Armentano in the WSJ, Abolition and Antitrust Fairy Tales …

Josh Wright —  28 December 2009

Leading antitrust critic and abolitionist, Dominick Armentano, has a letter to the editor in the WSJ.  The point of the letter to the editor is rather specific: that FTC’s attack on Intel is no outlier in the historical context of antitrust enforcement, contrary to the WSJ’s description.  To the contrary, Armentano argues that Intel is just another in a long line of misguided enforcement actions.   Here’s the letter:

Your editorial is correct to condemn the Federal Trade Commission’s attack on Intel (“The 100 Years Chip War,” Dec. 18), but it is dead wrong to conclude that the government’s antitrust intervention is “unprecedented” or that antitrust laws really “exist to promote business and price competition.”

Have we forgotten the FTC’s eight-year (1958-1966) campaign against the Borden Co. to stamp out lower prices for evaporated milk? Or its 10-year (1957-1967) legal assault to end the Procter & Gample-Clorox merger in which the FTC’s primary argument against the consolidation was that the probable “economies and efficiencies” of the merger could be passed along to consumers?  Or how about the Justice Department’s 15-year (1953-1968) war against United Shoe Machinery in which United was ultimately ordered to create a competitor with divested shoe machinery assets, license out all of its own patents to the competitor, and then refrain from active competition with the new-born company for five years?

And have we already forgotten that the Microsoft antitrust debacle started with a two-year investigation by the FTC back in 1990 or that the Justice Department pursued the company for another 10 years because Microsoft bundled its Web browser, Explorer, with its Windows operating system, much to the delight of willing buyers. Recall that in the 1999 trial verdict, lower court Judge Thomas Penfield Jackson even ordered the company divested until the D.C. Circuit Court of Appeals unceremoniously discarded that absurdity in 2001. In short, the FTC’s assault on Intel is hardly unprecedented.

What these cases (and hundreds of others) establish beyond any reasonable doubt is that antitrust does not exist to promote business and price competition. Never has, never will. The theoretical and case evidence, some of which I’ve cited, is all the other way.

The real mystery surrounding antitrust is why knowledgeable observers of the free-market process persist in believing this fairy tale.

I’m already on the record as publicly criticizing the FTC’s Intel complaint.  And to the extent the letter makes the general point that the past and present of monopolization enforcement is riddled with false positives and rent-seeking that dissipate any theoretically plausible efficiency gains, I’m on board.  But more generally, I was reminded by the letter of the antitrust abolition argument raised by Armentano and others (generally from Austrian or public choice traditions).  While I’m generally sympathetic to Armentano’s views  in so far as they express skepticism about the welfare benefits of antitrust enforcement, I do not favor the abolition of antitrust and never have.  I should note that I am especially sympathetic to the skeptical view with respect to Section 2 enforcement.  As an antitrust economist who has been highly critical of government intervention in his scholarship — particularly with respect to monopolization rather than cartel and merger enforcement — and who has been described as a “Chicago School apologist” by a sitting Federal Trade Commissioner, I’ve certainly been criticized by those favoring a “reinvigorated” antitrust regime for supporting a reduction in the scope of antitrust laws and a humble and cautious approach to their enforcement.  On the other side, I’ve also frequentlybeen asked why, if I take such a critical view, don’t I support the abolitionist position of Armentano and others who share his views (and criticized by them, see, e.g. the comments to this post)?  Indeed, I might even self-indulgently describe myself as one of the “knowledgeable observers of the free-market process” to whom Armentano ascribes a mysterious and persistent belief in fairy tales.

So why don’t I believe in abolishing antitrust in toto?  The last time the issue came up on the blog was in response to a similar question raised by my George Mason colleague Bryan Caplan (in regard to the new proposed law in Hong Kong).  In that post Bryan asserted:

Even if you’re a mainstream economist who thinks my general critique of antitrust is overblown, you should still grant that for Hong Kong, I’m right. And doesn’t the fact that Hong Kong’s made it this far without antitrust give you a moment’s pause about the domestic benefits of these laws?

My position then is my position now:

Bryan has overestimated the case in favor of abolition, or at least should take a more nuanced stance. In evaluating the social benefits and costs of antitrust enforcement (including rent-seeking, error and administrative costs) I think one really has to distinguish between cartel enforcement, mergers, and monopolization. The evidence that antitrust can generate net benefits in the first category is much stronger than that it is for either mergers or monopolization. Reasonable minds can differ about the state of evidence in those latter categories, as well as whether “real” antitrust enforcement in those categories results in social costs that swamp potential benefits.

Lets just take cartels as an example.  It would be tough to argue, based on the evidence, that there is enough there to support abolition of cartel prosecution.  And cartel prosecution is a substantial part of the modern competition policy landscape.  Nor do I believe that the fact that Hong Kong is a small open economy or that it has gone a long time without antitrust means that cartel prevention is ineffective in the U.S. or cannot be in Hong Kong.  This is not an optimistic or utopian view of antitrust.  I don’t think I’ve ever been accused of that.  I’m written quite skeptically about enforcement in the single firm conduct area and how little we know in these areas should inform our policy.  One can argue that in practice, cartel enforcement really amounts to consumer welfare decreasing activity by overzealous regulators. But thats an empirical question. And I think the evidence pretty strongly suggests that cartel enforcement is good for consumers. The evidence with respect to mergers is a mixed bag and there is no general consensus. The picture is much more bleak with respect to single firm conduct, where not much is known and there is very little empirical evidence to suggest that antitrust enforcement is producing the types of outcomes that would justify the social costs of enforcing and administering those laws.

Bottom line: the position for abolishing antitrust, if we are are basing this on the current state of theory and evidence, is weakest against cartels, uncertain with respect to mergers, and much stronger for single firm conduct.

The interventionists argument that is in theory, since monopolization can result in the same effects as cartels, it doesn’t make sense to prohibit one instead of the other.  Similarly, since a horizontal merger can be a substitute for a cartel agreement, it probably doesn’t make sense to have a cartel prohibition without merger law.  All this is true in theory.  But it does not necessarily follow that these theoretical connections justify adopting the entire antitrust machinery if the welfare losses from merger and monopolization policy exceed the gains from cartel enforcement (including administrative and error costs).  One can argue about the relative magnitudes of those values in theory.  And please note that nothing in such a hypothetical position would require one to believe that anticompetitive conduct doesn’t exist.  But my position is an evidence-based one.   Cartel enforcement, in my view, has largely proven its social value.  But I’m quite skeptical that the technology available to distinguish “single firm” conduct from its anti-competitive counterpart renders Section 2 a consumer-welfare increasing proposition.   In the meantime, in my opinion, the abolitionists’ refusal to confront the qualitative and quantitative evidence supporting the effects of cartel enforcement undermines their case generally, and shifts attention away from the much stronger case against monopolization enforcement.

17 responses to Armentano in the WSJ, Abolition and Antitrust Fairy Tales …

  1. 

    My letter to the WSJ did not speak directly to cartels; nonetheless, let me address several theoretical and empirical issues as raised above.

    Those who favor anti-cartel enforcement always assert that the theoretical case against cartels depends squarely on dead-weight welfare losses, i.e., standard neoclassical arguments. But this follows only if: a. cartel prices are “monopoly” prices, and b. if the “cost curves” and the nature of the “product” remain the same. In a disequilibrium world (the actual business world), however, none of this need be true and, if not, the (negative) welfare implications of the analysis simply DO NOT FOLLOW. And neither does the “logic” of antitrust enforcement.

    First, many of the cartel arrangements that I’ve studied attempt to prevent or remedy sub-competitive prices, that is, prices below long run average cost.(Much of this pricing is the result of the business cycle, another factor usually omitted in any welfare discussion of cartel behavior). It should be apparent that IF this is what cartel pricing does, there is no necessary “dead-weight” welfare loss. Indeed, during a severe downturn, price fixing might be the only way to secure “competitive” prices. (See Addyston Pipe, Trenton Potteries, and many other early cases).

    Second, standard static welfare analysis assumes “costs” are not reduced or restrained and that the nature of the product does not improve. Yet almost any cost reduction or product improvement within the cartel group (in manufacturing, transporting, selling, including advertising) could easily overwhelm any so-called static “welfare losses.” But since so-called naked price fixing is treated generally under a per se rule, little serious empirical work has fleshed out these “benefits.” So to assert that the empirical case against cartels is persuasive is misleading at best.

    Note that none of this is especially “Austrian.” Indeed, it is just standard price theory with an absolute Schumpeterian recognition that the real business world is never in any “equilibrium” and that price adjustments are only one aspect of a competitive market process.

  2. 

    I suspect that there is not much more to be gained in pushing the discussion further, and so I will only answer the question posed to be by Henry Manne in comment 14 and then move on:

    “All of this seems like a very slim empirical base on which to build your case for anti-cartel enforcement laws. Are you really comfortable with it?”

    I should make clear that, while the discussion here as focused largely on the empirics, the case I favor for cartel enforcement is based on BOTH economic theory and empirical evidence. With respect to the latter, it will have to suffice for here to say that I read the empirical literature as providing much more support for anti-cartel laws than you do. One thing that is absent from that literature, however, is evidence of the social benefits from cartels. Nor has that evidence documented the benefits of treating contracts among cartel members as any other contract, invoking the courts to enforce output restrictions and sanction cartel cheating.

    With respect to the theoretical support for a prohibition against price fixing, in addition to the fact that naked cartels appear to behave in practice as they do in the price theory textbook, i.e. restrict output, increase price, create deadweight loss (all with serious social consequences), I find some comfort in staking my position with antitrust hawks like Bork, Posner, Demsetz, Stigler, Ginsburg, Epstein, and even Hayek, who wrote that in spite of some reservations, in support of “a general prohibition of all cartels, if it were consistently carried through” … and that the state “declare invalid and legally unenforceable all agreements in restraint of trade, without any exceptions, and to prevent all attempts to enforce them.. by giving [injured parties] a claim for multiple damages.”

    Similarly, despite the imperfections of cartel enforcement specifically, for the reasons stated, I am comfortable with it.

  3. 
    Thurman Arnold 2 January 2010 at 5:49 am

    Good news, Josh — Henry Manne now concedes that cartels exist even outside bid-rigging in public procurement. He no longer is arguing about whether anticompetitive conduct of the sort the antitrust laws prohibit exists (we’ve won on that point), and has shifted to questions of how best to design public institutions to address the problem. That’s a good place for the argument to move. We want to design laws and enforcement approaches to maximize the (net) benefits of antitrust enforcement, recognizing that such rules can create both beneficial and harmful incentives for private behavior. (Even anti-cartel rules, if poorly designed, could deter beneficial conduct like efficiency-enhancing joint ventures – but first and foremost they deter harmful cartels.) The institutional design issues are important — they are why you are more skeptical of monopolization cases than I am, that’s the source of the debate over the benefits and costs of private vs. public enforcement, etc. — but they do not question the legitimacy of the antitrust enterprise, and are proper not to do so.

  4. 

    Well, let’s see. We still do not know the quantitative impact of private vs. public vs implosive cartel busting; we still do not know which mergers are for price fixing reasons and which for other reasons (and how to distinguish them) or what empirical impact this answer would have on the prior debate; we still do not know to what extent real price-fixing cartel agreements are substittues for bankruptcy (or worse, bailouts); and finally we do not know what effect this antitrust enforcement has on entrepreneurial attitudes and preferences for governmental industrial policy. Oh yes, and we do not know that impact of false positive or false negative prosecutions. All of this seems like a very slim empirical base on which to build your case for anti-cartel enforcement laws. Are you really comfortable with it?

    As to your and TA’s empirical point, I undoubtedly overstated the total number of cartels for public bid-rigging, though I am sure that I could come up with alternative explanations to why there have been so many private cartels in recent decades, none of which would implicate antitrust laws, cartel-inducing regulation, and anti-merger and antitakeover laws for instance. But I doubt seriously that the empirical studies you cite can be used to demonstrate the qualtitative difference it would make to your empirical case if, just for the sake of argument, we dropped the bid-rigging cases from the anti-trust sphere and put them in the public corruption sphere where they belong. Surely it would weaken your case considerably, and that is the point I was trying to make. Empirics too are very slippery.

    Finally, Josh, I do start “with the premise that the law has no business in seeking to reduce the deadweight loss associated with price-fixing” as long as I may add a short addendum: “so long as it is not occasioned by fraud, theft, or violation of any other recognized contractual or legal property right.” And please do not tell me that price fixing is theft. Whatever else it may be, it ain’t theft! If you don’t like the way my laissez faire producers behave, then let the government produce your goods and services.

  5. 
    Thurman Arnold 1 January 2010 at 5:26 pm

    Henry Manne claims “You will please notice that most of the cases cited (and most of the long-lived ones) are public bid-rigging cases, not really private cartel cases in the usual sense.” Really? I have copied below a list I found of the industries subject to the largest fines by US or European authorities as of two years ago (measured in nominal dollars, so the examples are recent): Vitamins, Elevators and Escalators, Gas Insulated Switchgear, Synthetic Rubber, DRAM, Plasterboard, Air Transportation, Hydrogen Peroxide and perborate, Methacrylates. I am surprised to learn that these were largely public procurement bid-rigging cases. Which government agenices bought all those vitamins?

  6. 

    Henry,

    I hope I make this one harder than shooting fish in a barrel — that does seem easy — but I’m not sure I’m up to the challenge if one starts with the premise that the law has no business in seeking to reduce the deadweight loss associated with price-fixing.

    Still, I’ll give it a shot.

    1. I do not think you accurately describe the empirical literature on the effects of cartels. Many of evidence comes from private cartel cases (and of course, there is enough bid-rigging evidence to go around as well).

    I do not think it controversial, based on the existing evidence, to conclude the cartels have the predicted microeconomic effects in terms of deadweight loss, and that they are sizeable.

    2. Let me now confront the point you raise at the end.

    The argument is that sanctioning cartel activity will just result in rational firms substituting toward merger activity, a less efficient alternative.

    It is true the price-fixing and anti-competitive mergers are substitutes — which now brings to two the count of things upon which we have come to agreement (with the first being that I did not make the analogy you inferred in the first instance).

    But not all mergers are created equal. Some mergers are motivated for efficiency reasons, others anticompetitive motives. I happen to think that the latter group is small but existent. Others disagree. The evidence in this score is, again, mixed.

    But yes, surely it would be perverse policy to stop price-fixing while legalizing ENTIRELY all mergers. Current merger policy should be based on its track record, which can be best (though imperfectly) be gleaned from serious empirical studies evaluated its effects. The performance policy of modern (lets say, the last decade or two) is certainly mixed, and needs improvement. The monopolization track record, I think, is pretty bad.

    So I’m happy to cop to the position that an appropriately restricted merger policy that aims at mergers that are the closest substitutes for price-fixing arrangements would be a necessary part of keeping the price-fixing regime I think is socially valuable. I don’t think section 2 is necessary for either of those purposes.

    3. If that position puts me on any slippery slopes, I guess I’ll have to live there with the following defense. Whether the slopes are disastrous or not, or whether having price-fixing laws lead to the consequences suggested have testable implications. Indeed, we’ve had both for awhile.

    I’m comfortable hanging my hat on the existing evidence that price-fixing prohibitions make society and consumers better off in the traditional efficiency sense. If that is not the metric we are using for this conversation, and it appears that it may not be, then I misunderstood the premise.

    4. “Just when did they become affected with a public interest, and when did society gain this right over private property, and when do you stop it?”

    Approximately 1890 is my answer. The real question seems to be whether restricting firms ability to exercise their property rights in this way (we must define the scope somehow, mustn’t we) has resulted in good or bad. Restrictions on price-fixing, the evidence shows, are positive for efficiency and consumer welfare. Restrictions on single firm contracting, i.e. exclusive contracts and such, I’d argue, do not. But I don’t think it so unreasonable to ask about the consequences of where we decide to draw those lines. Nor to follow the prescriptions of basic price theory re: the effects of cartels.

  7. 

    Josh,

    I will address this (final) comment at the quote below from your comment. In passing it also takes care of Thurman, but that is like shooting fish in a barrel.

    ” The example of enforcing property rights was simply that it was one obvious welfare improving law (law X, e.g. a prohibition on stealing) that came to mind — not that the consumer welfare losses associated with monopoly pricing are property. We can use other efficient rules and apply the same logic. But I’m not sure why it matters in the context of the claim here and so I’m going to stick to my guns on the appropriateness of the analogy for the purpose proffered.

    To cut to the chase, I do not think cartel enforcement is inextricably intertwined to merger and monopolization enforcement in the way that the hypothetical raises such that we should abandon it entirely given the empirical evidence of its beneficial effects.

    To cut to the chase further, you write that the cartelists “should be able to contract among themselves and with others so long as their doing so does not occasion some external damage to a third party.” But that’s exactly the point. The monopoly distortion and reduction in output and increase in price results in classic third party effects. Again, this is not conduct with redeeming social value.”

    Let’s grant that you did not actually make the analogy I imferred, but it was suggested by Aremntano’s, to me, irrefutable logic that business people, yes even incorporations, have transferable property rights in the goods they own and that includes the right to contract with competitors about pricing. To say anything less than that puts you directly on Geoff’s slippery slope (where I already find you in your suggestion that business people need to defend themselves with “redeeming social value” in their behavior. Just when did they become affected with a public interest, and when did society gain this right over private property, and when do you stop it? And since when is a failure to produce something that someone else wants a “classic third-party effect?”

    Now let me make two other points about cartels that you and others shoul be familiar with. The first relates to that empirical “proof” you cite for cartel harm. You will please notice that most of the cases cited (and most of the long-lived ones) are public bid-rigging cases, not really private cartel cases in the usual sense. Obviously private buyers have more incentive than do public buyers to disrupt cartel agreements, but this really suggests something fundamentally wrong with our public contracting laws and not something implicating antitrust.

    The final point is one you continue to skirt around and never confront directly. No matter how you slice it, any merger involves price fixing, and, since a cartel is only significant as a price fixing agreeement, you cannot outlaw the lesser matter while legalizing the larger one. If your cartel enforcement scheme were effective (which I always doubt) then all you would succeed in doing is creating more mergers, thus forcing coordination of much more than just prices and thereby, forcing a more expensive way of doing the same thing.

  8. 

    Although he proves my slippery slope point, Thurman (like Josh) is surely correct that there is a possible third-party harm from cartels and a possible efficient corrective from the government, but is there in fact a reliable economic analysis that demonstrates that the cost of cartel enforcement–including both direct costs as well as error costs–is worth the candle? Or is it just a matter of faith because, in fact, the direct costs aren’t that high, the error costs are left out of the calculation (as are the avoidance costs) and many of the benefits substantial? I guess the question comes down to what is the trade-off between (1) the cost of cartel prices times duration without enforcement, which surely is not infinity, minus the cost of cartels times duration with enforcement versus (2) the full costs of enforcement? Do we know?

  9. 

    To get back to antitrust: in questioning the application of antitrust to cartels, Henry Manne argues that firms “should be able to contract among themselves and with others so long as their doing so does not occasion some external damage to a third party.” It seems to me there is an obvious and systematic market failure when firms contract with each other to form a cartel – creating external damage to a third party – that antirust law corrects, even if we assume that the goal is efficiency (total welfare rather than consumer welfare). If sellers are concentrated and buyers numerous (think consumers), and the sellers form a cartel, the consumers have a collective action problem creating a market failure: they can’t be expected to act in concert to bribe the sellers to produce the efficient output. (And the seller won’t produce the efficient output on their own, as they cannot be expected to discriminate in price perfectly.) Antitrust corrects or deters that market failure. And, contrary to Henry Manne’s conjecture in a later comment, there is evidence that cartels can last for a long time. (See, for example, Levenstein & Suslow’s survey in the Journal of Economic Literature (March 2006), which documents the bimodal distribution of cartel duration.) I do not mean to suggest that the only competitive problem justifying antitrust intervention is with cartels. (I could explain why the same market failure justifies antitrust intervention against anticompetitive mergers and exclusion, take on Josh’s argument about whether government can make things better when the anticompetitive conduct is not collusive, and discuss whether there are settings in which buyers can protect themselves.) My point here is simply that antitrust prohibitions on cartel activity are justified on economic grounds.

  10. 

    [Intemperate and defamatory comment deleted, as with subsequent comments (including my own) responding to it]

  11. 

    I don’t follow. Of course there is no property right in the competitive market price. Nor did I say so. The analogy was in response to Geoff’s question about whether we should withhold support from welfare improving law X if it is inextricably linked to welfare decreasing law Y. The example of enforcing property rights was simply that it was one obvious welfare improving law (law X, e.g. a prohibition on stealing) that came to mind — not that the consumer welfare losses associated with monopoly pricing are property. We can use other efficient rules and apply the same logic. But I’m not sure why it matters in the context of the claim here and so I’m going to stick to my guns on the appropriateness of the analogy for the purpose proffered.

    To cut to the chase, I do not think cartel enforcement is inextricably intertwined to merger and monopolization enforcement in the way that the hypothetical raises such that we should abandon it entirely given the empirical evidence of its beneficial effects.

    To cut to the chase further, you write that the cartelists “should be able to contract among themselves and with others so long as their doing so does not occasion some external damage to a third party.” But that’s exactly the point. The monopoly distortion and reduction in output and increase in price results in classic third party effects. Again, this is not conduct with redeeming social value.

    I confess that I simply do not understand economically grounded defenses of naked price-fixing.

    There are defenses I would understand, e.g., the slippery slope argument Geoff raises, or the related argument that things like RPM and other efficient conduct will be swept up into the price-fixing net. The latter in particular has some appeal given that there is some history there and we rely merely on prosecutorial discretion to ensure that the DOJ doesn’t start pursuing jail sentences for monopolization violations or RPM contracts tomorrow. But I don’t think either of these are sufficient to warrant the abolition argument for naked cartel activity in light of the empirical evidence that the regulations thereof improves consumer welfare and efficiency.

  12. 

    And furthermore your argument by analogy to laws against stealing property is entirely inapproipriate. Thre is no property right that the purchaser of a cartel-priced item has lost by the purchase. There is no property right in the competitive market equilibrium (whatever those words mean) market price of anything. The property right is in the cartel members, and here I think that Armentano has the better of the argument, sans empirics. They should be able to contract among themselves and with others so long as their doing so does not occasion some external damage to a third party. Where’s the beef?

  13. 

    Well, presumably the reality is somewhere between never and always, and I admit it’s an empirical question but it’s a factor that seems to get zero weight and should probably get more than zero. And it’s not a “power hungry enforcers” story–the connection would seem to be much less tenuous, as you described yourself the logical, small steps from cartel enforcement to merger and monopolization enforcement. Your counter-hypothetical is far less interesting than my hypothetical. Nyah nyah.

  14. 

    Well, if you really mean “always and everywhere” cartel enforcement brings larger social losses than gains in the manner you suggest (and this could be empirically demonstrated), I don’t think its a very interesting question. Sure, the answer is none in that case. And I’m open to being persuaded by empirical evidence. But until then, the existing empirical evidence goes the other direction.

    And if you could be convinced that enforcement of property rights to prohibit theft led to a slippery slope wherein government enforcers became power hungry and overzealously enforced other, less social welfare increasing laws so that there was a net social loss? Would you give up on property rights?

    I don’t know how fruitful the hypothetical is.

    I’d probably just move to New Zealand.

  15. 

    What if you could be convinced that there was a slippery slope–that allowing cartel prosecutions always and everywhere leads to merger and monopolization prosecutions and that the sum total has more cost than benefit? I’m sure one can construct such a story, having to do with the intellectual dynamics you describe above, “epistemic communities” of antitrust experts, bureaucratic growth and rent seeking. Admittedly, this is more nuanced than what Armentano is suggesting, I suspect, but if you can’t have one without the other, would you rather have none or both?

  16. 

    Thanks Henry. A link to a discussion (with cites to some of the recent evidence) is here:

    http://www.justice.gov/atr/public/articles/240611.htm

    My own read of this evidence is that cartel activity imposes very significant social welfare losses, and when limited to naked price-fixing conduct, prohibitions on this sort of activity make quite a bit of economic sense. Of interest might be the finding that average cartel duration was something of the order of six years. More generally, I think the burden is on those advocating abolition to make their case in the face of both economic theory and evidence suggesting that the losses from hard-core price fixing to consumers are likely to be substantial.

    To the extent that the more general point you are making is that the evidence is imperfect, I agree. It is. But it is also difficult to establish beyond a reasonable doubt that prohibitions on stealing have net negative welfare consequences — yet I am quite comfortable with rules prohibiting such conduct and believe that the evidence is sufficient to shift the burden in the cartel area where the conduct does not have any redeeming social value.

    I think the burden flips in the monopolization / merger arena since the efficiency explanations for these forms of conduct are likely to apply in the vast majority of cases and so error cost concerns are prominent.

  17. 

    Joshua,

    I am not familiar with this evidence you allude to for the benefits of antitrust laws against cartels, but I would bet that, if it exists, it fails to deal with the difficult problem of establishing negatives, i.e. knowing what would have happened to any cartel, i.e. how long it would have survived, in the absence of antitrust prosecution. Make a good estimate for that, and I would wager that any apparent empirical argument for antitrust disappears.