This article provides the most comprehensive study to date of the Federal Trade Commission's Part 3 process. A unique and defining tool, Part 3 allows the FTC to challenge alleged section 5 violations in-house through administrative litigation. Highlighting the agency's expertise, Congress gave the FTC that authority in 1914 to develop antitrust and later consumer-protection law. Although the Commission has used its Part 3 authority to good effect, especially in some recent competition matters, the process is controversial on due-process grounds. After it finds “reason to believe” that a violation exists and authorizes staff to litigate before an independent Administrative Law Judge, the Commission reviews the ALJ's decision de novo. The result, some critics argue, is a foregone conclusion. Rather than advance the law, they contend, Part 3 frees the Commission to reach its favored result. Maligned in some quarters as a “kangaroo court” in which the FTC alleges and later summarily confirms section 5 violations, the key question is whether administrative litigation effectively fulfills the role Congress set out for it. To resolve the debate, we need a clearer picture of how Part 3 operates. Does the FTC rubber stamp its prior determinations, invariably siding with complaint counsel? Or does it change course at the appeal stage and, if so, how often? And, when it does find a section 5 violation on appeal, how often does the FTC nevertheless prune counts and allegations, suggesting that it scrutinizes the factual and legal record before it? Do the answers change with the political constitution of the Commission, such as when a Democratic majority votes out a Part 3 complaint, but a Republican Commission hears the appeal? How often does the Commission's composition change between voting out and later deciding a Part 3 matter, and might the answer resolve the due-process issue? And does the FTC fare better at the U.S. Courts of Appeal if it affirmed or reversed the ALJ? Until now, relatively little empirical work has scrutinized administrative litigation at the FTC. This study tracks every Part 3 case that produced a Commission Decision on or after January 20, 1977. It captures all 145 antitrust and consumer-protection matters falling within that period, tracking their development through initial vote at the “reason to believe” stage, disposition before the ALJ, appeal before the Commission, and petition for review at the U.S. Court of Appeals and beyond. The emerging statistics are illuminative. In hearing an appeal in the last 10 years, the FTC has never rejected an action that it had previously authorized complaint counsel to bring in Part 3. At first blush, that fact might suggest a preordained appellate process. But that phenomenon dissipates when one looks at a larger time horizon, suggesting that uniformity from initial vote to appeal is neither inevitable nor systemic. Indeed, it is possible that recent consistency may be a function of improved case selection, aided by effective factual, economic, and legal analysis by staff and the Commissioners before they authorize a complaint. Evaluating the ten matters voted out and affirmed by the Commission in the last decade lends some support to that proposition. Indeed, Part 3 has been an effective tool in developing complex antitrust questions. This article provides a host of new evidence with which to evaluate administrative litigation at the FTC. The article concludes by exploring some initial teaching points from the data.
The Federal Trade Commission is one of the world's leading antitrust and consumer-protection enforcers. Among its powers is “Part 3,” an administrative litigation procedure through which the FTC can challenge anticompetitive conduct and unfair or deceptive acts or practices.1 Congress's rationale for granting the FTC this power is that, as an expert agency, the Commission is well placed to resolve difficult questions of competition and consumer-protection law. The FTC has made great strides in doing so. In litigating seven cases before the Supreme Court in the last thirty years, for example, it has won six times.2 Five of those seven matters arose from Part 3.3 And, in the last decade alone, the FTC has used administrative litigation to advance the law on healthcare-merger enforcement, dominant-firm misconduct, pay-for-delay agreements in the pharmaceutical industry, state-action immunity, deceptive acts or practices, and beyond.4
Nevertheless, the FTC's administrative litigation process, examined in Part III.A, stands accused of being a rigged system. In a Part 3 proceeding, the FTC serves prosecutorial and adjudicative roles. After a staff investigation and recommendation, and following party meetings, the Commissioners may vote out a Part 3 complaint if they find “reason to believe” that a section 5 violation occurred and that the action would serve the public interest.5 An independent administrative law judge (ALJ) subsequently reaches an initial decision, based on a full trial. The Commissioners, however, then decide the merits of the case without deferring to the ALJ's factual or legal findings pursuant to their regulatory authority.6 Due-process objections result. The worry is that, once the FTC authorizes a Part 3 complaint, liability is inevitable no matter how the ALJ rules or what new facts or legal issues emerge. To evidence those claims, some commentators have argued that the Commission almost always rules in complaint counsel's favor.7
Is such criticism accurate?8 The answer lies in facts that have previously been elusive. To assess the value of Part 3, it is also crucial to gauge the effectiveness of administrative litigation at the FTC in serving its function to develop the law. This article provides the most exhaustive empirical study to date of the FTC's Part 3 process. It captures 145 Part 3 cases in which the FTC rendered a decision between January 20, 1977 and July 31, 2016. For each case, it tracks the composition of the Commission in voting out and later deciding a Part 3 matter, the result before the ALJ, the Commission, and on petition for review to the appellate courts. It explores trends across time and political affiliation. It identifies whether a dismissal was on the merits, reflected a change in the law or facts or lack of jurisdiction, or flowed from a determination that continuing the litigation was no longer in the public interest. As explained in the Appendix, to focus on the most illuminative matters, the study excludes Part 3 cases resolved by consent before a Commission decision, dismissed after the courts denied the FTC's motion to preliminarily enjoin an unconsummated merger, and other matters. The article marshals the data and presents it along dimensions most relevant to the debate on Part 3’s due-process implications and mission efficacy.
Interesting conclusions emerge from the data, which Part II presents in full. For the purposes of this brief introduction, however, consider five findings.
First, a recurring claim is that the FTC always imposes liability. That claim is true, but only for recent cases, which are relatively few. Although the Commission found liability in 92 percent of its 12 Part 3 decisions in the last decade, it dismissed 29 percent of the 143 Part 3 matters in which it made a liability-dismissal decision since January 1977.9 The Commission dismissed 16 percent of Part 3 matters on the merits. Those dismissals do not include the 21 matters in which the Commission found liability, but also trimmed counts or respondents, suggesting careful review. Notably, the Commission dismissed 40 percent of Part 3 antitrust complaints overall. And, of the antitrust matters adjudicated on the merits, 29 percent were dismissed.
Second, although the Commission dismissed more administrative cases historically—36 percent from 1987 to 1996, for example—its recent trend of finding liability coincides with a higher rate of success before the appellate courts. Although the judiciary affirmed Commission liability findings in 50 percent of cases from 1987 to 1996, since 2007 it has affirmed 100 percent of appealed Part 3 case to date in which the Commission found liability. Those findings are consistent with the seismic change in antitrust law, which shifted toward sophisticated economic analysis starting in the late 1970s and early 1980s.10 Though perhaps slow to adapt to this trend, the FTC's eventual embrace of modern industrial-organization economics likely led the Commission to select more meritorious cases over time, which would produce a higher rate of liability findings by the Commission and success in the appellate courts.
Third, what of the criticism that the Commission is biased in deciding the merits simply because it previously instituted the proceedings?11 In fact, in 72 percent of Part 3 cases, the commissioners who authorized the administrative litigation had either left or no longer formed a majority at the liability-dismissal stage. The argument for a systemic bias wobbles in light of the data because the same commissioners rarely voted out and later decided a Part 3 matter. For example, of the 51 cases in our study authorized between 1967 and 1976, a different commission majority made the liability-dismissal decision in 92 percent of them. Interestingly, however, only in the last decade—during which the Part 3 process been more expeditious—has the same commission majority generally authorized and resolved a Part 3 matter (90 percent), a period with both a high percentage of liability and a high success rate on appeal.
Fourth, a fascinating question is whether the Commission finds liability more often when the same majority authorized the Part 3 complaint. If claims of bias have merit, the FTC should be more likely to dismiss when different commissioners authorized the case. The results are the opposite. When the same Commission majority endured—that is, when “bias” would presumably exist—it dismissed 33 percent of cases. When a different majority decided the case than voted out the complaint, however, it was less likely to dismiss—doing so in 27 percent of matters. Put differently, respondents fared worse (losing 73 percent versus 67 percent of the time) when the commissioners who authorized the action no longer constituted a majority.12 Certainly, there are complications and subtleties behind these aggregate numbers, but that simply means that one must interpret statistics cautiously.
These findings are the tip of the iceberg. Part II lays out the empirical findings in detail, allowing readers to discern points of interest for themselves. It also addresses the factors that make it difficult to draw broad conclusions from the data. Part III explores the existing commentary and literature on Part 3, contextualizing this study's statistics. Part IV moves from the observational to the analytic and normative. It peers behind percentages to scrutinize a representative sample of the underlying cases, testing various propositions about the value and purported deficiencies of administrative litigation at the FTC. It mounts a firm, though qualified, defense of Part 3. An appendix explains the technical details underlying this empirical study.
CORE FINDINGS: THE FTC'S PART 3 PROCESS OVER THE LAST FORTY YEARS
This Part presents the article's core findings. Given the volume of results, it intersperses discussion with graphical and tabular presentations of the data. As explained in the attached appendix, the dataset comprises 145 matters on which the Commission rendered a Part 3 opinion between January 20, 1977 and July 31, 2016.
Contextualizing the Data
The population of administrative cases in the dataset is not equally distributed over time. To the contrary, the volume of Part 3 matters is heavily weighted toward the 1970s and 1980s. For the 145 matters captured in this study, the following chart shows the number of cases filed by decade, separately for consumer protection (BCP) and competition (BC).(Figure 1.
The asymmetric distribution of cases has important implications for this study and, indeed, for any empirical work on Part 3 that hopes to suggest meaningful conclusions.
First, recent cases are few in number, which limits inferences than one can reliably draw from them. Among other examples, critics have made much of the FTC's practice of finding liability in most cases in the last two decades. Without a richer body of cases to study, however, it is hard to know what to make of that phenomenon.
Second, one can find a larger volume of Part 3 cases by looking back to the 1970s and 1980s, but do those matters represent modern practice? That was an era when the Chicago School's contribution to economic analysis in antitrust cases was only beginning to seep into the case law and agency practice. Indeed, the FTC was strongly criticized in the 1970s and 1980s.13 Hence, it is easy to question whether the FTC's earlier “reason to believe” decisions and merits determinations capture the rigor of today's Part 3 complaints and opinions. Although there are lessons to learn from the 106 cases in the dataset filed between 1971 and 1986, we must be attuned to the dangers of overlooking the impact of sweeping changes in antitrust analysis. That point applies to the outcome of Commission Part 3 decisions themselves as well as their subsequent fates on appeal.
The following chart provides a more detailed picture of how the 145 cases played out over time. Each horizontal line represents a Part 3 matter in the database. The red portion reflects the time between when the Commissioners authorized the complaint and the ALJ issued an initial decision. The green bars represent the duration between the initial decision and the Commission's opinion. Finally, the purple segment captures the time between the FTC's opinion and the ruling on a petition for review from the appellate courts. Two features are striking. First, the duration of cases has shortened over time. Second, and consistent with the last graph, the FTC issued many more Part 3 decisions in the past than it has recently.(Figure 2.
Several dynamic trends emerge from the data. In particular, the FTC has brought many fewer Part 3 cases but become more likely to find liability in Part 3 over time. Counterbalancing that phenomenon, however, is the fact that the FTC's appellate record has also greatly improved.
Commission Decisions to Impose Liability
The core question is whether the Commission performs an objective and rigorous appellate review of the ALJ's initial decisions. This article's dataset sheds new light on the answer. Far from rubber stamping complaint counsel, the Commission dismissed 41 (29 percent) of the cases that it had authorized under Part 3. The FTC has been more willing to depart from complaint allegations in competition cases. There, the Commission dismissed 40 percent of (36 of the 90) competition matters on appeal from an ALJ decision. By contrast, the Commission dismissed only 9 percent of consumer-protection cases.(Figure 3.
The results depart sharply from the conventional narrative. In its more cynical version, the claim is that the Part 3 process—and a respondent's appeal to the Commission in particular—is essentially rigged. But reviewing four decades’ worth of Part 3 matters tells a different story. It is unlikely that the FTC has demonstrated a systemic bias if it dismisses 40 percent of antitrust cases and almost a third of all its cases.
Of course, one must peer behind the number to discern the relevant dynamics. The Commission did not dismiss all of the 41 cases on the merits. Only 23 matters—or approximately 16 percent of its 143 Part 3 liability determinations—fell on the merits.14 What accounted for the other dismissals? The FTC dismissed actions for a variety of reasons, including because they were no longer in the public interest, a lack of jurisdiction, a change in the law, and changed facts.15
Moreover, upholding or dismissing matters in full are not the only options for the Commission. It can also prune liability counts, allegations, and remedies that, in its view, the facts and law do not support, even though it finds liability on at least one count. In fact, of the 102 final orders where the Commission imposed liability, it struck allegations, counts, or a respondent in 21 of them, which in consistent with a careful review of the complaint allegations.16
Combining those observations contradicts recurring claims that the FTC always rules for complaint counsel. The Commission dismissed 29 percent of antitrust cases decided on the merits and cut allegations, respondents, or counts in 13 percent of matters where it found liability. There is ample room for debate about the proper implications of those findings, but the statistics undermine the claim that the FTC is simply a kangaroo court. So, what is the basis for claims that the Commission always finds in its own favor?
There are two answers to the last question. The first involves time. Some empirical work on Part 3 looks only at the last two decades, when the Commission issued only 28 decisions in administrative litigation as tracked in this study. This article looks back further, doubling the effective timeframe and capturing 117 additional cases. That broader picture tells a different story. For a period, the agency regularly dismissed cases that it had previously voted out. Dismissals have become rarer over time. Looking at the relatively small set of Part 3 matters from the last two decades, one could infer that the deck is stacked in complaint counsel's favor.(Figure 4.
Therein lies the claim that “the FTC has ruled for itself in 100 percent of its cases over the past three decades.”17 Technically, the Commission dismissed 22 percent of its Part 3 matters (11 cases) during that period.18 Not all of those dismissals, however, reflect a substantive merits determination. Earlier this year in Cabell, for instance, the FTC dismissed a Part 3 complaint challenging a healthcare-system acquisition because a change in the law created issues of state-action immunity.19 In R.J. Reynolds, the FTC dismissed its consumer-protection action after the tobacco firm accepted a binding settlement with the attorney generals of 46 states.20 And in Robert Koski the Commission dismissed its complaint against a doctor accused of conspiring to boycott a hospital. It did so with the support of complaint counsel after post-complaint evidence emerged that the physician had not conspired.21 Such dismissals may not shed much light on whether the Commission is predisposed to impose liability after authorizing a Part 3 complaint.
But of the eleven Commission dismissals in the last three decades, five were on the merits. For example, in R.R. Donnelley, the FTC challenged a commercial-printing merger in Part 3, after the district court denied its motion for a preliminary injunction.22 In its final order in 1995, the Commission rejected complaint counsel's proposed relevant market for “high volume gravure printing” and concluded that “neither coordinated nor unilateral anticompetitive effects are a likely result of the acquisition in the assumed market.”23 It reversed the ALJ and dismissed the complaint. Similarly, in 1994 in Adventist, the FTC dismissed a Part 3 complaint against a healthcare-system acquisition. Affirming the ALJ, the Commission held that “complaint counsel failed to carry the burden of proof” on its proposed relevant geographic market.24
In 1992, in Owens-Illinois, the FTC dismissed an administrative action challenging a merger in the glass-container industry. Overruling the ALJ and disagreeing with complaint counsel, the Commission found that “in each of the relevant product markets the evidence suggests that anticompetitive effects are unlikely.”25 In 1989 in Motor Transport, the Commission affirmed an ALJ determination that the state-action doctrine immunized the collective rate-making challenged by the FTC in Part 3 as price-fixing.26 And earlier that year in Midcon, the FTC sided with the ALJ in dismissing a Part 3 complaint against a consummated acquisition in the natural-gas industry.27 The agency found that “complaint counsel [had] failed to show that there was a substantial likelihood of anticompetitive effects in a section of the country.”28
In sum, over the past thirty years the Commission has held in favor of respondents and against complaint counsel in a variety of matters.29 The data does not support the claim that the FTC has demonstrated systemic bias in favor of complaint counsel regardless of the facts, economics, or the larger merits.30
It is true, however, that in the last decade the FTC has found liability in almost every Part 3 case that it had authorized. In LabMD, ECM BioFilms, Jerk, McWane, POM Wonderful, ProMedica, North Carolina Dental, Polypore, Daniel Chapter One, Realcomp II, and Evanston, the Commission found a section 5 violation.31 Only in Cabell-St. Mary's—a healthcare-acquisition that the FTC alleged to be anticompetitive—did the FTC dismiss the Part 3 complaint.32 After the FTC voted to authorize the Part 3 case in that matter, West Virginia passed a law that purported to shield certain healthcare cooperative agreements from antitrust scrutiny, raising potential questions of state-action immunity.33 Thus, the most recent experience is indeed one in which the FTC consistently finds liability.
Does the Commission's near-100 percent record in 12 matters over the last decade reflect careful case selection or a newly developed predisposition to rule in staff's favor? Part IV addresses that question. Evaluating each of those cases on its merits, it suggests that the more likely explanation for the Commission's imposition of liability in all recent cases is the selection of fewer and better cases. This explanation also finds support in the next subpart, which shows that the FTC's success on appeal has also increased over time.
More generally, the fact that the Commission has dismissed cases it previously authorized is important, even if the incidence varies over time. It suggests that the agency's recent track record in siding with complaint counsel is not necessarily evidence of systemic bias or an institutional flaw. Perhaps the FTC's case selection in the last decade, in which the Commission has affirmed liability in 11 of 12 Part 3 decisions, has been more rigorous and thus better. An examination of the merits of those decisions reveals that proposition to be plausible.
The FTC's Appellate Record
Of the 102 administrative matters where the Commission found liability in the last four decades, the parties appealed 75—or 74 percent—of them to the federal courts. Of those petitions for review, 62 challenged a liability determination by the FTC (as opposed to a procedural finding or remedial order). Antitrust matters represented a slight majority of the appeals: 44 of the 75 petitions for review.
The federal courts affirmed liability in 61 percent of petitions for review, reversed 33 percent of cases, remanded 3 percent, and had mixed rulings in the remaining 3 percent. That is not a dramatically impressive appellate record, particularly given that petitioners face an uphill battle because the courts defer to the FTC's factual findings and affirm if substantial evidence supports the agency's final decision.34 On the other hand, firms can generally appeal to any circuit, which often allows them to choose a venue with more favorable precedent.35
Yet, a potentially crucial trend is apparent from the data. The Commission's record on appeal has improved in the last two decades, at the same time it brings fewer cases in Part 3 and dismisses complaints less often. From both 1977 to 1986 and 1987 to 1996, the FTC won 50 percent of appeals. Things got better for Commission Part 3 decisions issued from 1997 to 2006, when the appellate courts affirmed in 67 percent of the cases. For cases decided by the Commission within the last decade, the FTC reached its zenith before the judiciary, winning 100 percent of appeals.36 That fact may help to explain why complaint counsel enjoyed a near-100 percent record before the Commission over the last decade.37 Recently, as examined in Part IV, the agency seems to be bringing strong cases in Part 3.(Figure 5.
It may also be helpful to break down appeals by competition and consumer-protection matters. Of the 40 antitrust matters appealed on liability across the full period of this study, the courts affirmed 20 of them, reversed 16 of them, and reached a mixed conclusion or remanded without a liability determination in the remainder. The Commission fared better in consumer-protection matters, winning 17 of the 22 cases appealed on liability grounds. There are distinct trends over time, however, which these high-level statistics obscure.
On the antitrust side, the FTC prevailed in appeals from liability decisions overtime as follows: 36 percent (1977 to 1986), 33 percent (1987 to 1996), 71 percent (1997 to 2006), and 100 percent (2007 to 2016). That trend shows significant improvement. That phenomenon may be especially notable on the antitrust side because the previously high rate of reversals is in some tension with the FTC's role as an expert adjudicator. The record on the consumer-protection side differs notably. There, the Commission won 73 percent, 100 percent, 50 percent, and 100 percent of petitions for review in each of the four preceding decades. Those percentages mask a small number of cases, however, because the database contains only seven consumer-protection matters in which the appellate courts reached a liability decision in the last 30 years.38 In the 1997 to 2006 period in which the FTC's appellate success rate dropped to 50 percent in consumer-protection cases, for instance, there were only two relevant matters.39
A final—but potentially revealing—statistic involves the role, if any, that affirming or reversing the ALJ plays in the FTC's success rate on appeal. In the 47 matters where the FTC affirmed the ALJ and an appeal on liability followed, the Commission won 64 percent of cases. By contrast, when the Commission overruled the ALJ, imposed liability, and the respondent appealed (14 cases), the courts affirmed 43 percent of the time. Based on that observation, there appears to be a benefit to a firm in prevailing before the ALJ, even if the Commission subsequently imposes liability. Bear in mind, however, that these comparative figures—64 percent to 43 percent—derive from limited case numbers.
As the FTC Has Become More Likely to Find Liability, Its Appellate Success Rate Has Improved
Another clear trend emerges regarding the FTC's liability findings and the agency's success on appeal. As the chart below demonstrates, the federal appellate courts have affirmed a higher proportion of the FTC's liability findings over time. By the same token, the FTC has dismissed fewer cases over time. The resulting dynamic is, at the very least, consistent with the proposition that the Commission is identifying superior cases in authorizing Part 3 complaints than it used to do in the past.(Figure 6.
Until Recently, the Same FTC Majority Rarely Authorized and Decided Part 3 Cases
The basic due-process criticism leveled at the FTC's institutional design is that the same people cannot impartially adjudicate a case that they were previously responsible for instituting.40 Combined with claims that the FTC always finds liability in Part 3 matters, it might be tempting to conclude that the Commission has its thumb on the scale in hearing appeals from the ALJ because it made up its mind when it authorized the complaint.
It is significant, then, that different Commissioners voted out and decided most of the administrative cases in this study. In 72 percent of the matters, the commissioners who had voted to authorize the Part 3 complaint no longer were present or formed a majority on the Commission by the liability-dismissal decision. That dynamic is a function of the duration of administrative litigation. In times past, the process was protracted and so, for example, different majority commissions decided 92 percent of the cases in this study filed between 1967 and 1976. Today, administrative litigation is more expeditious, which may explain why in the last decade the same majority Commission authorized and decided 90 percent of cases.41(Figure 7.
In a counter-intuitive finding, the FTC was actually more likely to dismiss a Part 3 complaint when the same majority commission initiated and resolved the case. When the voting majority of commissioners remained, 33 percent of the cases were dismissed. By contrast, when a different majority arose after approving the complaint, the FTC dismissed 27 percent of the cases. The explanation, in part, is that liability-dismissal decisions are more likely to involve the same majority commission when they come soon after the complaint vote. Hence, changes in the law42 or facts may produce more dismissals when the same commissioners responsible for authorizing a Part 3 complaint are still present.
High-Level Observations from the Aggregate Data
We now move beyond large trends in the data to consider several findings that weigh on the due-process issue and on the question whether the FTC's Part 3 process is fulfilling its congressional purpose.
First, the FTC draws particular criticism when it overturns an ALJ's initial decision to dismiss a Part 3 matter. Critics often point to that outcome as evidence of bias by the Commission in finding liability despite an independent fact-finder's determination to the contrary. Of the 35 matters in the dataset that the ALJ dismissed, however, the Commission overturned 19 of them (or in 54 percent of such cases). Hence, reversing the ALJ to find for complaint counsel does not occur as frequently as some commentators might suggest. And of the 95 administrative cases in which the ALJ found liability, the Commission dismissed 11 of them.
Second, when the parties litigate a Part 3 matter before the ALJ, complaint counsel win—meaning that they secure a liability verdict, on at least one count, in—73 percent of cases. Put differently, firms accused of violating section 5 win outright approximately a quarter of the time before the ALJ. Breaking down the data by antitrust and consumer-protection matters, a disparity emerges. Complaint counsel in antitrust matters won 52 of 80 cases or 65 percent of the time. When challenging unfair or deceptive acts or practices in Part 3, they prevailed in 43 of 50 matters or 86 percent of those cases.
The parties’ chances of success before the ALJ, however, vary over time. All four initial decisions on antitrust cases in the last decade found for complaint counsel on at least one count.43 FTC staff performed less well in the 1997 to 2006 period, losing 5 of the 10 competition matters litigated before the ALJ in that time. Complaint counsel fared worse still in the decade from 1987 to 1996, losing 5 of the 8 antitrust cases (63 percent) they tried before the ALJ. A mixed record on the competition side extends back into the 1977 to 1986 time frame, when FTC staff lost completely in 12 of 30 (or 40 percent of) competition matters in Part 3 before the ALJ.
Complaint counsel have fared better before the ALJ in consumer-protection cases. FTC staff have prevailed—at least in part—in all but one of twelve such matters filed between 1987 and 2016. Of the Part 3 matters in this study, complaint counsel only lost 6 of 38 consumer-protection cases (16 percent) filed between 1967 and 1986. It seems that ALJs are more receptive to the FTC's unfair and deceptive acts or practices cases. Maybe the Commission tends to authorize more compelling Part 3 cases on the consumer-protection side, such that the facts and law are clearer than in antitrust matters. Further, and relative to consumer protection, antitrust law saw sweeping changes during this time. Or perhaps competition matters are simply harder to prove, given the economic complexity and factual disputes that inevitably arise on market definition and anticompetitive effects. In either case, FTC staff's asymmetric success rate in competition and consumer-protection cases before the ALJ is striking.
Even in the cases where the ALJ finds liability—73 percent of cases—we can look more deeply at the nature of those holdings. Not all of those matters involved pure wins for complaint counsel. Of the Part 3 matters in which they imposed some form of liability, the ALJ rejected an allegation in 18 percent of those cases, rejected at least one count in 13 percent of those cases, and struck a respondent in 4 percent of those matters. In 61 of the 95 cases (64 percent) where the ALJ found liability, he did so finding in favor of FTC staff on all counts. In short, these statistics reveal that complaint counsel are consistently more likely to win a Part 3 matter before the ALJ, at least in part, but that ALJ findings against FTC staff are by no means rare, at least in competition cases. And, even where complaint counsel ultimately prevail in securing some liability finding, the ALJs are not shy about rejecting counts or allegations that they consider to be inconsistent with the evidence.
Third, the FTC brings almost twice (1.7 times) as many antitrust cases in Part 3 (91) than it does consumer-protection matters (54). If Part 3’s principal advantage over suing in federal court is to develop the law, the disparity in favor of competition enforcement may make sense. Often, though not invariably, competition cases pose more complex issues.
Fourth, when the Commission elects to litigate a matter administratively, it does so unanimously in 78 percent of cases. The level of unanimity in voting out complaints has generally increased over time, rising from 75 percent of the database cases filed between 1967 to 1976 to 100 percent in the last decade, with a dip to a low of 71 percent in 1987 to 1996. Interestingly, unanimity does not seem to depend on whether the Commission majority is Democratic (80 percent) or Republican (80 percent), though disagreement in voting out a Part 3 case rises to 31 percent when the Commission's political constitution is split 50/50. Finally, it bears noting that Commission unanimity extended to the decision stage, as well. There, 86 percent of decisions were unanimous, with 78 percent of decisions being unanimous in dismissing and 89 percent being unanimous in finding liability.
EVALUATING ADMINISTRATIVE ADJUDICATION AT THE FTC
The FTC's Part 3 authority is a powerful tool for developing or clarifying the law. For firms accused of violating section 5, however, the administrative process is rarely their preferred litigation forum. The perception is that the FTC decides the case when it authorizes staff to file a Part 3 complaint and hence is not impartial in later reviewing the ALJ's decision. Should the FTC reach a final decision in complaint counsel's favor, the respondent can petition for judicial review. The accused firm enjoys one significant advantage in that it may file its petition with any U.S. Court of Appeals in whose jurisdiction it conducts business. Nevertheless, the appellate court defers to the FTC's factual determinations, upholding final FTC decisions on a “substantial evidence” basis, even if the court does not defer to the Commission's legal analysis.44
Some companies subject to Part 3 adjudication claim due-process deficiencies. This has spurred a robust debate about the merits of Part 3. This part first examines ways to evaluate Part 3’s efficacy and due-process protections. The findings disclosed in the last part inform the answer. Second, to contextualize the larger debate, this part touches on a sampling of the literature that has questioned or criticized administrative litigation at the FTC. It focuses in particular on claims regarding the inevitability of liability findings when the Commission acts both to vote out a complaint and to review, de novo, the ALJ's legal and factual findings.
How should one evaluate Part 3 and its due-process protections? The answer turns on three questions. The challenge is that none of them is susceptible of direct empirical measurement. The issues are as follows: (1) How accurate are the FTC's complaints? (2) Does the FTC neutrally review the ALJ's initial decisions? and (3) Does judicial appellate review ensure accurate decision-making by the FTC? Each factor matters. First, the accuracy of the Commission's “reason to believe” decision controls much of the analysis that follows. In a world of perfect institutional competence, knowledge, and design, the FTC would base its decision to issue a Part 3 complaint on a solid foundation of legal and economic analysis. A trial on the merits before an independent ALJ would merely confirm the facts that the investigation had revealed. Regardless of the ALJ's ruling, the Commission would rule on appeal consistently with its initial vote. The FTC's decision would stand free of error. There may still be an issue of appearances, but the substantive critique would ultimately be hollow.
Of course, we do not occupy such a world. Hence, it is important that the Commission stand ready to revisit its prior determinations should a full record—bolstered not least by cross-examination of key witnesses and experts—reveal facts, economic insights, or legal principles that the initial investigation and complaint recommendation and deliberation did not adequately uncover. But how do we assess whether the Commission's “reason to believe” findings are well founded and accurate? There is no perfect answer, but the FTC's record before the appellate courts may provide the best available insight.
Second, the FTC's review of the ALJ decision frames the due-process issue. Perhaps it is preferable as a matter of institutional design to split the prosecutorial and adjudicative functions. Even then, however, an independent and careful review by the Commissioners—unmoved by an earlier vote to authorize the complaint—would largely or entirely solve the problem. Hence, the key question: does the FTC assess the merits blind to the Commission's earlier view or does it tip the scales in complaint counsel's favor, unconsciously or otherwise?
Unfortunately, teasing out an answer is not easy. When the Commission reverses the ALJ and imposes liability, for instance, does it (1) simply rubber stamp its prior decision, (2) fully scrutinize the record with an open mind to revisiting and rejecting its earlier, initial determinations,45 or (3) adopt a half-way position, approaching the appeal with a priori expectations consistent with its “reason to believe” decision, but willing to consider evidence or legal theories pointing the other way? The fact that a particular FTC decision favors complaint counsel sheds little light on the answer.
Nevertheless, if we assess a sufficiently large pool of Part 3 decisions, unwavering uniformity may look suspicious. Initial, “reason to believe” votes are of course susceptible to error. Trials, which include live testimony and cross-examination, expose truth more accurately. Even if the Commission makes well-informed decisions in voting out a complaint, we would expect to see some deviation at the Commission liability-dismissal stage over enough cases if a scrutinizing review takes place. Further, even if the FTC does not dismiss or otherwise reject a set of complaint counsel's cases in their entirety, one would expect some trimming of counts, respondents, or remedies to occur if review on appeal is in fact de novo, rigorous, and independent of the prior vote to authorize the complaint.
To answer this second question, however, one must first resolve the first: does the Commission identify meritorious cases in the first instance and screen out poor ones? Unless we know the answer, outcomes on appeal from the ALJ to the FTC tell us little. For instance, if the FTC did a poor job at Part 3 case selection, an unwavering adherence to complaint counsel would signal a problem. In that situation, one could reasonably infer that the Commission made its mind up when it voted out the complaint. By contrast, if the FTC overwhelmingly picked meritorious cases, the fact that it rarely dismissed a Part 3 case on appeal from the ALJ would be consistent with full, independent scrutiny on appeal. As noted, however, it is hard to gauge the quality of the Commission's initial votes. Thus, perhaps the best proxy is to consider how the FTC fares at the federal courts of appeals and at the Supreme Court, using those higher institutional bodies’ rulings as a qualitative metric.
Third, whether judicial review holds the FTC to its proof is an empirical question with no obvious answer. We gain important in sight, however, by looking at the depth of the appellate courts’ scrutiny, how often accused firms’ petitions for review succeed, and whether the FTC's success record has improved over time.
The statistics emerging from Part II above inform these factors and ought to be useful to the substantive debate. We now move to review some of the literature criticizing the FTC's Part 3 process.
Due-Process Criticism of Part 3
Commentators have long worried about the combined functions performed by the Commission in Part 3. In 1981, for instance, the ABA observed, “No thoughtful observer is entirely comfortable with the FTC's (or other agencies’) combining of prosecutory and adjudicatory functions. Whenever the same people who issued a complaint later decide whether it should be dismissed, concern about at least the appearance of fairness is inevitable.”46 A decade earlier, former FTC Commissioner Philip Elman noted that the “Commission plays many roles. It is investigator, prosecutor, judge, and jury.”47 He warned of “the subtle institutional influences which no law, regulations or codes of ethics can remove”—motivations that might include “fear that dismissal of its own complaint will be construed as an admission of costly error.”48
Turning to contemporary writings, in 2008, A. Douglas Melamed opined that FTC “Commissioners inherently and unavoidably lack the independence that we expect from adjudicative fact-finders.”49 Observing 60 FTC cases litigated in Part 3 between 1983 and 2007, he found that “the Commission won nearly 80 percent of the cases” and argued, “That result itself suggests a process that is tilted in favor of the Commission and against respondents.”50 He concluded that Part 3 “adjudication appears not to be sufficiently reliable to be counted upon to send the kind of sound, clear signals to the business community that competition policy requires.”51
In 2013, David Balto argued that the FTC's “administrative litigation role is particularly unsettling” because the agency “acts as both prosecutor and judge.”52 Noting the Commission's high-profile dismissal of the R.R. Donnelley merger litigation in 1995, he observed that since then “the FTC has always found a violation. In over 20 cases it has never found for the respondent and has reversed ALJ decisions that dismissed complaints. This FTC ‘winning streak’ is simply unprecedented.”53 It was not always that way. He noted, for instance, that “in the 1980s the Commission dismissed over 40 percent of its complaints on the merits.”54 Regardless of the root cause, he argues that the Commission's “almost two decade history of always ruling in its own favor creates a strong impression of unfairness.”55
What factors explain the FTC's consistently imposing liability in final decisions under Part 3? Perhaps the Commission brings stronger cases in-house than it used to do. Mr. Balto is dubious, however, because “some of its most important cases have been rejected by the appellate courts” and its Part 3 rulings “often do not stand up on appeal.”56 Other explanations include the FTC's “trying to establish new legal principles or explore new legal avenues” or that the ALJ's rulings are inadequate.57 Either way, Mr. Balto finds the Commission's consistent decisions favoring complaint counsel to be problematic.
Mr. Balto's voice is not the only one that questions the fairness of the FTC's Part 3 process. Former Commissioner Joshua Wright recently criticized Part 3 in connection with his larger view that the FTC eschews the “economic approach to antitrust enforcement.”58 He questions whether the Commission brings “unbiased rigor and expertise” through its administrative litigation.59 The culprit, he suggests, is that Part 3 gives the FTC “the option of playing prosecutor and judge. As prosecutor, the FTC Commissioners can vote to authorize its staff to initiate an administrative hearing before an administrative law judge. As judge, if the FTC loses in front of its own administrative law judge, the Commissioners themselves review the case.”60 He concludes that this ability “has been problematic and raises significant concerns that the deck is stacked against firms and in the agency's favor.”61
Empirical Studies on Administrative Litigation at the FTC
Due-process claims about the FTC's Part 3 process are mostly anecdotal or theoretical. That is, aside from the charge that appearances alone are problematic, the question whether the FTC neutrally reviews ALJ initial decisions is, at least in part, empirical. Hence, this part closes by addressing the limited empirical scholarship on Part 3. Does the evidence support charges of FTC bias in favor of complaint counsel? Given the relatively thin factual record to date, there is no clear answer to that question. This article, of course, hopes to contribute to a more complete discussion.
The most in-depth study to date appears to be Nicole Durkin's 2013 piece, Rates of Dismissal in FTC Competition Cases from 1950–2011 and Implications for Fairness.62 It pursued three substantive questions.
First, how often did the FTC dismiss Part 3 antitrust cases on appeal from the ALJ?63 Studying competition matters resolved on the merits between 1950 and 2011, Ms. Durkin found that the Commission dismissed 20 percent of its Part 3 antitrust cases in the 1950s, 15 percent in 1960s, 13 percent in the 1970s, 40 percent in the 1980s, 24 percent in the 1990s, and 0 percent in the 2000s.64 Viewing the 1980s as an outlier, she cautioned against extrapolating general principles from that decade. In her view, the dismissal rates “cannot be relied upon in concluding that FTC adjudication is free from bias resulting from the Commission's dual functions.”65 She acknowledged, however, that superior case selection may explain the FTC's tendency not to dismiss on the merits.66
Second, although the FTC is an independent agency, Ms. Durkin looked for political bias in Part 3 final orders.67 She scrutinized “straddle cases” where a majority Commission associated with one administration authorized an action and a majority Commission associated with the other party heard the appeal from the ALJ.68 Three presidential administrations fell within her review, specifically those of Eisenhower, JFK/LBJ, and Reagan.69 She found that, on average, the FTC was 1.6 times more likely to dismiss straddle cases than non-straddle matters.70 That phenomenon dissipated, however, between the Eisenhower and Reagan administrations.71 As with the first question she studied, she found the data ambiguous. Her takeaway point was similar: “this data does not offer support for the conclusion that the Commission enjoys the political independence that would make any bias resulting from its dual functions tolerable.”72
Finally, Ms. Durkin asked when the appellate courts were more likely to affirm FTC decisions than the final orders of other federal agencies.73 She observed that the courts reversed 29 percent of the FTC's antitrust final orders in 1970s, 64 percent of those cases in the 1980s, 20 percent in the 1990s, and 22 percent in the 2000s.74 Those percentages compared unfavorably to the judiciary's 13.06-percent reversal rate of all agency decisions between 1997 and 2011.75 She cautioned, however, that the FTC decisions are too few to employ statistically reliable cross-agency comparisons.76
Although a valuable contribution to the literature, Ms. Durkin's work is a starting point. She did not study consumer-protection matters, which provide a larger window into the Commission's Part 3 final decision-making.77 Nor did she track individual counts or ask whether the FTC's track record on appeal varied depending on whether the Commission affirmed or reversed the ALJ. This article explores all of those points and more, including the political constitution of the Commission at the time of authorizing a Part 3 complaint and reaching a final decision across the full temporal range of the study.
A study by Andrew Ewalt, and referenced by A. Douglas Melamed in 2008, examined a 25-year universe of Part 3 decisions between 1983 and 2007.78 It found that, of the 16 matters, “respondents did not win a single such case.”79 The unpublished study reportedly covered only “Sherman Act cases litigated and decided on the basis of disputed facts,”80 and we do not know which cases fall within and outside of that criterion.81 Given the study's limited coverage and unclear case selection, it is difficult to evaluate its results. Certainly, the study is not a substitute for this article's in-depth examination of 145 Part 3 antitrust and consumer-protection matters over a 40-year period.
In conclusion, there is a real need for more a more nuanced analysis of the FTC's Part 3 process. This article marks a step forward in empirical understanding of administrative litigation at the FTC. The next part takes advantage of the limited pool of recent cases to delve behind the numbers to evaluate the individual underlying matters. Doing so suggests that contemporary Part 3 practice reflects improved case selection, which casts a better light on the FTC's recent liability determinations, especially given the relatively small number of recent administrative cases.
DELVING FURTHER INTO PART 3: PAST, PRESENT, AND FUTURE
Claims that the FTC invariably imposes liability in Part 3 are understandable.82 They reflect the reality of recent administrative litigation. In the last decade, the Commission has found liability on at least one count in every appeal from an ALJ's initial decision. For firms accused of wrongdoing, it is easy to infer that liability is a foregone conclusion. Yet, I do not think that that inference is justified, for at least two reasons.
First, the last decade of Commission decisions provides only a small number of cases—twelve—making it an unreliable sample from which to draw broad inferences about the FTC's institutional design. In fact, looking at the merits of the underlying cases provides insight that data alone may overlook. Accordingly, the first subpart here analyzes the Commission's last ten years of Part 3 decisions to gauge their quality.
Second, if the design of Part 3 led the Commission always to impose liability after an initial “reason to believe” determination because of a systemic bias, then the numbers over time should bear out that phenomenon. But they do not. As Part II explained, the FTC dismissed 40 percent of its competition matters in Part 3 and did so for 24 percent of such cases on the merits. Coupled with the Commission's strengthening record on appeal before the courts over time, the more likely explanation is that the FTC is doing an increasingly good job at picking meritorious cases. The FTC's staff pursues exhaustive, data-rich investigations, coupled with empirical and theoretical analysis by the Bureau of Economics, which provides a separate recommendation. The Commission thoroughly scrutinizes all the recommendations, hears the parties’ arguments in detail, and often engages in robust deliberations among Commissioners. And an evaluation of the FTC's propensity to impose liability in Part 3 is incomplete without also considering the fact that the Commission frequently closes investigations rather than pursue litigation, whether administratively or in court.
We now turn to the last decade's worth of Part 3 decisions to discern whether criticism of the Commission's de facto 100 percent record in that period has accurately identified a systemic-bias problem.
Is the FTC a Kangaroo Court? Evaluating the Substantive Cases
The cost of a small dataset is that the ensuing statistics only allow a modest range of reasonable inferences. For instance, the fact that the Commission essentially found liability in all cases in the last decade does not necessarily imply deficient due process or suggest that the FTC has its thumb on the scales. Twelve cases simply are not enough to warrant such a conclusion. But a limited data pool also has benefits. It allows us to look beyond the numbers to evaluate the underlying cases on a substantive basis.
If the FTC's recent practice of siding with complaint counsel is problematic, it must be because the Commission failed to dismiss an unmeritorious action. We can test that proposition by looking at the relevant cases. Of the twelve study cases that were the subject of an FTC dismissal/liability decision in the last decade, seven were antitrust matters (Cabell, North Carolina State Board, McWane, Evanston, ProMedica, Polypore, and Realcomp II) and five were consumer-protection cases (LabMD, ECM BioFilms, Jerk, POM Wonderful, and Daniel Chapter One).83
Some of those enforcement actions stand among the FTC's most influential and pro-consumer victories to date. In North Carolina Dental, the FTC challenged a state regulatory board controlled by active market participants in the field of dentistry for prohibiting non-dentists from offering teeth-whitening services.84 The Commission secured an influential victory at the Supreme Court, which limited the scope of state-action immunity.85McWane was a rare example of a monopolization suit that was both meritorious and successful. In 2015, the Eleventh Circuit affirmed the FTC's liability finding based on the dominant firm's use of exclusive-dealing contracts to exclude its competitors.86
Evanston marked a turning point in antitrust scrutiny of healthcare mergers.87 After the FTC and DOJ lost seven hospital-merger challenges in a row, the FTC conducted a section 6(b) study to understand the competitive effects of such deals, finding effects to be more localized and geographic markets to be narrower than previously thought.88 One of four consummated transactions examined in that study was Evanston, which the FTC subsequently challenged through a ground-breaking Part 3 action. The FTC prevailed in every healthcare-merger challenge after Evanston until May 2016, when two district courts denied the agency's motions for preliminary injunction.89 The FTC obtained stays in both cases and has appealed to the relevant circuits. Regardless of its success in those appeals, the antitrust community widely regard Evanston as a ground-breaking case.90
The other three Part 3 antitrust matters from the last decade are also noteworthy. In ProMedica, the FTC challenged the hospital's consummated acquisition of St. Luke's in Part 3.91 The agency won an important appeal at the Sixth Circuit in 2014.92 In Polypore the FTC challenged a consummated deal in the battery-component market,93 and the appellate court affirmed.94 And in Realcomp II, the FTC challenged an entity comprised of real-estate agents and brokers that published a database of listings, for refusing to publish information regarding discount real-estate brokers.95 The Sixth Circuit affirmed.96
This consistently positive record in the appellate courts weakens the argument that the Commission finds liability in unmeritorious cases due to systemic bias. The same is true on the consumer-protection side. Although I disagreed with some aspects of two of those Commission decisions,97 the Commission appropriately found liability for the core behavior challenged in each case.
In Jerk, the Commission filed a Part 3 complaint against “Jerk LLC” and its founder, John Fanning, based on misrepresentations contained on “Jerk.com” about the source of its content and the benefits of subscribing as a member. FTC staff moved for summary decision, which a unanimous Commission granted.98 In May 2016, the First Circuit affirmed liability.99 In POM Wonderful, the D.C. Circuit upheld an FTC decision that a firm's disease-prevention claims pertaining to its pomegranate juice misled consumers.100 The court also rejected the petitioner's constitutional arguments, except insofar as the FTC's remedial order “imposed an across-the-board, two-RCT [randomized clinical trial] substantiation requirement for any future disease-related claims by petitioners.”101 The Supreme Court denied certiorari in May 2016.102 And in Daniel Chapter One, the D.C. Circuit affirmed an FTC decision that the company and its sole “overseer” and member deceived customers into thinking that its herb products variously inhibited tumor growth and treats cancers.103
The merits and the appellate success of these cases do not support a narrative that the Commission blindly supports ill-conceived cases because of systemic bias. To the contrary, they show a recent history of solid, well-supported enforcement actions in Part 3, even if one does not agree with every choice the Commission has made in each matter. The courts have affirmed all 8 of the Commission liability decisions that have been appealed.104 Trying to use the last decade as evidence of the FTC's blindly siding with complaint counsel simply does not square with an evaluation of the underlying matters.
Indeed, to find Part 3 antitrust liability decisions that succumbed on appeal, one must reach back to Rambus and Schering-Plough.105 Those cases warrant mention, if only because critics of the FTC's Part 3 record have focused on them in the past. For example, in 2013, one observer wrote, “recently the appellate courts have been critical of the FTC's decision-making where it has substituted its fact finding for the ALJ” and illustrated his point with the Schering-Plough and Rambus cases.106 In criticizing the FTC's Part 3 process, others have made a similar point.107 In my view, however, Schering-Plough and Rambus are poor support for these critiques.
The weakness of these claims is most evident in Schering-Plough. There, the Eleventh Circuit adopted a scope-of-the-patent test for antitrust liability governing pay-for-delay agreements in the pharmaceutical sector.108 Finding that the branded-drug manufacturer's patent lawsuit could have excluded the allegedly infringing generic, the court found that Schering-Plough's agreement to settle the lawsuit by paying the generic to abandon its validity challenge fell within the patent's scope and was therefore lawful.109 That ruling marked a serious setback for the FTC's antitrust-enforcement agenda in the life sciences, but it was not the end of the story. In 2013, the Supreme Court rejected the scope-of-the-patent test and upheld the FTC's position that such agreements may violate the antitrust laws. Far from being evidence of a poor case selection and a deficient administrative process, Schering-Plough is a positive example of the FTC's pursuing the development of antitrust law through Part 3.
Rambus raises more interesting questions.110 The matter challenged alleged holdup by Rambus of JEDEC, a microelectronics standard-setting organization, by concealing and later enforcing its standard-essential patents (SEPs) against firms implementing JEDEC standards. The antitrust questions that emerge from the use of SEPs in industries subject to voluntary standards are among the most difficult in the field today. For the last several years, leading experts have debated the theoretical and empirical support for holdup and holdout by owners and users of proprietary technology.111 In its 2006 decision in Rambus, a unanimous Commission imposed liability in Part 3, finding that—due to the firm's deception—JEDEC lost the opportunity either to design a standard without Rambus's technology or to secure a commitment from Rambus to license its technology on reasonable and nondiscriminatory (RAND) terms.112
The D.C. Circuit granted the petition for review.113 It assumed, without deciding, that tricking an SSO into adopting one's technology over a substitute technology would harm competition.114 There was insufficient evidence, however, that JEDEC would have chosen a different technology had Rambus disclosed its patents. Hence, liability for monopolization could exist only if evading a RAND-licensing promise was anticompetitive. The D.C. Circuit held that it was not because the deception did not impair a rival or otherwise harm the competitive process.115 It merely resulted in a higher price, which under the court's reading of NYNEX was not actionable anticompetitive conduct.116
It is clear that Rambus raises difficult issues at the cutting edge of antitrust law. But it is in complex cases where Part 3 has its greatest value. This brings me to a nuanced point. Regardless of whether one agrees with the FTC's decision in Rambus, it is scant evidence of a predisposition by the Commission stubbornly to stick to its prior position in voting out a complaint. And even if the Commission erred in holding as it did—a question that the antitrust bar still debates—Rambus is not a liability decision that plausibly indicates bias or prejudgment.
Finding the Value in Administrative Litigation: Time to Develop the Law
As my account of recent Part 3 cases reveals, the Commission has recently employed its administrative process rather well. It halted and reversed a losing streak in healthcare-merger enforcement with its Part 3 action in Evanston.117 The FTC magnified that success in ProMedica.118 It brought a rare modern example of a successful monopolization action in McWane.119 It fought an early battle in an ultimately victorious campaign against pay-for-delay agreements in Schering-Plough.120 And it persuaded the Supreme Court to narrow the state-action immunity doctrine in North Carolina Dental.121 Part 3 underlay each of those enforcement actions.
As a former FTC Commissioner observed 45 years ago, the agency “was conceived not as a prosecutorial or enforcement body, but as an expert administrative tribunal vested with the responsibility for developing an enlightened antitrust policy and given the tools to carry out that task.”122 It should not be controversial to suggest that the FTC as an institution123 has greater substantive expertise in antitrust law and economics than many—though certainly not all—federal district judges.124 In modern practice, the FTC has effectively deployed its administrative-litigation process to develop the law for consummated deals. In clear-cut conduct matters, district court is usually the preferable forum. But with new questions continuing to arise, especially—but by no means only—in patent-related cases, Part 3 is a unique asset for contemporary antitrust policy. For that reason, I have generally favored bringing difficult conduct matters in Part 3, rather than in federal court. Thus, for example, I dissented from a recent Commission vote to litigate a pay-for-delay case in federal court, instead of through administrative adjudication.125
In short, frontier antitrust cases are what Congress designed Part 3 to tackle. Not everyone agrees, however, that the FTC is endowed with unique substantive expertise vis-à-vis the courts. Wright and Diveley question, for example, whether the FTC outperforms district courts in antitrust cases.126 And indeed former-Commissioner Wright has previously found evidence that judges do struggle with antitrust matters.127 Thus, the authors’ findings about the FTC's relative performance warrant attention.
Wright and Diveley found that “cases decided by the Commission are 14 per cent more likely to be appealed than are cases decided by Article II[I] judges.”128 They regard the decision to appeal as the principal indicator of the quality of the underlying decision.129 The theory is that, by appealing, a party credibly signals a lack of faith in the integrity of the underlying order. Wright and Diveley also determined that “Commission decisions are reversed 20 per cent of the time and decisions by Article III judges are reversed only 5 per cent of the time.”130 Subsequently running a regression, they “consistently observe higher appeal and reversal rates for the Commission that are robust to controls for type of case, time trends, and a variety of robustness checks designed to control for unobservable differences in cases brought through administrative litigation rather than in federal district court.”131
Although welcoming the study as the start of much-needed empirical work in this field, I take the authors’ findings with a grain of salt. Two considerations, in particular, warrant caution in inferring that federal district courts typically reach superior antitrust decisions. First, it is unclear whether the quality of FTC determinations in the 1970s and 1980s—during a time of seismic change in antitrust law, when the Commission brought and lost many more Part 3 actions than it does today—is representative of the sophistication of the agency's decision-making today. As explained above, the FTC has fared extremely well on appeal in the last two decades and a close look at the cases taken in the last ten years reveal consistent merit. By contrast, the FTC won antitrust appeals in just 37 percent of cases from 1977 to 1986 and 34 percent the following decade. Summing the data on antitrust Part 3 matters from 1976 to 2010, as the authors did, may not accurately convey the reality of the modern Commission. It is possible and indeed likely, for instance, that the FTC's performance was historically less strong than it is today. Certainly, there is a view that today's agency is a more sophisticated entity than it was before the Chicago revolution took hold.
Second, it is unclear whether a party's decision to appeal is a reliable proxy of the quality of underlying decision, and especially when using those choices to compare two dissimilar contexts. Based on criticism already discussed, there is likely a perception—warranted or not—that the FTC will find liability in a Part 3 proceeding. Companies that litigate through the ALJ trial and appeal to the Commission, may surmise that the petition for review to a U.S. Court of Appeals is where they should fight the battle. Put differently, it would not be surprising if respondents are more likely to appeal from an FTC final decision than the average disappointed party is in a federal-district-court proceeding. Finally, it is not at all obvious that the Commission and district courts hear comparable antitrust matters.
In conclusion, there is a need for more nuanced analysis of the institutional expertise of the antitrust agencies vis-à-vis district court. Indeed, the reality of the FTC's modern appellate record in Part 3—winning 100 percent of all antitrust appeals from Part 3 in the last decade and 71 perceent the decade before—, the impact of Part 3 decisions on recent developments in antitrust law, and the mixed record of recent district-court opinions in difficult competition matters raise substantial challenges to other narratives.132
Although criticized in some quarters, administrative adjudication is a useful tool for advancing competition policy. Despite a checkered history, the FTC recently has used its Part 3 authority well, transforming the law on healthcare mergers, pay-for-delay agreements, and state-action immunity in the last 15 years. The ability to bring to bear institutional expertise in antitrust law and economics on the day's most difficult competition questions has clear value. That is why I have favored bringing more complex antitrust cases into Part 3.133
But any system of law enforcement only works if it incorporates safeguards that ensure accurate decision-making, both by giving accused entities a fair hearing and through a review process that holds the agency to its proof. It is here that Part 3 stands accused of deficient due-process protections. Certainly, firms enjoy a number of rights when defending a Part 3 action. They present their views of the law, facts, and economics to the Commission before it determines whether there is “reason to believe” that a violation occurred and that an action would be in the public interest. Afterward, they litigate a trial before an ALJ, who is independent of the Commission, and can cross-examine witnesses and experts. Meanwhile, the Commissioners remain cordoned off from the case until the parties argue on appeal from the ALJ. The Commission holds oral argument before issuing a final decision that, if it imposes liability, respondents can usually appeal to any U.S. Court of Appeals in the country. That choice of forum is an unusual advantage.
Nevertheless, skeptics are dissatisfied. They claim that, by voting out a complaint in the first instance and then later hearing an appeal from the ALJ, the Commission inevitably compromises its independence. In its most hostile expression, the charge accuses the FTC of being a kangaroo court.
There are two dimensions to that criticism. The first one concerns appearances, such that an agency acting as both prosecutor and adjudicator invites questions about partiality. The second issue, however, is more substantive and ultimately more important. Does the FTC in fact make up its mind at the case's inception and put its thumb on complaint counsel's side of the scales on appeal? Or does it take its independent, appellate function seriously and scrutinize the record de novo with an open mind to dismissing should the merits warrant it?
Some commentators have offered, as evidence of a faulty system, the fact that the Commission often sides with staff on appeal from the ALJ in Part 3 matters. Skeptics infer that a high success rate for complaint counsel in final FTC decisions reflects a foregone conclusion, at worst, or an unconscious bias in favor of liability, at best. This article has sought to address and to supplement the empirical literature on which critics base those claims. Its findings are notable. Despite recurring claims that the FTC “always” sides with complaint counsel, or rules for itself in 100 percent of cases, the record over the last four decades tells a different story.
Statistics can be misleading when one disregards the actual phenomena being studied. It is true that the FTC has found liability in almost every Part 3 case in the last decade. But the appellate record suggests that those matters had merit, meaning that the law and facts likely warranted liability. Indeed, the federal courts have affirmed 100 percent of the FTC administrative final orders issued in the last decade and challenged on appeal. The FTC's track record in the last 10 years is impressive. Influential wins at the Supreme Court during that period only enhance the point.134 To get a clearer picture of whether the Commission dogmatically imposes liability, regardless of the merits or the ALJ's initial decision, one must go back further in time.
This article goes back four decades. 145 administrative cases in that time show that the Commission does not always rule for complaint counsel. Indeed, the agency dismissed 29 percent of the Part 3 matters it previously authorized. It is difficult to reconcile that fact with the claim that the FTC's institutional design renders it systemically predisposed to impose liability. Indeed, in competition matters, the Commission rejected liability even more often, doing so in 40 percent of Part 3 cases. And although one can legitimately debate the implications from this statistic, the FTC dismissed 16 percent of all appeals (both antitrust and consumer protection) from the ALJ's initial decision on the merits.
What is also striking from the data is that, as time went on, the FTC brings fewer Part 3 cases, is less likely to dismiss, and is also more likely to prevail on appeal. Although commentators will discuss the implications of the data, to my mind the emerging picture is largely one of improved case selection, likely in response to changes in antitrust law that demanded more economic rigor.
VI. APPENDIX: METHODOLOGY
This study tracks 145 Part 3 cases with Commission decisions on or after January 20, 1977, the beginning of the Carter presidency, to Commission decisions before August 1, 2016. A relevant question is why the past four decades. The study could have gone back farther, but 1977 provides a good approximate starting point of modern antitrust analysis taking hold with a growing focus at the Commission and within the antitrust bar on applying rigorous economic analysis to antitrust cases, as well as a more disciplined focus on cases likely to pose antitrust harm affecting consumer welfare.135 Additionally, until the 1960s, the Commission's adjudicative procedures were significantly different, with discovery largely consolidated with the hearing, which occurred in different locations and over an extended time, and diverging in other ways from federal procedure.136 Alternatively, if the study only included the last decade of Commission decisions, it would have examined only 12 Commission decisions, a very limited group of cases that would reveal little about trends or the intrinsic characteristics of the Commission's adjudicative process and outcomes.
The next step involved determining which cases would provide the most accurate view of the Commission's record. This question presented complex issues. As a default, I included all matters in which the Commission issued a complaint and subsequently issued a Commission decision. In most of the matters with Commission decisions included in the study the ALJ also issued a decision, but not in all (130 out of 145).
From the baseline of including all matters before the Commission in Part 3, I excluded matters: (1) in which cases settled after the filing of a complaint and before the ALJ decision, which involved more than 40 cases, or (2) settled after the ALJ decision but before a Commission decision, a significantly smaller number. I also excluded cases that (3) involved an ALJ decision that was not timely appealed and thus issued as a final order of the Commission, (4) the Commission accepted or rejected proposed consent orders, (5) the Commission dismissed a complaint following denial of a preliminary injunction in federal court, and (6) the Commission dismissed a complaint after the parties abandoned a transaction. In many of these circumstances, the records show some form of a Commission decision but the decision is not particularly meaningful.
From a more technical perspective, I also excluded (7) decisions regarding requests to reopen or modify existing orders; (8) dismissals of related matters when the primary matter already was included in the dataset of matters; and (9) interim decisions that subsequent, final Commission decisions in the same matter superseded (and which the dataset included).
The study characterizes outcomes at the ALJ and Commission decision stages as liability or dismissal. At the Commission decision stage, it also records whether the Commission trimmed or expanded charges or allegations, or respondent, and whether its decision affirmed or reversed the ALJ. If dismissed by the Commission, it records whether the dismissal was on the merits or for a different reason such as a change of law, change of fact, or no longer in the public interest. The database includes any appeal of a Commission decision, the federal appellate court's decision, and any subsequent Supreme Court record. For each appellate ruling, the database contains the date and court of appeals; whether the appeals court affirmed, reversed or had a mixed decision; and whether it was a liability, remedy, or procedural ruling.
Beyond outcome, the study tracks many variables to provide relevant and more granular insights, such as how frequently the same majority of Commissioners that voted a complaint was still present and voted on the Commission decision, and whether the Commission was more or less likely to find liability in this situation. At the time of the complaint, the study tracks the date/decade; whether it was a competition (merger, non-merger, or both) or consumer protection matter; the charges/counts; the Commissioners voting, the vote count, and whether it was unanimous; and the party of the majority of Commissioners voting. At the time of the Commission decision, beyond outcome variables, the database contains information on the date/decade of the vote; Commission composition and vote count (notation of dissents); whether the Commission trimmed counts, allegations, or respondents relative to the ALJ decision; the number of Commissioners that left the Commission between complaint and decision votes; whether a complaint majority still existed among Commissioners at time of decision; whether the vote was unanimous; and the party of the majority voting, as well as the party of those in the decision majority. The database also tracks the times between complaint, ALJ decision, and Commission decision.
For sources for more recent Commission matters, generally those of the past two decades, I relied on publicly available records on the FTC web site, records available through Lexis and Westlaw, and Commission records. For older records, particularly the identity and vote of Commissioners voting on complaints and decisions, I relied upon information contained in the Commission's “Minutes” records.