This essay is a reply to Daniel A. Crane’s Has the Obama Justice Department Reinvigorated Antitrust Enforcement?.
We recently concluded that government merger enforcement statistics “provide clear evidence that the Obama Administration reinvigorated merger enforcement, as it set out to do.” 1 Jonathan B. Baker & Carl Shapiro, Reinvigorated Merger Enforcement in the Obama Administration, Antitrust & Competition Pol’y Blog (June 25, 2012), http://lawprofessors.typepad.com/antitrustprof_blog/2012/06/reinvigorated-merger-enforcement-in-the-obama-administration.html. Three weeks later, in an article published in the Stanford Law Review Online, Professor Daniel A. Crane reached the opposite conclusion, claiming that “[t]he merger statistics do not evidence ‘reinvigoration’ of merger enforcement under Obama.” 2 Daniel A. Crane, Has the Obama Justice Department Reinvigorated Antitrust Enforcement?, 65 Stan. L. Rev. Online 13, 16 (July 18, 2012), http://www.stanfordlawreview.org/online/obama-antitrust-enforcement.
Crane is simply wrong. The data regarding merger enforcement unambiguously support our conclusion and cannot reasonably be read to support Crane’s assertions. Crane’s conclusion regarding merger enforcement is inaccurate because he relies upon flawed metrics and overlooks or misinterprets other important evidence.
We should disclose that in evaluating the Obama Administration’s antitrust policy, we are not entirely disinterested academics. Both of us served in the Obama Administration in positions involving competition policy and enforcement. 3 Baker was Chief Economist at the FCC for two years during the Obama Administration, and in that capacity, he worked closely with the Antitrust Division of the Justice Department in reviewing two mergers Crane mentions (Comcast/NBC Universal and AT&T/T-Mobile). Shapiro was the Deputy Assistant Attorney General for Economics at the Antitrust Division for two years during the Obama Administration, and he subsequently served on the President’s Council of Economic Advisers. We also held senior antitrust enforcement positions during the Clinton administration: Baker as Director of the Bureau of Economics at the Federal Trade Commission and Shapiro as Deputy Assistant Attorney General for Economics in the Antitrust Division. But our academic work on merger enforcement trends (other than the brief update quoted above) was written before Barack Obama was elected President, so we developed the standards we apply now without reference to the current debate. Because our work focused on merger enforcement trends across administrations, we limit our comments on Crane’s article to its discussion of mergers—the longest of his three substantive sections.
Comparing Crane’s discussion with the approach we took in our book chapter and subsequent article reveals the problems with Crane’s analysis. 4 See Jonathan B. Baker & Carl Shapiro, Reinvigorating Horizontal Merger Enforcement, in How the Chicago School Overshot the Mark: The Effect of Conservative Economic Analysis on U.S. Antitrust 235, 244-46 (Robert Pitofsky ed., 2008); Jonathan B. Baker & Carl Shapiro, Detecting and Reversing the Decline in Horizontal Merger Enforcement, Antitrust, Summer 2008, at 29, 30-32. Our key statistic was the ratio of agency merger enforcement actions (litigation, consent settlements, and abandonments) to merger filings. This measure had previously been employed to analyze agency enforcement trends in a study by FTC Commissioner Thomas Leary. A low value for this statistic indicates an unanticipated decrease in merger enforcement, and a large and sustained dip to a level below the norm identifies an extended period of substantially more lax merger enforcement. As we reported in our book chapter, this measure showed that merger enforcement at the DOJ during George W. Bush’s first term and the first half of his second term was surprisingly low, even after accounting for expectations that a new Republican administration would resolve close cases more in favor of permitting mergers than would the prior Democratic administration. We found that the depressed enforcement level at the DOJ was comparable to the low rate observed there during the second term of the Reagan Administration. 5 We also analyzed comparable statistics for the FTC. Professor Crane focuses solely on DOJ, so we limit our attention here to DOJ as well.
We now have two years of data on the Obama Administration, enough to make a preliminary comparison. 6 For reasons discussed in our articles, FY 2009 is attributed to the Bush administration, just as FY 2001 was attributed to the Clinton administration. The table below also updates the Bush second term data to include the last two years. Our previous work provides two benchmarks based on past experience: a 1.8% rate is the long-term average since the start of the Reagan Administration, and a 0.75% rate indicates severely reduced enforcement levels. 7 These benchmarks account for a change in Hart-Scott-Rodino reporting rules that took effect in 2001, and reduced the number of mergers filed by 60%. Baker & Shapiro, Reinvigorating, supra note 4 at 246 n. 85. The 1.8% figure corresponds to the long term average and the 0.75% figure corresponds to the severely depressed rate at the DOJ during the second term of the Reagan Administration. The data from that period are consistent with contemporaneous reports that senior officials frequently overruled staff recommendations to challenge acquisitions, and the few mergers that were challenged were typically mergers to very high levels of concentration.
Ratio of Agency Merger Enforcement Actions to Merger Filings
|Bush 1st term (FY 2002-05)||0.75%|
|Bush 2nd term (FY 2006-09)||0.9%|
|Obama 1st term (FY 2010-11)||1.5%|
These data show a clear change of course at the DOJ, from severely lax merger enforcement during the Bush Administration to a level during the Obama Administration that we described as close to the average when previously discussing the Bush-era FTC figures.
As this sketch suggests, we adopted a well-defined measure of enforcement activity previously used in the academic literature. In our previous articles, we explained its theoretical justification—why persistent deviations in that measure from its long-term average should reflect unanticipated changes in agency merger enforcement activity—based on ideas from the law and economics literature. 8 As shifts in merger enforcement standards come to be understood by antitrust counsel and merging firms take those changing standards into account, the frequency of challenges should tend to revert to the long term average regardless of whether enforcement standards are tough or lax. Accordingly, sustained periods in which enforcement rates are substantially below the norm reflect unusual laxity in agency enforcement standards. Id. at 245. We established benchmarks for the measure, and showed that the measure provides a reasonable interpretation of the enforcement history of both antitrust agencies. 9 Id. at 245-46. We showed quantitatively why our conclusions were not called into question by various measurement issues. 10 These issues include variation in the mix of mergers presented to the enforcement agencies, and changes in merging firm expectations about the severity of merger enforcement. Id. at 245; Baker & Shapiro, Detecting, supra note 4 at 30-31. When we updated the statistics with the preliminary data available for the Obama Administration, we applied the methodology we had previously developed and tested, and compared the statistics about Justice Department enforcement rates against benchmarks we had previously established, tested, and employed when analyzing the enforcement trends under previous administrations.
Crane did not approach the analysis of merger enforcement data rigorously by academic standards. He did not even mention the merger enforcement statistic on which we relied, notwithstanding its prior use in the academic literature. Instead, he primarily compared the change between the Bush and Obama years in the number of merger investigations and the number of second requests. His analysis of these data is flawed for four reasons.
First, Crane gives more attention to the raw numbers than to rates normalized by the number of merger filings, even though he recognizes that trends in the raw figures poorly measure variation in enforcement attitude because they are heavily influenced by fluctuations in merger activity. This is particularly problematic for interpreting merger data during the Bush and Obama Administrations because merger filings dropped dramatically due to the financial crisis.
Second, Crane’s yardsticks, whether viewed as raw numbers or as ratios to the number of filings (which he also calculates), are unreliable as measures of agency enforcement. A decline in either could mean that the agencies have grown more lax in enforcement (his implicit interpretation), but such a decline could also mean the agencies have become more efficient in targeting potential problems. If enforcers open fewer investigations or continue fewer investigations after an initial round of information gathering, that does not necessarily mean that they have relaxed their enforcement standards. It might instead mean they have become more successful at weeding out transactions that do not harm competition at an early stage, thus avoiding a more extensive review. We pointed out this problem with respect to the second request rate in our academic articles, but Crane relies on these yardsticks without acknowledging or addressing our critique.
Third, if Crane’s measures are nevertheless taken seriously, the trends he reports support our position. He reports a 25% increase in the rate of investigations per filing and a 50% increase in the rate of second requests per filing—but downplays the large percentage increases by terming the figures “comparable” across administrations notwithstanding a “tick up” in the second request rate.
Fourth, Crane did not look to see whether the statistics he analyzed give sensible or anomalous interpretations of merger enforcement patterns before the George W. Bush Administration. By contrast, in developing our enforcement measure, we examined trends from the Reagan Administration forward. Nor did Crane establish benchmarks (determining what would count as a high number and what would count as a low number) with which to evaluate the measures he employed.
In addition to discussing data on merger investigations and second requests, Crane noted trends in the raw numbers of two other measures: challenges and transactions restructured or abandoned to avoid a complaint. However, he did not calculate the ratio of those figures to the number of filings, even though he calculated the ratio for the other measures on which he relied. Doing so would have created measures related to the statistic we employed to reach the opposite conclusion from his.
We agree with Crane that qualitative measures should be used along with merger statistics to understand enforcement patterns at the antitrust agencies. Our conclusions were also based on the results of a survey of experienced practitioners. 11 Baker & Shapiro, Reinvigorating, supra note 4 at 247-48. We also corroborated our interpretation of enforcement patterns in the George W. Bush Administration through a detailed analysis of the decision not to challenge a high-profile transaction, and we connected trends in agency enforcement with trends in merger review in the courts. 12 See generally id.; Baker & Shapiro, Detecting, supra note 4.
Crane did look beyond the merger statistics when analyzing merger enforcement patterns as a whole. He did not survey practitioners, but he did offer an informal and subjective review of some high-profile DOJ enforcement decisions and guidance documents from the Obama Administration. The latter review missed the mark throughout, though, often because it exhibited little awareness of the context in which decisions were made and did not take into account information that was well known to merger experts in the bar and other close students of agency merger enforcement.
Crane called the decision to challenge AT&T/T-Mobile “conventional” and not “theoretically or factually adventurous.” However, this does not give the DOJ credit for developing a strong case and bringing suit in a high-profile matter. Crane did not note that, in the wake of the DOJ’s Bush-era loss in its suit challenging Oracle’s acquisition of PeopleSoft, the Department had commonly been viewed as gun-shy about merger litigation, especially in unilateral effects cases. 13 As one of the experienced antitrust practitioners we surveyed near the end of the George W. Bush administration explained, “Oracle has been a major factor in DOJ decisions not to bring a case.” Baker & Shapiro, Reinvigorating, supra note 4 at 248. He also did not note that the DOJ’s success in litigating its challenge to the H&R Block/Tax Act transaction during the Obama Administration changed its reputation, while establishing an important pro-enforcement case precedent. 14 See James A. Keyte, United States v. H&R Block: The DOJ Invokes Brown Shoe to Shed the Oracle Albatross, Antitrust, Spring 2012, at 32, 32 (“[T]he desire for a litigated win in the shadow of Oracle had become palpable in the hallways of the Antitrust Division.”); Scott A. Sher & Andrea Agathoklis Murino, Unilateral Effects in Technology Markets: Oracle, H&R Block, and What It All Means, Antitrust, Summer 2012, at 46, 46 (“[T]he DOJ’s recent victory in H&R Block has reinvigorated a mode of unilateral effects analysis that had been seriously undermined when the DOJ lost the Oracle case.”) (footnotes omitted).
Crane chided the Obama DOJ for taking settlements in two high-profile vertical mergers, LiveNation/TicketMaster and Comcast/NBC Universal, rather than challenging those mergers in court, but his brief discussion of these cases recognizes that “the Administration required significant procompetitive structural and/or conduct commitments in both cases.” Although Crane purported to compare antitrust DOJ enforcement in the Bush and Obama Administrations, he never asked whether the Bush DOJ would have sought remedies as strong as the Obama DOJ obtained in these two high-profile cases—or any relief at all. 15 See 2012 Antitrust Merger Enforcement Update and Outlook, Gibson Dunn 7 (March 9, 2012), http://www.gibsondunn.com/publications/Documents/2012AntitrustMergerEnforcementUpdate-Outlook.pdf (“It is now conventional wisdom that merger enforcement has been and will continue to be a priority under the Obama Administration. In particular, vertical mergers . . . have received far more scrutiny than they had under prior administrations.”). Yet this should have been an obvious question given Crane’s description of the theories of harm in these cases as “more adventurous,” a term rarely used to describe the enforcement approach of the Bush DOJ.
In the same vein, Crane criticizes the Obama DOJ for promulgating revised remedies guidelines that he reads as being more receptive to conduct remedies in vertical cases than the remedies guidelines promulgated by the Bush Administration. But Crane never considers whether those remedy guidelines have served to facilitate conduct remedies in cases where the Bush DOJ would have sought no relief at all. Vertical mergers are much harder for the antitrust agencies to challenge than horizontal mergers. 16 See Herbert Hovenkamp, Federal Antitrust Policy: The Law of Competition and its Practice 430 (4th ed. 2011) (“Prevailing judicial opinion now seems to be that vertical mergers should be condemned only in the most extreme circumstances.”). Our intimate knowledge of the LiveNation/Ticketmaster and Comcast/NBC Universal cases allows us to say with confidence that the DOJ pressed hard for the strongest remedies that were both feasible and desirable. In the Comcast/NBC Universal case, the DOJ cooperated effectively with the FCC to achieve that end.
Crane’s discussion of the 2010 revision of the Horizontal Merger Guidelines also cannot be credited. Most striking is his statement that raising the HHI thresholds “suggests that greater levels of concentration resulting from a horizontal merger will be necessary to trigger antitrust scrutiny than under the previous regime.” This “suggestion,” while sensible in the abstract, does not reflect reality. The extensive public record surrounding the development and release of those guidelines makes it clear that neither the DOJ nor the FTC had applied those thresholds for many years, and the primary purpose of changing the HHI thresholds in the Guidelines was to align the Horizontal Merger Guidelines with enforcement reality, not to signal a more permissive policy. 17 See, e.g., Hill Wellford & Gregory Wells, The “Litigation Mulligan” in the 2010 Merger Guidelines: Better Economics but Not (Necessarily) More Clarity Before the Agencies and the Courts, 10 CPI Antitrust Chron., no. 2, Oct. 7, 2010, at 7 (“The 1992 HHI thresholds were well known to be out of date, and the new thresholds merely recognize longstanding and uncontroversial Agency practice.”). This is analogous to a situation where the posted speed limit has long been 50 miles per hour, but most cars are going 70 miles per hour or faster, and few or no tickets are being issued. Raising the speed limit to 60 miles per hour and then enforcing that limit is unquestionably a stricter enforcement regime. All in all, Crane’s review of Obama Administration merger cases and guidance does nothing to rehabilitate his unconvincing interpretation of the merger statistics.
Our analysis of merger enforcement at the DOJ during the George W. Bush Administration—based on the enforcement statistics and more—showed that it was unusually lax and in need of reinvigoration. It is too early to reach a comparably definitive conclusion about merger enforcement at the DOJ during the Obama Administration, but nothing in Daniel Crane’s article seriously challenges our interpretation of the preliminary data as demonstrating that the necessary reinvigoration has taken place.