Despite rumors, slogans, and “new” conventional wisdom to the contrary (See, e.g. here, here and the Obama campaign promise to “reinvigorate merger enforcement), it is apparently not the case that the current DOJ is not interested in enforcing the antitrust laws. Perhaps it never was. This interesting interview (HT: Danny Sokol) suggests that the DOJ pressured Google and Yahoo to abandon their deal with a threat to file in federal court to enjoin the transaction. Sandy Litvack (Hogan & Hartson), whom the DOJ hired to prepare to litigate the case if necessary reveals that “We told them we were going to file the complaint at that time of day. Three hours before, they told us they were abandoning the agreement.” Of course, critics of the Bush Administrations’ antitrust policies can say that this is but one instance of aggressive antitrust enforcement — albeit a big one — which should be judged alongside refusals to pursue cases against the Whirlpool/Maytag and Sirius/XM mergers. But we’re dealing with small numbers and anecdotes here which are a woefully insufficient basis from which to reject the alternative — but less newsworthy — hypothesis that the DOJ pursues cases that they believe violate the antitrust laws and harm consumers but does not otherwise.
By the way, do you think the fact that the Google-Yahoo “almost” complaint included a Section 2 claim mean antitrust commentators and critics will count this a monopolization case?
Thanks again for the thoughtful comment. You wrote:
“In the absence of better evidence … isn’t the burden of proof on those like the current DOJ who advocate deviation from the bipartisan approach (e.g. embraced by the current FTC and past DOJ) of the past twenty-five years to justify a reluctance to enforce the prohibition on monopolization?”
No, its not. Not unless we ignore everything we know about error-costs and antitrust.
Your claim is that the default position, without more or better evidence, should be that more is better. To be fair to your position, this is only true so far as we know when “more” means something like a case per year.
We appear to now agree that there is not any evidence to support the notion that this level of activity is good for consumers, other than that it persisted for some time. I’m sorry. But I think that is rather weak evidence of that current levels of enforcement are optimal. 1960s style Warren Court antitrust was bipartisan too, and it persisted for some time as well. So I’m not quite sure why one would think that this is a good equilibrium and deviation from it is bad.
The good news about the integration of economic science and empiricism into antitrust is that we can resolve these disputes with data and evidence rather than assertion. If there were evidence that the “marginal” monopolization or merger cases currently was anticompetitive, I would support expansion of the antitrust laws. I haven’t seen that evidence.
That takes us back to your point — what’s the right default in the absence of better evidence? The question here, oversimplified, is what we do we when we don’t know what were doing in terms of the impact of monopolization cases on consumers?
Your answer is do more. Mine is do less. I don’t think there is any theoretical or empirical reason to believe the right default is a case per year and not significantly less. The theoretical underpinnings of my conclusion are the error-cost approach embraced by many modern economists for understanding antitrust rules and enforcement. To the extent that false positives are more costly than false negatives in the monopolization (and other contexts), if we are unsure — and we are — we ought to err on the side of caution unless we know quite a bit about how to identify monopolization cases that are highly likely to harm consumers — btw, the state of the empirical vertical exclusion literature on this point is not hopeful for your position.
Let me close with what I hope is a clarifying point. There are two issues I’ve tried to separate out along the way. The first is whether the optimal level of monopolization (or other) enforcement. The second is whether the attacks on the DOJ are sensible from a consumer welfare perspective.
You might think that the DOJ is below the optimal level on question 1, but it seems that if you agree that we just don’t have evidence on how good or bad the case per year approach is for consumers (other than looking busy) and don’t have confidence we are doing more good than harm, then one also must think that the attacks on the DOJ are overstated. You might have a prior that is consistent with those attacks, e.g. that the DOJ isn’t doing enough and one case per year is indeed good for consumers. But from an empirical perspective, there just isn’t much support for that conclusion (not that the hypothesis can be rejected either of course). My initial point was that the rhetoric was overblown and we don’t have any proof that is what is going on.
I respect those whose priors differ than mine on the answer to the first question (and thus the second), but I also respectfully submit that this is an empirical question, not a political one, and one that should be resolved as such.
Whether the current DOJ should be counted as having brought no monopolization cases in eight years or one, that rate is less than that of the current FTC and the Clinton DOJ. The former brought more than one a year; the latter nearly so. Prof. Wright isn’t sure the difference is meaningful, but monopolization cases aren’t like mergers, which come across the transom in HSR filings. They take time and effort to investigate, and at the current FTC and Clinton DOJ, there were doubtless many investigations that did not turn up a violation. Based on these statistics alone, it is obvious that the current DOJ is less interested in pursing monopolization cases than the current FTC and the Clinton DOJ. That interpretation is confirmed by the speeches of the DOJ leadership and their policy statements, which the current FTC majority couldn’t stomach. It is hard to depict the current FTC as a rogue agency given that most of the Commissioners where appointed by the current President. It is patently obvious that it is DOJ during the current administration that is out of the mainstream on monopolization (mergers too, but that is off topic except insofar as the President-Elect’s statement that Prof. Wright quotes is based on the full range of non-cartel cases ).
Chairman Kovacic is right to emphasize that enforcement agencies can advance the cause of competition by doing more than bringing cases, and that any evaluation of agency success limited to a review of enforcement actions misses part of the story. His agency, the FTC, has a strong record of undertaking non-case efforts to promote competition. But it is striking that no one defends the non-enforcement record of DOJ during the current administration on the ground that DOJ is doing more for competition by devoting resources that might otherwise have gone to enforcement to other activities.
Both Prof. Wright and Prof. Manne make two other, related arguments. One is that the simple counts of enforcement actions don’t control for case quality. The other is that even if DOJ is in a different place from the past and from the FTC, perhaps DOJ is in the right place and the FTC and Clinton administration is in the wrong place. In the absence of better evidence (we would all like to see many more careful retrospectives that account for case quality) or some compelling recent developments in the economy, isn’t the burden of proof on those like the current DOJ who advocate deviation from the bipartisan approach (e.g. embraced by the current FTC and past DOJ) of the past twenty-five years to justify a reluctance to enforce the prohibition on monopolization?
Thurman, your comment about whether the evidence of smaller number of monopolization actions is sufficient to simply show that the DOJ is “reluctant” to bring monopolization cases moves the goalposts and is a significantly less bold claim than the one you started with describing the DOJ as a .200 hitter and no Babe Ruth. I think its fair to read that statement of an endorsement of the proposition that more is better and that we can draw inferences from activity level to quality of enforcement or gains for consumers or some such metric of good output. It is that position which I think is incorrect.
To the extent that the claim not that, but merely that from the fact that 1 < 7 we can infer that the DOJ brings fewer monopolization cases than previous administrations … we'll, I can do nothing but agree with the mathematics. But this brings us full circle to my earlier declaration that 1<7, without knowing more about the welfare effects of the 1 and the 7, does nothing to reject the alternative "hypothesis that the DOJ pursues cases that they believe violate the antitrust laws and harm consumers but does not otherwise."
Two quick points.
The first is that the difference between zero and one is, I would think, quite significant in this context where an administration that brought 3 significant monopolization cases would, I think, be considered as quite active. Notice again that I’d like to distinguish between activity levels and quality of enforcement for all the reasons described above, in separate posts, and that Geoff reiterates.
The second is that while you claim that nobody is *really* arguing that the DOJ has literally zero interest in monopolization or mergers, I think its actually WORSE than that. Indeed, many have argued that the DOJ wants to make monopolization enforcement more difficult than the status quo, see e.g. the FTC/DOJ scuffle. And this isn’t just the NY Times (though its them too). To quote the FTC Majority statement: the DOJ wants to “place a thumb on the scales in favor of firms with monopoly or near-monopoly power.”
And here’s a quote from the NY Times:
“in light of the decision and the flurry of corporate dealers, there are could really be any mergers that this administration would challenge.”
Here’s Jon Baker in the TNR piece linked above:
“Justice has brought at most one monopolization case during the current administration, compared to at least seven during the Clinton administration. Justice has even sought to make bringing such cases harder, by calling on the Supreme Court to adopt legal rules that privilege dominant firms in monopoly cases … with luck, the damage will be minimal, and we can rid the Justice Department of the deregulatory radicalism that allows monopolies to spin out of control if a new administration rolls into town in January.”
Here’s President Elect Obama:
“Regrettably, the current administration has what may be the weakest record of antitrust enforcement of any administration in the last half century … in seven years, the Bush Justice Department has not brought a single monopolization case.”
To the extent that the claim is not a claim about the EFFECT of the bringing or not bringing of cases, but an inference about intentions, then I happily agree: the Bush DOJ seems “reluctant to bring monopolization cases.” But the relative level of enforcement does not, on its own, tell us anything about the relative value of enforcement practices under different administrations–this is the inference Josh and I are endeavoring to combat.
To your first point, I think your re-statement of my position is a bit strong. I do not claim that absent full analysis of all cases brought, etc., one cannot in any way assess the likely propriety of the DOJ’s enforcement history. Rather, my claim is that a) this sort of analysis would be helpful and, more important, b) those who criticize the Bush DOJ on the basis of the number of cases brought do not, even in their casual analysis, incorporate sufficient (any?) assessment of the costs to social welfare of over- and misguided-enforcement. There is, instead, a presumption that more enforcement is better, at least up to a point (for many, a point we have never reached) and that the low level of enforcement in the Bush DOJ is conclusive evidence of damaging under-enforcement. This assessment, however, presupposes an optimal, and it is the claim to know this optimal level that I question, particularly when those claiming to know it seem to think it likely lies at the end point on the far right. I will admit that my own inclination places the optimal to the left of center, but I take seriously the institutional problems that I have never seen most critics discuss at all (except, I guess, to decry the current administration’s ideological reluctance to bring cases).
Prof. Wright: I don’t think anyone has argued that DOJ today has literally no interest in enforcing the antitrust laws. You’ve set up a straw man if that’s all you aim to disprove. The question is whether we have under-enforcement today, not whether we have zero enforcement.
Prof. Manne: It sounds like you are saying that it is impossible to evaluate DOJ’s record without a full analysis of cases actually brought and possible enforcement actions not brought. That standard would place a very high burden on the critics of any administration. Do you think that it is no more possible to criticize DOJ during the Clinton administration for over-enforcement (as I suspect you might want to do) than to criticize the current DOJ for under-enforcement?
If you don’t want to accept that implication of what appears to be your position, and want to criticize, say, case selection during the Clinton administration, then don’t you have to accept that we should do the best we can to evaluate the enforcement record with the evidence we’ve got? Part of that evidence is that DOJ has brought no monopolization cases over the past eight years, while earlier administrations at DOJ and the current FTC both found monopolization cases they thought were worth bringing. Why isn’t that statistic enough to infer that the current DOJ is reluctant to bring monopolization cases, even if it they might have been willing to challenge Google recently?
Josh’s last point can’t be stressed enough. It is beyond meaningless to point to the number of cases brought and to infer anything from that number about the extent to which the DOJ’s enforcement decisions help or hurt consumers. And yet there are plenty of seemingly-reasonable people out there who equate a high number of cases with effective enforcement, full stop. It’s nonsensical, as Josh says–and yet it’s de rigeur. Oh, but it’s explainable. You see, those people tend to believe, explicitly or implicitly, that the risk and cost of over- and misguided-enforcement is low, while the cost of under-enforcement is very high. I happen to believe precisely the opposite, but either way, we needn’t rely on such crude measures–we should talk about actual error costs, actual cases and actual results. Do the assessment Josh suggests. Be comprehensive in our understanding of institutions and political incentives, as well as our information limitations. Until the folks tarring the DOJ with harming the consumers for not bringing enough cases demonstrate any appreciation for these issues, there is no reason to take them seriously.
Thurman (great handle, btw…), thanks for the comment. The passage you quote, the post in general, and a series of posts I’ve done on this topic makes pretty clear that I’m responding to the series of allegations that the DOJ just isn’t interested in enforcing the antitrust laws — and especially monopolization cases. There is nothing inconsistent with saying that the information that they brought the case is sufficient to demonstrate that the DOJ is interested in enforcing the antitrust laws — but not sufficient to figure out whether the enforcement action is a good idea. If I didn’t make that clear in the post, I hope it is now.
To your question, I think more work needs to be done on figuring out the report card grade of the current DOJ. Namely, we need to know something about the competitive effects of “marginal” actions, e.g. enforcement actions that were close to the line that they didn’t bring or did and lost (e.g. Oracle, Sirius, Whirpool). I think retrospectives could be valuable in providing information along those lines.
All of this is consistent with the basic point enforcement agency activity levels and win rates simply are not sufficient to figure out whether our enforcers are doing a good job in terms of protecting consumer welfare.
Prof. Wright: If it is “nonsensical” to rely on “small numbers and no information about the competitive effects of the transactions at issue” to judge enforcement agencies, what other information are you relying on when you say “it is apparently not the case that the current DOJ is not interested in enforcing the antitrust laws. Perhaps it never was.” If those sentences aren’t implying that you see the current DOJ as an all-star performer, what are they saying? And what about the larger question your post flirts with: would you defend the record of DOJ during the current administration as strong on antitrust enforcement beyond cartels?
One more point. I should be clear that I’m not defending or attacking the DOJ’s action here. I don’t know enough about the transaction or have the data that they had to make the decision. That G/Y abandoned the transaction might be good or bad for consumers. I just don’t know. As the comment above indicates, whether this is a strike out or a home run, that’s not the point of the post.
Thurman,I would have thought that enforcement against G/Y but not Sirius or Whirlpool would be a .333 average —- no? I kid. But the larger point is that relying on these small numbers and no information about the competitive effects of the transactions at issues is just a nonsensical way about going about judging antitrust enforcement agencies.
A .200 hitter hits one home run and Prof. Wright says that proves he’s been Babe Ruth all along. I don’t think so.