Institutional Factors Contributing to the Under-Enforcement of Merger Law

A past chair of the ABA’s Antitrust Law Section observed “a sense even in many of the mainstream segments of the antitrust community that there has been a degree of under-enforcement of antitrust laws especially in merger control,” and a former Antitrust Division chief economist has written that “[t]he clearest area where antitrust enforcement has been overly lax is the treatment of mergers.” Federal Trade Commissioners, authors of merger retrospectives, prominent antitrust scholars, and a former Director of the FTC’s Bureau of Economics, among others, have voiced similar concerns.

This article focuses on the process by which the antitrust agencies review mergers as a source of weakness in merger enforcement. We identify three aspects of that process that have contributed to a weakening of merger control in the United States: (1) unreported transactions that go unreviewed; (2) transactions that are reported but overlooked, that is, transactions presenting significant antitrust issues that the agencies fail to detect because the issue is not apparent from the parties’ Hart-Scott-Rodino filings or readily available public information; and (3) transactions that should be challenged but are not because the agencies do not have the resources to do so. We propose modest changes to remedy each of these problems.

Most of what has been written about under-enforcement of merger law addresses the standards the agencies and courts use to review potentially anticompetitive mergers. One view is that the structural presumption––the presumption that mergers between competitors that significantly increase concentration in a properly defined market are likely to harm competition––has become weaker as courts place more emphasis on direct evidence of competitive harm and less emphasis on market concentration, making it more difficult for the government to block mergers. Others have criticized the emphasis that agencies and courts place on consumer welfare and economic efficiency, arguing that broader tests that account for a variety of other factors should be considered as well. Another view is that merger reviews have become so complicated that district court judges do not understand them, which favors the private parties and makes the reviewing agency more reluctant to challenge a merger unless it is highly confident of prevailing in court. Most likely, some combination of these factors is at play.

While these are important issues, the institutional aspects of the merger review process that we discuss have largely been overlooked as contributing factors to the under-enforcement of merger law. Even if changes were made to substantive merger law that led to more successful government challenges to mergers, these changes would have little or no effect on a large number of potentially anticompetitive mergers. The agencies exercise very little scrutiny over transactions that fall below the Hart-Scott-Rodino reporting thresholds, so it is difficult to see how changes in the substantive standards would have any significant effect on those transactions. Similarly, changing legal standards would not alter the number of potentially anticompetitive transactions that are inadvertently overlooked. And most importantly, changing legal standards would not enable the agencies to issue more Second Requests. Because the agencies are capacity constrained, they generally do not issue Second Requests except in those cases in which they conclude after their initial 30-day review that a challenge is highly likely, thus allowing closer cases to slip through, since transactions that do not receive a Second Request are not challenged in court and are not subject to consent decrees. For these reasons, the changes in substantive legal standards many scholars have advocated would only solve part of the problem of under-enforcement.

Transactions Below the HSR Reporting Thresholds

When the statutory size-of-person and size-of-transaction tests are met, the Hart-Scott-Rodino Act requires parties to a proposed acquisition of voting securities, assets, or other non-corporate interests to file notification with the FTC and the Department of Justice prior to consummation. The parties must then observe a 30-day waiting period (15 days for cash tender offers and bankruptcy sales) before they may consummate the transaction. If one of the agencies issues a Second Request—a detailed set of interrogatories combined with requests for a large amount of documents and data—the waiting period is extended until 30 days after the parties substantially comply with the Second Request, something that in practice can take several months.

The legislative history of the HSR Act makes clear that its purpose is to provide the antitrust agencies with the opportunity to review and, if appropriate, challenge acquisitions prior to consummation. This avoids the problem of “unscrambling the eggs”––constructing effective post-acquisition relief is more difficult if the parties have already merged. It also gives the agencies greater ability to prevent consumer harm by challenging acquisitions before the harm occurs.

When the HSR size-of-transaction threshold was raised from $15 million to $50 million in 2001, it resulted in a sharp decrease in the number of transactions that triggered HSR filings. In the five fiscal years prior to 2001, the agencies received an average of 3,993 filings per year. After the thresholds were raised, filings dropped by over 50 percent to an average of only 1,414 filings per year for the five fiscal years from 2002 to 2006.

While it is impossible to know for certain how many more post-2001 transactions would have been challenged had the filing threshold not been increased, some approximation is possible. We can estimate the number of transactions valued below $50 million that were challenged when such transactions were reportable: in the five fiscal years prior to the 2001 threshold change, about sixteen such transactions were challenged each year.

Following the change, the agencies have challenged a much smaller number of now non-­reportable transactions. From 2001 to 2010 the agencies challenged only 22 non-reportable transactions––an average of little over two per year––and from 2011 to 2020 they challenged 19, slightly fewer than two per year. This suggests that the agencies have recently challenged a much smaller number of transactions below $50 million than they did prior to the change.

The agencies monitor newspapers and industry publications for announcements about below-the-threshold transactions, and they investigate complaints they receive from customers and competitors. But inevitably many problematic non-reportable transactions escape their notice. There is no requirement that acquisitions by privately held companies be publicly announced. Public corporations are required to disclose transactions only if they are material, and many below-the-threshold transactions are not material to large public companies. Acquiring companies are well aware that post-closing conduct can affect the likelihood of an investigation, so they tend to avoid actions that may cause customers to complain to the agencies.

There are two additional reasons why the change in the size-of-transaction threshold has led to more anticompetitive mergers each year than the roughly 16 transactions per year that were challenged before the Amendment.

First, the Amendment provided that beginning in 2004, all dollar-based thresholds in the HSR Act would be indexed to changes in the gross national product. As a result, the size-of-transaction test is now $94 million; the threshold would be $76.4 million had it been indexed to the Consumer Price Index. It is by no means a perfect comparison, but it is telling that in the five years before the change in the size-of-transaction threshold, 38 percent of Second Requests were issued in reported transactions below $100 million.

Second, after the first full year of operation of the HSR premerger notification program, the FTC reported to Congress in 1979 that it is “possible that the inception of the premerger notification program itself has deterred companies from entering into merger agreements which might violate the antitrust laws because of the parties’ awareness that their transactions will be subjected to more careful scrutiny than in the past.” There is now evidence that the converse is true: raising the size-of-transaction threshold appears to have led to an increase in horizontal mergers below the filing threshold in direct response to a reduction in enforcement activity. This is not a surprising result: it is common sense that as the probability of being subject to an enforcement action falls, the likelihood that direct competitors will merge increases. Indeed, it appears the current situation may be the exact outcome that Attorney General Janet Reno cautioned Congress about during considerations on raising the threshold, noting that if an increase

did in fact cause anticompetitive transactions to escape HSR review, antitrust enforcers would spend additional resources to find out about—and then investigate—those transactions. It may well be significantly more costly—not only to the government but to the parties and, most importantly, consumers—
to investigate and remedy those anticompetitive transactions than it would be utilizing HSR as a premerger investigative tool.

Transactions under $94 million that can have anticompetitive effects include, for example, transactions in specialized national markets, such as medical devices, pharmaceuticals, and semiconductors; transactions in local markets, such as hospitals and clinics, grocery stores, and funeral homes; and so-called killer acquisitions, where incumbent firms acquire innovative firms to shut them down and preempt future competition from them—or to internalize the value of their innovations. Earlier this year, the FTC announced a retrospective study of all non-­reportable acquisitions by five leading technology companies—Alphabet (including Google), Amazon, Apple, Facebook, and Microsoft—during the decade beginning 2010, noting the study would aid its considerations of “whether additional transactions should be subject to premerger notification requirements.”

An article about killer acquisitions shows that in the pharmaceutical industry, incumbents’ acquisitions of targets with drug development projects in the same therapeutic class that use the same mechanism of action, which the authors call overlapping acquisitions, cluster just below the HSR reporting threshold, while non-overlapping transactions do not, suggesting that incumbents intentionally conduct more killer acquisitions in non-reportable transactions where they can expect to avoid antitrust scrutiny. Another recent paper found that more than 80 percent of acquisitions of kidney dialysis facilities that would result in local monopolies were blocked when they are part of reportable transactions, but less than 2 percent were blocked when the transactions were not reportable.

While it may be politically impossible to turn back the clock and reduce the size-of-transaction threshold, at a minimum the current threshold should be frozen instead of allowing it to continue to climb, and the agencies should dramatically increase their efforts to identify and challenge transactions below the threshold that violate Section 7.

A better solution would be to amend the HSR Act to lower the size-of-transaction threshold to $50 million but to have the FTC amend its HSR regulations to permit a new short-form filing for transactions between $50 and $100 million. The 2001 Amendment was enacted because the thresholds had not been adjusted for 25 years and there was a feeling that the premerger notification process was imposing undue delay and burden on small transactions. Allowing a short-form filing for smaller transactions would balance the need for enforcement against the burden on merging parties in those smaller deals if the HSR threshold were reduced.

For those smaller transactions below $100 million, the filing fee could be eliminated or substantially reduced, and the parties would only have to fill out a simplified Notification and Report Form that calls for the most basic information. This might include a description of the transaction, the location of the parties’ facilities, and the North American Industrial Classification System (NAICS) codes in which the parties derived revenues in their most recent fiscal year. The initial waiting period for these smaller transactions could also be reduced from 30 days to 15 days, which could be extended to for an additional 30 days if the agencies issue a voluntary request for more information. This would provide the agencies with notice and an opportunity to investigate transactions that may impermissibly reduce competition while imposing a minimal burden on the merging parties.

Problematic Transactions that Are Reported but Overlooked

Although it does not happen often, occasionally the parties will submit HSR filings for a transaction that raises significant antitrust issues under the existing merger guidelines and enforcement policies, yet the waiting period will expire or the parties will be granted early termination without the agencies contacting the parties or conducting an investigation. As we discuss below, the agencies are capacity constrained. However, the scenario we just described does not appear to be driven by capacity constraints. If it were, it seems likely the staff would issue access letters to gather more information about the transaction and the harm it could potentially cause in order to decide whether it merits a Second Request. Instead, we think that occasionally the staff fails to detect a problematic transaction because there is nothing in the Hart-Scott-Rodino filings or publicly available information suggesting the transaction is problematic.

In our experience, the agencies do not fail to scrutinize large transactions or transactions in key sectors, such as food, energy, and health care, but inevitably there are some transactions in other less visible industries presenting substantive issues that appear to go unnoticed by the agencies. Sometimes the agencies catch their mistakes and bring a post-closing challenge to a transaction that was reported, but that is very rare (only five times in past two decades).

One reason problematic transactions are overlooked is that the HSR Notification and Report Form does not directly ask whether the parties compete with each other. The two items on the Form that provide the most information relevant to antitrust analysis are Item 4(c), which requires the submission of all “studies, surveys, analyses and reports which were prepared by or for any officer(s) or director(s) . . . for the purpose of evaluating or analyzing the acquisition with respect to market shares, competition, competitors, markets, potential for sales growth or expansion into product or geographic markets” and Item 7, which requires parties to list any six-digit NAICS code in which both parties to the transaction derive revenues.

In contrast, Form CO, which is the form used for pre-merger filings with the European Commission and is the model for pre-merger filing forms used in many other jurisdictions, requires parties to provide detailed information about competition between the filing parties. This includes the requirement to state the economic rationale for the transaction (Section 3.2), define the relevant product and geographic markets and explain which are likely to be affected by the transaction (Section 6), and provide market share and capacity estimates (Section 7).

Standing alone, the HSR Notification and Report Form is often not a reliable guide to whether a transaction raises serious antitrust issues. First, well-counseled companies do not create bad Item 4(c) documents––that is, documents that convey the impression that a transaction raises antitrust issues. Second, Item 7 is not always a reliable source of information. Two competitors may legitimately report their revenues under different NAICS codes; in that case Item 7 will not disclose any overlap. Third, many of the six-digit NAICS codes encompass broad industry segments. This can make it difficult to discern how closely two companies compete. For example, two relatively small companies seeking to merge could both derive revenues in a NAICS code such as 332722, Bolt, Nut, Screw, Rivet and Washer Manufacturing, where numerous large competitors derive revenue and the merging parties’ revenue is therefore only a very small percentage of the total revenue all companies reported in that code. This could suggest to a busy agency staffer reviewing the filing that the transaction raises no issues. However, the merging parties could be close competitors in a niche with few other competitors, such as the supply of titanium fasteners to airframe manufacturers.

Of course, staff attorneys at the DOJ and FTC do not rely solely on HSR filings to decide whether a transaction warrants scrutiny. They typically review the merging parties’ websites and other publicly available information, rely on industry expertise they have developed, and in some cases may have access to relevant information from agency files on other transactions. Nevertheless, these sources may not always reveal antitrust problems in transactions that have them. The resulting error rate, even if already low, could be reduced by simple changes. Without creating the burdens Form CO imposes, the HSR Form could be revised to require the parties to a transaction to disclose whether they are, in fact, direct horizontal competitors. This could be accomplished with a simple question such as, “Do the parties to the transaction compete against each other?” If the parties answer yes, the form could require them to identify the products and geographic areas in which they compete.

Limited Resources

Although the agencies’ caseload has increased, their resources have not. In its most recent budget request the Antitrust Division stated that “between FY2013 and FY2019, the number of mergers the Division reviewed annually increased by more than 50 percent, from 1,326 in FY2013 to 2,091 in FY2019.” The Division also reported that “[t]he number of mergers per year valued at more than $1 billion has also increased dramatically, from 128 in FY2010, to 225 in FY2014, to 284 in FY2019.” Yet from FY2013 to FY2019 the total number of employees at the two agencies charged with enforcing the merger laws likely fell and almost certainly did not increase.

The Federal Trade Commission’s budget documents show how many employees are assigned to discrete tasks. In FY2013 a total of 209 fulltime equivalents (FTEs) were assigned to premerger notification, merger and joint venture enforcement, and merger and joint venture compliance. In FY2019 there were 211 FTEs assigned to the same tasks. The Antitrust Division does not provide similar detail. Instead, it reports that approximately 60 percent of its FTEs are assigned to civil enforcement and 40 percent to criminal enforcement. The number of civil enforcement FTEs includes employees assigned to merger enforcement, as well as those assigned to civil non-merger enforcement and those who work on both. It therefore overstates the resources available for merger enforcement and masks any shift of resources between merger enforcement and civil non-merger enforcement. In any event, the number of FTEs at the Antitrust Division devoted to civil enforcement dropped from 392 in FY2013 to 344 in FY2019.

It is striking that during a period when reported mergers increased by over 50 percent, the number of agency personnel devoted to merger control likely decreased, perhaps substantially at the Antitrust Division. But it is even more striking when one looks back a further ten years and sees that the resources available for merger enforcement in 2013 were already substantially less than they were in 2003, even though there were more reportable transactions in 2013.

Figure 1