So-called “revolving doors” between government and the private sector raise the specter of potential corruption (if not in the strict legal sense, then in the broader sense), and some anticorruption advocates have called for much more aggressive restrictions on former government officials’ ability to work for the sectors they used to regulate. (See, for example, here, here, and here.) Though the concerns are legitimate, I argued in a post a little while back that the issue is much more complex: many of the concerns about the harms of the revolving door may be overblown, and revolving doors might in some cases have beneficial effects.
I thought I’d revisit the issue in light of two very interesting recent contributions on this topic: a blog post last week by Transparency International Programme Manager Dieter Zinnbauer on the pros and cons of the revolving door (along with a companion post on measurement issues), and an article by Wharton School Professor David Zaring. Mr. Zinnbauer concludes that the weight of the evidence suggests that the revolving door is indeed a serious problem, and that for the most part the costs outweigh the benefits; Professor Zaring reaches more or less the opposite conclusion.
Although I think the first half of Mr. Zinnbauer’s post is an excellent, succinct, evenhanded summary of the main issues, I respectfully disagree with the inferences that he draws from the existing evidence. That’s not to say that his conclusions are wrong, or that revolving doors are nothing to worry about. But when Mr. Zinnbauer says that “a much larger body of new evidence comes down quite distinctively on the negative impact of the revolving door,” I think he’s overstating his case. Here’s why:
First, let’s take a step back and clarify what, exactly, we’re worried about, and therefore what we’re looking for in the data. Though there are many potential concerns with the revolving door, the big one is that government officials’ past or potential future private sector employment will distort their decision-making in ways that are harmful to the public interest. The usual version of this concern emphasizes the possibility that public sector employees’ desire for future private employment will cause them to go easy on those potential employers. That’s why some suggest the revolving door encourages a kind of “corruption” — even if not illegal, and even if not conscious, when a regulator does something that’s not in the public interest, in order to help the regulator’s own future job prospects, the regulator has abused her public power for private gain. Another version of the revolving door concern is that regulators who come from the regulated industry will have a natural tendency to identify and sympathize with their former employers; this version of the problem doesn’t so much involve the corrupting interest in future employment, but rather a form of what’s sometimes called “epistemic capture” — a fancy term for the idea that if you’ve spent most of your life hanging out with bankers (or whatever), and being one yourself, you start seeing things from their point of view.
The big question is the extent to which this possible concern is, in fact, a serious problem in the real world. To Mr. Zinnbauer’s credit, he points out that the answer is not obvious, and that many commentators have pointed out reasons why the revolving door might not only not be bad, but might actually be good: First, public officials might improve their job prospects more not by going easy on the regulated industry, but by being quite aggressive — and highly competent — so that firms think that this is the kind of person they’ll want on their side in future interactions with the government (what Mr. Zinnbauer calls a “signaling competence effect”). Second, the prospect of future private employment makes the relatively lower salaries in most public service jobs less of a disincentive for the most talented people to take those jobs, at least for a while (a “talent attraction” effect). (It’s also possible that regular movement of individuals back and forth will lead to better communication between the public and private sectors and a better understanding of the issues in the industry, though these are less directly relevant to the question of whether the revolving door might distort government decisions.)
Mr. Zinnbauer, as noted above, acknowledges this complexity but says that the weight of the evidence supports the conclusion that the downsides of the revolving door — the potential for corruption — outweigh the good. But the evidence he offers doesn’t really seem to support that conclusion:
- Most of the evidence cited indicates that firms that hire lobbyists with government experience receive benefits — faster approvals, more funding, more favorable results, and a higher stock price. That’s plausible, but it is not evidence that the officials going through the revolving door made worse decisions while in government. Now, it does suggest that firms that hire such officials do better than firms that don’t, and perhaps that’s troubling for other reasons. But it’s not clear exactly what the problem is: Unless we know the right baseline, we can’t say whether the problem is that firms with revolving door lobbyists get (undeserved) special benefits, or rather whether firms that fail to hire such lobbyists are failing to get what they deserve. Another difficulty is the fact that hiring revolving-door lobbyists is not random, and so it’s hard to tell whether it’s the fact that these firms hired such lobbyists that leads to the better outcomes, or some other factor about the firms that tend to hire them.
- Mr. Zinnbauer also points out that revolving door lobbyists tend to work on a broader range of issues than other lobbyists, indicating their value is in their contacts rather than their expertise. But, first, that still doesn’t show that these lobbyists’ behavior in government was distorted, and, second, it may just mean that lobbyists without government experience need to specialize to make up for the fact that they don’t have as much general understanding of how government works.
- Perhaps the most compelling evidence that suggests some distortion effect is a working paper that does provide some fairly strong evidence that analysts at credit rating agencies (not technically government employers, but performing a public function) give more favorable ratings to financial institutions when they are about to switch to managerial jobs at those institutions. That result is indeed troubling, and consistent with the negative assessment of the revolving door. But it’s not clear that the result is generalizable.
And that last point is where Professor Zaring’s article comes in. In addition to discussing many of the same criticisms of the anti-revolving-door position (in much greater depth), he reports a study of 152 prosecutors who were in the criminal division of the US attorney’s office in the Southern District of New York in 2001. At the risk of oversimplifying Professor Zaring’s nuanced findings, the headline result is that there was no statistically significant correlation between a prosecutor’s aggressiveness (measured by number of cases brought, prison sentences imposed, publicity) and that prosecutor’s future private sector employment.
All this leads me to the somewhat unsatisfying conclusion that we still don’t know all that much empirically about how much the revolving door phenomenon corrupts government decision-making. Mr. Zinnbauer has done the anticorruption community a great service in providing such a lucid overview of the debate, and he is admirably careful to qualify his arguments with all of the appropriate disclaimers. Nonetheless,I would push back a bit on his main conclusions as to the weight and strength of the existing evidence on this issue.
GAB | The Global Anticorruption Blog 
A possible counterpoint to the first response–that former government employees are often successful in securing better treatment for their private employers: Assume first that the new advantages are not merit-based (stronger contract proposals, better knowledge of internal government process). (That’s a big assumption, and I think it is not warranted in lots of cases. But for the sake of argument …) If so, then the benefits flow from personal ties developed while in public office. If that’s the case, then the revolving door arguably is encouraging corruption by creating a thick web of interpersonal relationships between current and former government employees working in the same field.
In other words, the revolving door is an indirect problem–and you could imagine that though a former public employee didn’t make worse decision while in government, his close professional relationships could encourage other government employees to do so once he has left.
In that case, though, the problem would be the closeness of certain professional relationships, rather than individualized incentives to cut regulated industries slack in order to secure a lucrative position later on. And I’d be inclined to think that the benefits of moving expertise between the public and private sector would outweigh the potential risk that some professional connections would improperly influence public servants.
I agree entirely. Though I touched on this briefly in the post, I probably should have emphasized it a bit more. Even if public officials are not distorting their decision-making in order to improve their employment prospects (the most frequent version of the revolving door concern), the fact that former public servants make for effective lobbyists–and may be able to use their connections and understanding of how to work the system in order to get special (undeserved) benefits for their clients–is _also_ cause for concern. (And there may be a hybrid of the two concerns: If X and Y both work at a government agency, and X goes off to work as a private sector lobbyist, X and Y’s personal relationship might lead Y to behave more favorably to X.)
But — and I think we both agree on this point — that legitimate concern is premised on the assumption that the benefits that former public officials can get for their private clients are not, in fact deserved. But that’s not always clear. Suppose a private client hires an ex-DOJ prosecutor as defense counsel, because the client expects that an ex-DOJ official will be able to negotiate a more favorable deal (because she understands DOJ’s priorities and concerns, because current DOJ staff know her and think of her as a reasonable person, etc.). Maybe that’s bad, but maybe that’s good — we can’t really tell unless we have some sense of what the right outcome should be.
(The inequality issue — that those firms that can afford to hire ex-government lawyers or lobbyists will do better than those that don’t — is another independent concern, but one that takes us a bit far afield from worries about “corruption,” even under an expansive understanding of that term.)
Matthew, I’d like to pick up on your “right outcome” point as it pertains to epistemic capture. Although I am wary of capture, there is a lot to be said for hiring a former banker – who thinks like a banker and knows a lot of other bankers – as a banking regulator. Even _if_ that regulator goes easy compared to the current baseline, the result may be the optimal one.
I can image that public servants who only ever regulate/prosecute implicitly view their issue area as the be-all and end-all. By contrast, businesses deal with, say, banking requirements as one set of regulations among many, all of which are subordinate to making profits. This schism could result in over-enforcement, which might chill business activity and discourage voluntary cooperation and information-sharing on behalf of companies.
Rather than going in with a saw, regulators with experience in the industry can use a knife. They can demand efficient penalties that achieve optimal deterrence and can structure settlements to focus on areas that need the most work.
I have to think more about whether I support this view. It rests on a couple of challenging assumptions, namely (1) that regulators/prosecutors who don’t come from the private sector over-enforce, (2) that the regulator/prosecutor who does come from the private sector thinks enough like a banker but not too much like one, and (3) that the approach would be systematic, an unlikely proposition.
All good points. Perhaps one thing your observations suggest is that diversity in government agencies may be desirable: you don’t want all your banking regulators to be former (and/or future) bankers, but you may well want _some_ of them to be. The devil, though, is always in the details.
One question I have is whether different types of private sector jobs 1) invite more of the problems associated with the revolving door (whether conscious or unconscious); and 2) have particular attributes of which we should be more wary. It seems that the general idea of assessing the costs and benefits of the revolving door between the “private sector” and “government” at large is too much of an abstraction, and that some lessons might be gleaned by looking at certain types of positions in the private or public sectors.
Perhaps, for example, the organizational structure of the U.S. Attorney’s Office in SDNY could account for the lack of correlation Professor Zaring points out. While prosecutors execute a great deal of discretion, major strategy decisions regarding aggressiveness, plea bargaining, etc., generally do go up some chain of command, increasing the amount of oversight in those areas. When compared to the oversight involved with lobbying or credit rating agencies, are there structural arrangements that allow for, or even encourage improper behavior (perhaps due to heightened discretion or focus on face time)?
Very nice point. Another aspect of this might be whether, as a public sector employee, you improve your later private sector job prospects by (A) demonstrating aggressiveness, tenacity, expertise, etc., or (B) demonstrating sympathy with the industry’s interests, showing that one is reasonable, has values that align with those of the regulated sector, etc. The relative importance of these factors might differ across job types and industry types. I’d conjecture that for a government lawyer, you increase rather than decrease your value by winning a lot, getting big penalties, being a tough negotiator, etc. Maybe if you’re at a credit rating agency things are different.
This is somewhat similar to your point, but a bit different too, in that I’m suggesting prosecutors might be different not because they’re subject to so much internal review and control, but because for a prosecutor being tough doesn’t diminish, and may enhance, perceived value to the private sector.
Dear Matthew (if I may),
Thanks so much for your really thoughtful comments that help to greatly advance a debate that as we probably both agree all too often gets stuck in polarized and rather unproductive exchange of speculative opinions. Here some rather spontaneous reactions, all in the spirit of jointly further unpacking the issues at hand.
Let me focus on the point that I felt raised the most serious questions about my argument: the missing baseline, which is really a very important one. Yes, superior treatment when hiring former government insiders might in some cases be welfare-enhancing when it is about cutting more efficiently through red tape, or about getting what one should be entitled to in the first place. Yet, I would suggest that many contexts, including several of the ones that are being discussed in the empirical studies at hand, are not so much about inefficiencies but more on the zero some side: either directly adversarial (when competing interests are adjudicated about, e.g. should company x or company y (or the public) shoulder the consequences of some risky deal gone wrong, or at least with some characteristics of rivalry or competition (should limited public funding available for universities go to university x instead of university y or does expedited treatment for company y when it comes to specific approval processes constitute a competitive advantage over some corporate rival). In such situations the lack of a baseline is not so material, since relative advantage is what counts and that is what those studies clearly established. Now, if one wanted to follow this line of reasoning the issue about who can or cannot afford revolving door lobbyists becomes actually pivotal, since it turns the quest for edge from an annoying collective bad (each party locked into the competition grudgingly paying up for expensive revolvers to level the playing field) into a structural disadvantage. And I think the empirical picture here is troubling: Cain and Drutman (2014), for example, show in their analysis of the future career trajectories of US House of Representative staffers that overwhelming majorities end up with (typically expensive and thus difficult to afford to non-profits) lobbying firms, corporations or business and trade associations, while less than 10% become advocates for non-profit cases and no one was found to work for unions. The more senior the staff members, the more pronounced this effect with close to 90% of chiefs of staff pursuing the former career track, while literally none is being employed by non-profits or unions. These observed structural asymmetries in the deployment of revolving door lobbyists is closely mirrored by spending asymmetries. As de Figueiredo and Richter (2014) find in their review of a larger body of empirical work in this area, corporations and trade associations account for well beyond 80% of total lobbying spend both at national and local level. And their review also shows that such troublesome asymmetries also exist within the business community with larger corporations far outspending smaller ones.
So against this backdrop, the revolving door looks to me like a catalyst or aggravating factor for structural asymmetries in access with the risk of significant adverse consequences – even in the absence of a proper baseline. In my view this warrants a) more effort in actually monitoring the phenomenon in key areas of policy-making or regulation (thus my follow up posts on measurement approaches b) serious efforts to diffuse the problems associated with the phenomenon by tightening regulation and stepping up enforcement. Let’s bear in mind, no one subscribes to the unrealistic ambition to abolish revolving door practices. The focus on our advocacy is on reasonable efforts to take the sting out of it and prevent the all-too-direct and seamless switching in and out through longer cooling-off periods, better enforcement and the closing of loopholes. A continuous stream of well-documented case studies (see for example today’s NYT on the amazing lobbying of state attorney generals with the revolving door as a central catalyst http://nyti.ms/1wbLOoc) in conjunction with the types of bigger picture empirical studies that I reviewed provide in my view a more than sufficient evidence base and red-flag situation for taking such action – and do so in careful, context-sensitive and targeted manner – even if all-around conclusive evidence is still (and perhaps permanently) elusive.
Best wishes from Berlin
dieter
Dieter,
Thanks very much for this thoughtful, persuasive response. While I can’t go into much depth on all the points you raise right now, one of the things that’s really useful about your response and elaboration is how it highlights that there are several different — possibly related but analytically distinct — concerns we might have about the “revolving door”, and we need to be careful not to conflate them. In particular, the concern about structural inequality — the firms that can hire former government officials are treated better than firms that can’t — is distinct from the concern about government officials deliberately “going easy” on past or potential future employers. I’d assumed that your main concern — like that of much of the existing discussion — was on the latter problem, but now I think I understand more clearly that you’re actually placing more emphasis on the former problem. Fair enough. The baseline question is still an issue, though I accept that in some cases the inequality is itself a sufficiently significant concern that the baseline concern is less relevant (or, to put this anther way, in some cases the baseline should be “like cases treated alike”, so that deviation is prima facie undesirable). Of course, we’d need to carefully identify those cases in order to tailor an appropriate response.
One additional question this does raise: How much of your concern is really about the revolving door as such, and how much of your concern is about the inequality in lobbying power across firms or groups? The revolving door may contribute to the latter problem, but it would presumably only be one factor among many. Even if no firms ever hired former DOJ lawyers or former SEC regulators, those firms might still have a big structural advantage over other firms, and still be able to hire better lawyers and regulators. And if the problem is lobbying inequality, is the revolving door the right focus, or is it more like a peripheral aspect of a broader and much more deeply entrenched problem?
Dear Matthew, once more thanks for your great comments. My concern is both with revolving in and out and I opted to focus on the baseline issue, since I felt it poked the biggest hole in my reasoning. And yes, your questions at the end are spot on; the revolving door is certainly one issue in the broader setting of protecting against undue influence, a significant stand-alone problem at times, but ultimately more consequential as an amplifier of broader structural problems. As such it is one of the issues that we and some of our country groups here at TI are working on as part of a comprehensive policy package that includes efforts such as making political financing more transparent, prod the lobbying industry to make their codes of conduct stricter, regulators to close related loopholes, corporate actors to be more forthcoming about their policy engagement etc. Related scholarly research and the type of critically-constructive discussion that you regularly “instigate” on your blog are really crucial for helping us sharpen our tools in this ambitious quest. So I am very much looking forward to more such discussions in the future.
All the best,
dieter
Matthew, you mentioned the potential that the revolving door has to attract talent to public service by doing something to close the wage gap between public and private career trajectories. I’d like to highlight just how important this consideration is. Policy makers and regulators are paid a middle class salary, but it pales in comparison to what many of those people could be making in the private sector. Most senior level public officials make significantly less than a first year associate at a large law firm. Even when someone does work in the public sector and then ‘cash out’ through the revolving door, their career earnings are likely to be significantly less than if they had spent the whole time in the private sector. Many of my friends only see public sector work as a viable option for them if they know there is the potential to exit to higher salaries; otherwise they would never start there in the first place.
Because even those who plan to make use of the revolving door are intentionally setting themselves up for lower career earnings, the vast majority of them must care about public service. Therefore, perhaps we could eliminate many of the problems of the revolving door by paying policy makers and regulators significantly more. Even if we couldn’t match the public sector, reducing the gap would reduce the incentive to leave the public sector for those who are publicly minded.
I am very much in sympathy with your argument here, and I’ve heard much the same thing from friends and former students: The wage gap between the public and private sectors–at least for very talented individuals–is enormous in many countries–perhaps especially in the United States–and this is a significant problem in attracting and retaining top talent to public service. The “revolving door” ameliorates this problem somewhat (as Dieter also noted in his original post), but the revolving door may also create a number of other serious concerns. (Again, while I think the evidence is more ambiguous than Dieter implies, I don’t dispute that there’s a genuine reason for concern.)
You propose what I think of as the “Singapore Solution” — pay senior public officials very high salaries — comparable, and perhaps even higher than, what they could make in the private sector. This might not only reduce the temptation to take bribes or engage in other forms of corruption (the usual argument in favor of Singapore’s approach) but also means that there may be much less of a revolving door. (Here the Singapore example may start to become inapt, though, because there are such close ties between Singaporean business and government elites that it’s hard to say the high salaries are really ensuring that public officials are not excessively sympathetic to private sector interest groups.)
I like the idea. If I could wave a magic wand, I’d be tempted to combine much stricter limits on the revolving door with substantially higher salaries for public servants. The problem is feasibility. At least in the United States, there seems to be so much knee-jerk opposition to spending more money on the public sector that it’s hard for me to imagine a proposal to, say, triple the salaries of a large class of government officials going anywhere. But I do think you’re on to something, and perhaps the activists and advocates who criticize the revolving door might do well to combine those criticisms with a call for substantially higher public sector salaries. In addition to addressing the talent-attraction concern, it would also make the anti-revolving-door position less like an attack on public servants (and former public servants), which, from a political/tactical standpoint, might also be desirable.
The revolving door phenomenon may be keeping government salaries low. The drivers of policy in the American federal government are almost all political appointees. They all leave government service for the private sector where they pay and benefits are generally equal (Washington think-tanks) to significantly higher than government. So why would they push for better salaries for career employees?
The phenomenon of appointees actually brings up another revolving door problem, too: it’s nearly impossible for people to stay in high level appointed government positions when their party is not in power. Therefore, many public-service inclined people attempting to sustain a high-level career must find private sector options to have as a back up when the other guys win, even if they plan to eventually ‘revolve back’ into public service.
Unfortunately, even if we suddenly had higher government salaries it wouldn’t fix that problem. There’s an issue both of Administrations being understandably disinclined to appoint those who disagree with them ideologically as well as individual
reluctance to work in an administration whose policies you disagree with. Perhaps the only way to tackle this would be to significantly raise the salaries for other not-in-government options to make them more attractive as a place to bide time until your party wins again (think tanks, academia etc…). This may be an even thornier problem than raising government salaries, as most of those alternative employers are non-profits without the funding to raise salaries like that. But as long as we’re waving our magic wands…