Merger Guidelines and the Limits of Our Understanding

But, as the recent emergence of antitrust populism shows, advocates for stricter enforcement can also embrace a more open-ended approach to merger analysis; see, e.g., Khan (2017).

See Arrow-Hart (1934) (acquisition of assets not subject to Clayton 7); and Western Meat (1926) (if corporation acquired the assets of a firm whose stock it had acquired, then there was no longer any Clayton Act jurisdiction).

The two largest firms would have just under 50% of the total volume and the six largest in combination held slightly over 70% of the volume in the year prior to the merger (Id. at 461, n. 11).

In evaluating bottle and can competition, the product market question is whether plastic or paper were also part of the relevant container market. The scope of the product market relevant to this case highlights the fact that in some merger analyses market definition—either product market or geographic market—can be quite contentious and potentially ambiguous (Id. at 447–48). The case was ultimately remanded for further evaluation of the market definition issue.

Two other decisions in this period confirmed that overcoming the presumption required convincing evidence that either the buyer was not a plausible competitor in the market of the target or the target was not and would not compete with the acquirer because they were under common control. Thus, the successful rebuttals all involved proof that the combination would not create the kind of market structure that would impair competition. See Marine Bancorporation (1974) (no potential competition possible because the acquirer could not enter or expand in the target’s market in the absence of the merger in question); and Citizens and Southern (1975) (bank established affiliates to avoid restrictive state law; when law changed, no competitive harm would result from consolidation with the parent bank).

The different shares and market structures in Philadelphia National Bank and Continental Can further illustrate that the presumption should vary with market share and market structure.

A market HHI is the sum of the squares of the market shares of all participants, and the change from a merger is measured by two times the product of the market shares of the two firms.

I confess to having been unenthusiastic about this change when Baxter imposed it, having learned merger analysis with the use of the CR4. Over time I have concluded that this was an important improvement in the description of market concentration and changes in that concentration. Thus, not all change in the Guidelines is necessarily undesirable.

The premise that is set forth—most notably by Easterbrook (1984) is that Type I error (over-enforcement) results in long-term or permanent harm to the economy, while the dynamics of free markets will quickly obliterate any Type II error (under-enforcement).

Critics often point to decisions—such as Brown Shoe and Von’s Grocery—as examples. These decisions pre-date the 1968 Guidelines. Moreover, when one looks at the markets involved—shoes and grocery retailing in Southern California—it is impossible to discern any cognizable harm. The Brown Shoe company itself still exists under a different name, with total sales of nearly $2.6 billion in 2017 and net profits that exceeded $65 million; and Kinney Shoes is still in the shoe retailing business and competes with Brown—although it too trades under a different name.

The Von’s story is more complex, but basically the company survived, grew, merged, and ultimately was a component of Albertson’s. Shopping Bag, the acquired firm, was divested to a new entrant into the market that expanded as well. The unrestrained combinations in that market resulted in Albertson’s and Safeway combining subject to an FTC order that required the divestiture of many stores to retain local market competition; but the buyer went bankrupt, which left the market with high concentration. See Albertson’s (2015).

The GE-Honeywell merger that was blocked by the EU is a contemporary transaction that is often described in terms that suggest error. The European Commission blocked the merger in 2001 based on a conglomerate effects theory. The reviewing court rejected the Commission’s primary theory but affirmed the decision based on the overall dominance of aircraft engines (Buck 2005). This basis resonates with the concerns of the Supreme Court in the GM-DuPont case in 1957. Despite predictions of doom, both companies have survived and prospered. Moreover, there is no evidence that aircraft assemblers—including Boeing and Airbus—have suffered any cognizable disadvantage because of the lack of a consolidated aeronautic instrument package that was to be the merger’s contribution to efficiency.

Mueller and Sirower (2003, p. 374) (citing five studies “that suggest that acquisitions significantly impair the long-term profitability or market share of the acquired businesses.”); that study also found that there was a strong tendency to overpay for acquisitions: “several of our findings actually imply that mergers destroy more of the value of the bidding firms than is paid as premium to the target” (Id. at 380, 388).

Chasan and Murphy (2013) (reporting a study by Duff & Phelps); see also Lipin and Deogun (2000) (reporting that Salomon Smith Barney’s analysis of major mergers showed that the acquirers “on average underperformed” measured by both the S&P 500 stock index and their peer group); Bahree (1999) (“Megamergers often flop, and oil mergers especially are prone to failure.”).

Weiss (1989, p. 267, Table 13.1) found that 62.8% of the 121 studies showed significant positive price effects that resulted from concentration and another 24.8% had non-significant positive effects. Only 3.3% of the studies had significant negative correlations of price and concentration, while another 9.1% had non-significant negative effects.

Alternative explanations exist for this phenomenon in that the market might have bid up the value of the enterprise, thus capitalizing the excess profit. In addition, accounting profits bear only a passing relationship to actual economic profits.

See, e.g., Schmalensee (1989, p. 988) (“In cross-section comparisons involving markets in the same industry, seller concentration is positively related to the level of price.”); Bresnahan and Suslow (1989) (for every 100 point increase in the HHI index that resulted from mergers in the North American aluminum industry in the 1960s and 1970s, there would be a 2.7% price increase during cyclical downturns); Martin (2002, pp. 217–20) reports several additional studies with the same result: Concentration and price are correlated. See also Baker (1999) (predicting price effects of proposed merger Staples/Office Depot merger).

Ghosh (2001, p. 151) (finding generally “no evidence that operating performance improves following acquisitions”); Moss (2013) (airlines); Cummins et al. (1999, p. 327) (“larger [life insurance] firms generally are found to exhibit decreasing returns to sale”); Milbourn et al. (1999, pp. 197, 198) (banking: “little or no improvement in cost efficiency” and “there is also a lack of empirical evidence that expansion of scope in banking has been beneficial.”).

Id. at 113, tbl. 7.3. But five of the negatives were in industries (railroads and airlines), where the agencies had no authority to challenge the actual merger (Id. at 115).

In the case of structural remedies, post-merger prices increased on average more than 6% and conduct remedies resulted in average increases of nearly 13% (Id. at 120).

Morton et al. (2015) (substantial increase in air fares following the American Airlines merger with US Airways, as well as overall evidence of increased margins that resulted from higher prices and decreasing costs such as fuel). Other scholarly studies confirm that air fares have increased significantly because of increased concentration in the industry; see, e.g., Peters (2006) (study of six airline mergers showing average post-merger price increases of 8–30%); Brueckner et al. (2013) (empirical data show that price competition among legacy carriers is weak, but low-cost airlines do stimulate price competition); Kwoka et al. (2013) (finding competition among legacy carriers weak and prices high unless there is competition from low-cost carriers).

DeGraba (2009) (a rail line outage may have contributed to these price effects, but the observed prices seem inconsistent with that as the primary explanation for higher prices).

See also, Anheuser Busch (2013, Compl. pp. 2–3, 18).

See Chen (2008, pp. 56–61, 71–77) (describing major consolidation mergers in telecommunications; FCC approval conditioned on acceptance of more open access). The relatively recent emergence of T-Mobile as a price competitive market participant in wireless communication was therefore news (Manjoo 2014, B1; Lohr 2011).

See Pear et al. (2015, A12) (“In most of … [New Hampshire] the number of insurers is increasing to five in 2015, from just one this year [2014]. Prices for the lowest-cost silver plan have fallen by 14%.”).

See Id. (presenting a model that demonstrates that monopolists will expend almost all of the gains to avoid competition thus eliminating any “excess” profit).

Another false economy comes from exploiting enhanced buyer power to drive down the price of inputs. Such buyer power can offset an upstream oligopoly’s seller power, but it is often used to exploit powerless suppliers; see Domina and Taylor (2010). A study that was done for the U.S. Department of Agriculture on hog markets found that there was significant evidence of monopsony buying power viewing the market nationally (GIPSA 2007, ES-10). Despite this evidence, the Department of Justice allowed a further significant increase in national and regional concentration by permitting Smithfield to acquire Premium Standard Brands; see DOJ (2007); see also Strom (2015, B1) (reporting major food processors use their buyer power to force suppliers to wait much longer for payment thereby forcing the seller to finance the buyer’s operating expenses).

I am indebted to George Hay for reminding me of these additional adverse impacts of the contemporary merger enforcement system.

In statistics, such a one-tailed test is used when the value of interest is in only one side of the distribution; see Mendenhall (1971, pp. 163–167).

See Kwoka and Gu (2015, p. 541) (“[W]e find that agency decisions do in fact achieve a better balance between errors of commission and errors of omission than do either event studies or purely structural criteria, demonstrating the added value of the merger-specific process of investigation that they conduct.”).

If we convert HHI changes to market share, this standard would only prohibit mergers if the two firms each had 10% or more of the market or one had at least 20% and the other 5% or more.

In the range between 1500 and 2000, a merger that involves firms that each have more than 8% of the market would be presumptively illegal; or if one had 15%, it could not acquire a competitor that has more than 5%. Above 2000 HHI, firms with mid-range market shares (5–10%) are unlikely to exist given the overall HHI; hence the standard would bar mergers in 2000–2500 level by a firm with 25% of the market with any other competitor that has more than 2%; while above 2500, the implication is that such a firm could not acquire another firm with more than 1% of the market.