Merger Analysis
Because the analysis of planned mergers inherently involves projections into the future, conditional prediction markets that forecast consumer surplus could help. Currently in the The government’s role in merger analysis is predictive; it anticipates the effects of a merger on competition. Because open-ended government decision making might promote considerable uncertainty among businesses considering whether to merge, the Antitrust Division and the FTC have issued joint Horizontal Merger Guidelines.39 The ultimate goal of the guidelines is to determine whether merged entities would have “market power,” that is, “the ability profitably to maintain prices above competitive levels for a significant period of time.”40 The guidelines provide an overview of how the government will define the relevant product and geographical markets that the merger might affect and how the government will calculate the market share of the firms. It also indicates how the government will consider the possibility that competition might be lessened by coordination and the possibility that entry in the future will limit the possibility that a business will operate monopolistically. In each of these areas, the guidelines provide indications of the approach that the government will take in assessing the issue, not mathematical formulas that potential merger partners can apply to predict the government’s approach with certainty. The guidelines thus fall somewhere in the middle of the rules-standards continuum. To the extent that the guidelines indeed constrain the government, they risk leading it to make decisions that it would not make if all factors were considered. To the extent that the guidelines allow open-ended decision making, they risk uncertainty and caprice. An ideal system of merger regulation would provide both more predictability and greater assurance that all relevant factors will be considered. The predictive decision making approach might use conditional prediction markets to forecast the effect of a merger on consumer surplus in the relevant market. I admit that this specification itself leaves considerable ambiguity by, for example, requiring some assessment of what the relevant product and geographical markets in fact are. But it at least saves the government from the challenge of modeling the effect of the merger on consumer prices. All the government would need to do is to identify a basket of products and promise to conduct consumer surveys or use other economic methodologies to calculate consumer surplus for the basket, whether the merger takes place or not. For example, in a planned merger of office supply stores, the basket might include pens, paper, and so on. The government also would need to provide an appropriate subsidy for the market. Prediction market participants would then have incentives to develop their own models of the merged company’s pricing. The approach might function effectively even without an ex ante identification of the relevant basket of goods. The government could simply provide that in measuring consumer surplus ex post, the agency will attempt to consider a representative basket of goods. Whether the merger takes place or not, years later governmental decision makers would conduct the analysis, focusing perhaps on a few representative products. There is always the danger that the government ex post will pick an unrepresentative basket of goods. The ex post measurement would matter only in disciplining the prediction market payouts, however, not in actually determining whether the merger takes place. So there would be little incentive for ideologically motivated governmental officials to manipulate the decision to advance a particular agenda. To the extent that manipulation or simply poor measurement might take place, it will be difficult ex ante to predict whether errors are more likely to be made in one direction or in another, especially if the effects of the merger are not to be evaluated for a long period of time. It thus might be the case that any ex post errors or manipulations will generally cancel each other out. This reveals a general point about predictive decision making: even if what is being predicted is only a noisy proxy, the ex ante prediction of it might be relatively stable and predictable. Of course, it is hard to be sure in the absence of experimentation, but it would not necessarily require adoption of this approach to merger analysis. Initial experiments could be purely informational. If this approach were adopted, prospective merger partners such as utility operators would have incentives to tie their hands in ways that would make the merger appear more attractive to market participants. They might commit to selling off particular divisions, the current practice. An advantage of the market approach, however, is that merger partners might have an incentive to develop other ways of committing not to charge monopoly prices. For example, they might develop a formula, perhaps a very sophisticated one, determining the maximum amount they would charge in the future for particular products. They could specify whether any consumer or only particular groups would be authorized to sue if the merged company arguably ignored the formula. Merging companies might agree to give representatives of consumer groups some role in managing the firm or in making pricing decisions. Prediction market participants once again would have incentives to assess the effectiveness of these various commitments.
Leave a Reply