Professor of Economics
DOJ/FTC Workshop on Merger
February 17, 2004
I am grateful for the opportunity to participate in this Merger Enforcement Workshop. Referring to our Monopsony panel, the press release announcing the Workshop states that the Agencies seek comment on "how, if at all, they should assess the creation for buying market power differently than selling power." In line with the Workshop's focus, I will frame my remarks largely in the context of horizontal mergers, though many of the points apply also to assessing the formation of market power through other means, such as exclusionary conduct.
I interpret "assess" to encompass two things: (a) the method of analysis, e.g., the types of information employed to predict the likely effects (such as the minimal level and changes in concentration that trigger concerns); and (b) the criteria used to decide whether to bring a challenge - e.g., whether a price decrease (in the case of buyer power) is enough.
The thrust of my remarks is that I know of no economic reason to justify different antitrust treatment of the formation of market power on the buying side than on the selling side. The reasons offered to justify differential treatment turn out, on closer look, to apply equally to seller power or to derive from other characteristics of the environment rather than the buyer / seller distinction.
I will develop this thesis by examining, first, arguments for treating buyer power less stringently than seller power, then arguments for treating it more stringently.
I. Less Stringent Treatment of Buyer than Seller Power?
During my term at the Antitrust Division and since then, I have encountered various questions about whether a merger should be challenged on grounds that it increases buying power. The following is a sampling, along with my reactions.1
"To Justify a Challenge, Must there be Harm to Consumers?"
The claim that "antitrust protects consumers not competitors" should not be interpreted as saying that a merger warrants antitrust concern only if it imposes harm on final consumers. Rather, its thrust is that harm to competitors of the merging firm is obviously not enough to warrant a challenge, since such harm can arise also if the merger increases efficiency and benefits consumers. However, the term "consumers" is, in my view, a metaphor for "trading partners" - be they buyers of the merging firms' products or sellers of inputs to them. A merger that increases market power and enables the merged firm to impose worse terms on its trading partners is equally objectionable if the trading partners in question are suppliers to, or buyers from, that firm.
To make this point sharply, consider a hypothetical example where final consumers buy a product directly from resource owners, such as corn from farmers. Suppose farmers are competitive, and their industry supply curve is upward sloping, meaning that a lower price can be obtained if buyers collectively agree on a lower quantity. Finally, suppose that a sufficient fraction of consumers band together in a buyers' group and reduce the quantity they purchase precisely in order to drive down price. The reduction in output produces a decrease in overall economic welfare ('total surplus'), meaning that the dollar value of the gains to consumers is less than the loss to producers. On economic grounds, there is no reason to be more tolerant of such a merger than if the shoe were on the other foot - where sellers merged to raise price to consumers.
Having argued that antitrust concern is warranted even if increased buyer power benefits consumers, it is worth stressing that the typical case of increased buying power involves consolidations among intermediaries (rather than among final consumers as above). In such a case, it is difficult to see why consumers would benefit, and easy to see why they would be harmed by a reduction in input purchases stemming from increased monopsony power of the merging firms.
At best, consumers may remain unaffected. This would occur if the merging firms obtain increased market power in the input market but remain price-takers in the output market, e.g., if on the selling side the geographic market is much broader and concentration is trivial. In such a case, the merger can result in a decrease in the input amount purchased by those firms and in its price, but any reduced output by those firms is made up from other perfectly-substitutable sources leaving output price and consumer welfare unchanged.2
Outside of the above case, where the merging firms obtain increased power in the input market but remain perfectly competitive on the output side, one would expect consumers to be harmed if the merger reduces input price and quantity. When the merging firms are imperfectly competitive (or local monopolists) on the output side, a reduction in their output, expected to flow from their reduced input purchases, is likely to raise price to consumers.3
"Is Countervailing Power an Acceptable Defense?"
Suppose that a merger is likely to reduce price to suppliers. Is it an acceptable defense to argue that the merger will countervail preexisting selling power of suppliers and, thus, is likely to reduce price towards the competitive level rather than below it?
There are arguments to be made on both sides. Under the assumed facts, the countervailing merger may improve welfare, at least in the short run. Countervailing power, however, is unlikely to duplicate the competitive outcome and raises its own risks. Instead of trying to combat one distortion by allowing an opposite one, an alternative approach is to resist any merger that increases market power, and rely on market forces to erode market power on the other side or, in those cases where competition is deemed incapable of doing so (perhaps because of exclusionary practices or formidable structural barriers to entry), consider tackling such market power through government intervention, be it antitrust or regulation as the case may warrant.
Whatever one's view of the relative merits of these arguments, countervailing power can arise also when a merger increases seller market power. Thus, if one accepts countervailing power as a potential defense, it applies also where the merging firms sell to buyers that have market power.4
"To Justify a Challenge, is a Monopsony-Induced Decrease in Price Sufficient, or Must there Also Be a Significant Loss of Output?"
Moving away from countervailing power, suppose that the merging firms face sellers that are competitive and that the merger is expected to depress price. However, suppose that in the short run there is unlikely to be a significant reduction in the quantity of the input purchased. Such an outcome could arise for at least two reason, (a) The elasticity of supply facing the merging firms (over the range of the predicted price cut) may be quite low. (b) Another possibility is that contracting between the merging firms and suppliers involves more than just per-unit prices, e.g., it could encompass two-part tariffs, quantity discounts, or other forms of non-linear pricing. Increased buyer power then may depress the total revenue to the sellers, but leave the quantity transacted unchanged (e.g., under a two-part tariff, if the merging firms continue paying the same price per unit but depress the fixed fee).
Should a buyer-power merger be opposed in such circumstances? This question is sometimes framed as "Are wealth transfers are a sufficient reason to oppose a merger?"
One response is that a decrease in the economic return to producers is likely to reduce not only their wealth but also output ultimately, even if it does not do so over the foreseeable horizon. Another potential response is that, even if one could be confident that there would be no losses in economic efficiency even in the long run, antitrust policy should be concerned also with the distributional impacts created by market power.
These questions, like those concerning countervailing power, are legitimate matters for debate. The key point again, however, is that they apply no less when dealing with a merger that increases selling power. Whether to oppose a merger that benefits the firms at the expense of its trading partners without reducing output significantly is an equally valid question when the trading partners are buyers as when they are sellers.
II. More Stringent Treatment of Buyer than Seller Power?
Having tried to dispel the notion that antitrust treatment should be less stringent in the case of buyer power, I now turn to whether it should be more stringent.
The Designation 'Buyer' or 'Seller' Is Often Arbitrary
Economically, it is not always clear which party is the 'buyer' and which is the 'seller'. Barter transactions provide an obvious illustration.5 A second example involves financial intermediaries. In dealing with depositors, an intermediary buys current dollars and sells future dollars. When dealing with borrowers, it sells current dollars and buys future ones. It is arbitrary whether the intermediary should be designated the 'buyer' in its role as the borrower or as the lender.6
Even where the designation of the buyer is seemingly clear, as the party that pays 'money' (generalized purchasing power) in exchange for a specific good or service, a simple reformulation of the transaction can switch the identities. For example, suppose that a manufacturer initially sells its good to a distributor. Now it moves to retain ownership of the goods until they are sold to final consumers, and pays the distributor a percentage of the final price - thereby purchasing distribution services. A merger of distributors that increased their market power versus manufacturers creates buyer power in the first case and seller power in the second, with no economically relevant difference between the two cases.
The point of these examples is to convey that the nature of the economic analysis in antitrust cases or the criteria for making decisions cannot turn on the labels 'buyer' or 'seller'. Let me illustrate this position further with reference to the following arguments that have been made for why buyer power may warrant stricter treatment.7
"The Anti-Competitive Harm Is Sometimes Felt Only at Several Levels Away."
A merger of powerful retailers in a concentrated market that depresses prices to suppliers may not harm primarily the immediate suppliers but suppliers several layers away in the economic chain, as the lower price is passed-through. One implication is that antitrust enforcers should not take a lack of complaints by trading partners as evidence that the merger will not produce an anti-competitive increase in buyer power, because these effects will be felt only by indirect sellers. This point is certainly valid. However, it applies with equal force when a merger increases the ability of sellers to raise price, and that price increase to direct purchasers is likely, for various reasons, to be passed through further down the chain, possibly to final consumers.
"Certain Buyer Power 'Abuse' is Not Reached by Antitrust."
Certain practices that may be perceived as objectionable, such as 'unfairly low' prices paid to farmers or lower prices paid to independent farmers on a spot basis than to those on longer-term contracts, are likely to lie outside the reach of antitrust. However, the same is true of such perceived abuses when undertaken by sellers. Antitrust does not make it unlawful, for example, to charge a monopoly price provided the monopoly position was attained and maintained through legitimate means. Therefore, to the extent one believes that perceived problems in a particular industry warrant intervention beyond the scope of antitrust, such a position can apply equally to buyer power as to seller power.
"Lower Market Shares Can Suffice for Harm in Buyer Power Cases."
It has been noted that low concentration nationally is consistent with the presence of buyer market power, because the relevant geographic markets for purchasing are often localized, hence buyer concentration in those markets is considerably higher than it is nationally. While economically correct, this observation merely states that one must be careful in properly identifying the relevant geographic market. If adopting an overly broad definition reduces the measured level of concentration, this can understate the actual degree of market power that may exist in a narrower market. But this caveat applies equally when gauging seller market power. Moreover, it is not self evident that geographic markets are always narrower on the buying than on the selling side, so that measuring concentration at a national level would systematically understate the relevant market concentration by more in buyer-side cases.8
Even assuming that the identified market is not overly broad, it has been argued that competitive harm from buyer market power requires lower market shares than for seller power, e.g., because coordination among buyers is easier. However, the evidence offered in support of this position is anecdotal, and may reflect not the buyer-seller distinction but other factors specific to the industry or the case.9 I am not aware of systematic economic evidence from which one could reliably generalize that buyer power concerns arise at lower concentration thresholds than for seller power.
III. Concluding Remarks
The economic approach to analyzing seller or buyer market power is the same. In both cases, the fundamental inquiry revolves around the competitive alternatives available to trading partners as substitutes to dealing with the firm(s) in question. Answering this question often requires a complex and fact-intensive analysis of numerous subsidiary factors, such as the similarity of product offerings, transportation costs, capacity constraints, and the like. But the basic method of inquiry is the same for buyer power as for seller power.
Nevertheless, there is a perception that the Agencies bring relatively fewer buying power cases. What might explain this? In my experience at the DOT, it is not a lack of concern with buyer power nor a failure to understand its sources or effects. The Economic Analysis Group at the DOJ's Antitrust Division has some fifty Ph.D. economists who are fully versed in the economics of monopsony power.
Any difference in enforcement patterns is unlikely to arise from the buyer-seller distinction per se, but more likely traced to other factors that may be correlated with this distinction. One hypothesis is that concentration levels decrease as one moves 'down' the economic vertical chain of activity, closer to final consumers. Consider a stylized example in which firms at each stage are symmetric, the number of firms increases monotonically moving down the chain, and firms at any stage only buy from and sell to firms at the immediate adjacent stages. The 'typical' vertical stage, then, would be less concentrated than the previous 'upstream' stage from which it buys, and more concentrated than the next 'downstream' stage to which it sells. If so, this may explain why, on average, market power is more likely to arise on the selling side. For instance, in the interface between final consumers and firms with which they deal, the market power is unlikely to reside with consumers, so the typical concern would be 'seller power.'
Along those lines, an interesting factual question is whether the Agencies' enforcement patterns remain significantly skewed towards seller power even after excluding consumer good industries and considering only intermediate producers. A second question is whether concentration indeed is lower on average the closer the activity is to final consumers (recognizing that the measure of 'closeness' is not always unambiguous).
In conclusion, I recognize that when scholars or practitioners call for assessing buyer power differently from seller power, they may be using these labels as shorthand for particular industries or for certain industry characteristics. However, it is important to understand what those fundamental economic characteristics might be and incorporate them explicitly into the analysis. Doing so allows for sharper thinking. Sound antitrust policy demands this, and would be undermined by framing distinctions based on the 'buyer' or 'seller' labels.
1 Some of these points are discussed more fully in Marius Schwartz, "Buyer Power Concerns and the Aetna-Prudential Merger," Address presented at 5th Annual Health Care Antitrust Forum, Northwestern University School of Law, October 20,1999, posted on web site of Antitrust Division, Department of Justice: <http://www.usdoj.gov/atr/public/speeches/3924.htm>
2 For example, in the Cargill-Continental merger of two major grain traders, the Department of Justice alleged harm on the buying side, to farmers and other grain suppliers in certain localities, but not on the selling side, since grain prices to consumers (or to national processors that sell to consumers) are determined in world markets. For a summary of this case, see Susan Davies and Marius Schwartz, "Monopsony Concerns in Merger Review," American Bar Association Antitrust Section, Clayton Act Committee Newsletter, vol. II, no. 1, Winter 2002.
3 Consider the following stylized example. Suppose that two firms are local monopolists in the sale of some outputs, but compete for purchasing an input. A merger between them does not increase their market power on the output side, because they were monopolists initially. However, the increased monopsony power can lead them to reduce input purchases, which results in lower quantity and higher prices in both output markets.
4 If such a defense is acceptable, there remains the question of who should have the burden of proof, antitrust enforcers or the merging parties. Again, the appropriate standard should be symmetric in buy-side and sell-side mergers. Thus, if one puts the onus on antitrust enforcers to show that a merger that increases buying power not only would depress price to suppliers but would move it further from the competitive level, a similar burden would seem appropriate for challenging a selling side merger. Conversely, if showing a price increase is deemed sufficient to challenge a sellers' merger, showing a price decrease should suffice for challenging a buyers' merger.
5 For example, consider a content provider supplying programming and a television broadcaster supplying distribution, with the advertising revenue split between the parties. There is no obvious sense in which one party is the buyer rather than the seller.
6 Moreover, whatever convention is adopted, the intermediary will be both a buyer from and a seller to consumers. So a merger that increased intermediaries' market power towards both depositors and borrowers will harm some consumers by raising price and others by lowering price, further illustrating the pitfalls of hinging the economic analysis on the labels 'buyer power' vs. 'seller power'.
7 Some of these arguments can be found in Peter Carstensen, "Monopsony in Markets for Agricultural Products: A Serious Problem in Need of a Remedy," Statement prepared for the hearing on "Monopsony Issues in Agriculture: Buying Power of Processors in Our Nation's Agricultural Markets," Senate Judiciary Committee, October 30,2003, <http://www.senate.gov/~judiciary/testimony.cfm?id=975&wit_id=2782> I take no position on Professor Carstensen's specific points in the context of agriculture, but am addressing whether they would justify a stricter antitrust treatment of buyer power than seller power generally.
8 One can think of examples where the relevant geographic market is narrower on the buying side and others where it is narrower on the selling side. For instance, slaughter plants purchase hogs or cattle and sell meat. Transportation costs are likely to be higher for livestock than for meat, suggesting a narrower geographic market on the purchasing side. However, the reverse can occur in the case of retailers that sell in localized markets but obtain their products (e.g., clothing or electronic goods) on world markets.
9 The Toys "R" Us case has been cited as an example where low market share was sufficient to generate significant buyer power. However, that case was colored by special circumstances, such as the alleged role of Toys "R Us in facilitating a conspiracy among toy manufacturers. Toys "R" Us v. FTC, 221 F.3d 928 (7th Circ. 2000).
When such aggravating factors are present, challenges have been brought at relatively low market shares also against sellers. For example, the Department of Justice recently sued to block UPM's acquisition of MACtac, a subsidiary of Bemis that competed with UPM"s Raflatac division in the sale of label stock. The Complaint states that the merging parties had market shares of about 12 percent each. In challenging the merger, a key factor stressed in the Complaint is a history of attempted coordination between UPM and the market leader (whose share was estimated at 50 percent). United States v. UPM-Kymmene, OYJ, Raflatac, Inc., Bemis Company, Inc. & Morgan Adhesives Company, Verified Complaint, April 15,2003: <http://www.usdoj.gov/atr/cases/f200900/200942.htm> For a further discussion of this case see David S. Sibley and Ken Heyer, "Selected Economic Analysis at the Antitrust Division: The Year in Review," Review of Industrial Organization, vol. 23, no. 2, September 2003,95-119. In July 2003, Judge Zagel of the US District Court for the Northern District of Illinois granted the DOJ's motion for a preliminary injunction, and shortly thereafter the Parties abandoned the transaction.