired and profitable new land. Or any changing or enacting laws that have a much greater direct benefit for them. It is also not unheard for unions and private industries to help each other. If I remember correctly, it happened with the bailouts of the airline industries. On a much smaller scale it happened with the building of a new hospital at UCLA.

While it possible that they both would think that the other will do it, this situation will only be a short term one. Eventually one party will figure out the other isn't planning on doing and will then begin to lobby or attempt to form a partnership. I doubt either side will leave money on the table in the long term.
3.29.2007 12:14am
Sasha Volokh (mail) (www):
Brian K: A lot of (perhaps most?) prison privatization is not the private sector taking over an existing public prison. Rather -- because the prison population is growing -- it's the private sector building a whole nother prison and bringing in all their own employees. Public prison guards are unionized as part of a public employee union -- sometimes they're actually part of AFSCME, the general public employee union -- so there's no reason why the newly hired guards at a newly built private prison would be unionized. So the issue of prison guards unions getting concessions during the privatization process doesn't seem to be empirically significant.

On the free-riding point, Part II of my paper goes into this in depth. Even if two people stand to gain directly from a policy and their goals are totally aligned, their contributions to the goal will still fall (relative to the one-person case) because part of each one's contributions benefit the other. This is standard in the theory of public goods.

Now if you break up a single actor into two, but then the two parts collude with each other, then that can just reestablish the old scenario. But, as I explain later in my paper, here, even under collusion advocacy might fall, because the private sector in fact gets less benefits from prison provision than the public-sector actors do. So the total industry benefit is less than under full government provision, so even full collusion can't reestablish the old level of benefits.
3.29.2007 10:36am