VoxEU Column Competition Policy

The effect of mergers on consumer prices: evidence from five selected case studies

US antitrust authorities block very few mergers. This column presents estimates of the consumer price impact of five large mergers that were allowed. Prices increased, providing one piece of evidence to support those who say US antitrust authorities are too acquiescent.

More than a thousand merger requests are filed with U.S. antitrust authorities each year, but only a small number of these transactions are blocked or modified because of concerns that they might result in higher, anticompetitive consumer prices. Recent examples of highly debated proposed mergers include the combination of Delta and Northwest airlines and Microsoft’s bid for Yahoo!. Although very significant public and private resources are devoted to the administrative review of the potential anticompetitive effects of mergers before they are approved, there has been surprisingly little evaluation of whether the mergers that have been permitted are actually anticompetitive. Absent this information, it is impossible to determine whether government policies are either too stringent or too restrained or to make rational adjustments to these policies.

More generally, the value and effectiveness of antitrust policy itself is the subject of intense debate that must rely on very little factual information. Crandall and Winston (2003), for example, argue that antitrust policy has not been beneficial for consumers, while Baker (2003) argues to the contrary.

Which mergers would be anticompetitive?

In the late 1990s, the U.S. economy experienced the largest merger wave in its history with the number of merger filings more than doubling between 1994 and 1999. In recent research, we selected the consummated mergers to study from those that, based on the public record, appeared to be the most problematic from the antitrust agencies’ point of view: those most likely to result in anticompetitive price increases (Ashenfelter and Hosken 2008). The price increases from these mergers provide an upper bound on the price increases that other permitted mergers may have produced and a lower bound on the price increases that might otherwise have occurred in mergers that were blocked. We thus provide an indication of whether government merger policy may have been too hostile or too acquiescent. To identify potential mergers we searched the Merger Yearbook and identified five consummated mergers where the merging parties produced products that were substitutes in markets that appeared from press accounts to be somewhat concentrated. The five cases are Proctor and Gamble’s purchase of Tambrands (feminine hygiene products), Aurora Food’s purchase of Kraft’s Log Cabin breakfast syrup business, Pennzoil’s purchase of Quaker State motor oil, General Mills’ purchase of the branded cereal business of Ralcorp, and the merger of the distilled spirits businesses of Guinness and Grand Metropolitan.

Difference in differences

To control for other confounding factors that may also have changed at the time of the event such as the effect of possible changes in demand or costs on prices, we use a “difference-in-differences” estimation to measure the consumer price effect of mergers. We consider multiple control groups and windows of data surrounding the events we study. Our preferred control group consists of the private label products sold in each industry. The advantage of private label products as a control group is that they are likely to be distant substitutes to consumers for the higher quality branded products affected by the merger. Yet private label products should share many of the same inputs (with the exception of advertising) that are used to manufacture branded products. Assuming that these private label products are supplied perfectly elastically, the prices of these products should serve as good controls for costs, while at the same time they should be relatively unaffected by any anticompetitive price increase by the merging parties. Our preferred merger window contains a symmetric amount of data pre- and post-merger. By dropping the data within three months of the merger date, we avoid the issue of exactly when the firms start coordinating their pricing behaviour, leaving us with uncontaminated measures of the firms’ pre- and post-merger pricing.

Results

Our empirical results indicate that four of the five mergers we study resulted in some increases in some consumer prices. The estimated price increases might be considered relatively modest, as they are typically between 3% and 7%. However, given the large amount of commerce in these industries, the implied transfer from consumers to manufacturers is substantial. While the magnitude of some price effects varies with our empirical specification, the conclusion that consumer prices did not decrease, and most likely increased, was not altered when we changed the way we measured consumer prices or the control group, or when we changed the choice of the window of time surrounding the merger during which price changes might be implemented.

Too acquiescent regulators

Some advocates of less intervention may be surprised to learn that our best estimate of the price effects of the marginal merger are positive, not negative as would be the case if the marginal merger were producing large benefits to consumers through the efficiency of the enlarged firm.

Our results have several limitations, and thus it would be premature to conclude either that the mergers we studied were, on net, harmful to consumers or that our evaluation is comprehensive. First, we are not able to analyze either the possible longer-term effects of mergers on prices that may result from the increased economic efficiency of merged firms, nor are we able to study the role of mergers in the development of new products. Second, we have restricted our analysis to consumer products mergers, which, although they attract much attention, constituted a small fraction of mergers in the period we study. Finally, our study has only examined the implications of Type II errors, the failure to block a merger when it might be useful to do so. Our study does not analyze Type I errors, rejecting mergers that would not have resulted in higher prices. It is possible that by allowing some anticompetitive mergers to take place, the government may also allow many efficient mergers to take place that would have been challenged by a stricter antitrust policy. A complete evaluation of optimal enforcement behaviour must be based on a consideration of all these issues.

References

Ashenfelter, Orley and Daniel Hosken (2008). “The Effect of Mergers on Consumers Prices: Evidence from Five Selected Case Studies,” NBER Working Paper 13859, February.

Baker, Jonathan. “The Case for Antitrust Enforcement.” Journal of Economic Perspectives Vol. 17 (2003) pp. 27-50.

Crandall, Robert W. and Clifford Winston. “Does Antitrust Policy Improve Consumer Welfare? Assessing the Evidence.” Journal of Economic Perspectives Vol. 17 (2003) pp. 3-26.

5,144 Reads