FTC Staff Testimony Before the Antitrust Modernization Commission Concerning the Treatment of Efficiencies in Merger Enforcement

PREPARED REMARKS OF
DR. MICHAEL A. SALINGER,
DIRECTOR, BUREAU OF ECONOMICS,
FEDERAL TRADE COMMISSION,
BEFORE THE
ANTITRUST MODERNIZATION COMMISSION:
TREATMENT OF EFFICIENCIES IN MERGER ENFORCEMENT
NOVEMBER 17, 2005
Thank you for inviting me to be here today. I am pleased to share with you my
views on this most important topic – the treatment of efficiencies in merger review. I
must note, of course, that they are my views – not necessarily the views of the Federal
Trade Commission or any of the individual commissioners.
As you are no doubt aware, I am relatively new to the Commission. I can,
however, attest to the fact that there is a clear message from the Chairman that
efficiencies matter. I can also attest that my staff analyzes efficiencies.
One case in which efficiencies were the explicit reason for not challenging a
merger was the Genzyme-Novazyme merger. In that case, it declined to take action even
though it was a merger to monopoly.1 The two firms were the only companies involved
in the development of a treatment for Pompe disease, a rare condition that affects about
10,000 people annually. As in any patent race, both firms had an incentive to be first.
1 Statement of Chairman Timothy J. Muris, Genzyme Corporation/Novazyme Pharmaceuticals, Inc. (Jan.
13, 2004), available at www.ftc.gov/os/2004/01/murisgenzymestmt.pdf; Dissenting Statement of
Commissioner Mozelle W. Thompson, Genzyme Corporation/Novazyme Pharmaceuticals, Inc. (Jan. 13,
2004), available at www.ftc.gov/os/2004/01/thompsongenzymestmt.pdf; Statement of Commissioner
Pamela Jones Harbour, Genzyme Corporation/Novazyme Pharmaceuticals, Inc. (Jan. 13, 2004), available at
www.ftc.gov/os/2004/01/harbourgenzymestmt.pdf.
1

All else equal, therefore, the existence of a competing developer would be a spur to
completing the development process quickly. Yet, all else was not equal. There was no
guarantee that either project would be successful, and both of the merging parties had
knowledge that could help the other’s development. The Commission allowed the
merger to go unchallenged on the grounds that the sharing of knowledge would speed up
the development more than the reduction in the competitive threat would slow it down.
To be sure, it is rare for the Commission to make efficiencies the explicit reason
for its decision. While this might create the impression that we do not give great weight
to efficiencies, I think such a conclusion would be unwarranted. Efficiencies do play a
key role in our analysis, although the way they are considered is perhaps less formal than
is suggested by the guidelines. In the evaluation of horizontal mergers, a key question
lurking in the back of everyone’s mind is, “Why do the parties want to do this deal?” with
efficiencies and the creation, extension, or preservation of market power being the
principal competing hypotheses. As the merger guidelines have developed through their
various iterations, efficiencies have moved, in part, from a possible “defense” to part of
an integrated analysis of competitive effects.2 In principle, our integrated analysis of
competitive effects is done scientifically – i.e., as part of a model. That approach
requires an estimate of the size of efficiencies, which in turn requires reliable data. In
practice, due to data and other limitations, we often do not have a formal model of the
effect of the merger on the market outcome. Rather, the weighing of competitive and
anticompetitive effects is less rigorous, requiring judgment. Efficiencies affect the
judgments we make even if they are not cognizable.
2 United States Dep’t of Justice and Federal Trade Commission, Horizontal Merger Guidelines (Apr. 2,
1992), reprinted in 4 Trade Reg. Rep (CCH) ¶ 13,104 at § 4.
2

If I am right that efficiencies matter quite a bit but in a relatively informal way, it
is interesting to consider why they do not play a more formal role. Let me suggest a few
reasons why that might be the case. The first is the combined effect of our “substantial
lessening of competition” standard – particularly as it has evolved over time – and the
continued use of consumer welfare as the basis for enforcement. In a two-to-one merger,
which would presumably result in monopoly pricing or a three-to-two merger in which
collusion to something like the monopoly price is plausible, it would take very large
marginal cost savings for the net effect of the merger to be a price reduction. Many
economists believe that a total welfare standard would be more appropriate. I am not
convinced that they are right, but I am confident that the measurement of efficiencies
would become more central if we did move to a total welfare standard. 	If you will
indulge me for a moment while I lapse into economics jargon, marginal cost savings are
what economists call “first order” effects and the deadweight loss from market power is a
“second order” effect. The graphical representation of this point is that deadweight loss
is a triangle in the graph showing the Williamsonian trade-off while marginal cost
savings are a rectangle. To see that this point is not mere abstract theorizing, look at the
Superior Propane case in Canada, where gains in producer surplus are included in the
criterion under which mergers are evaluated.3 In that case, the Canadian Competition
Commission estimated that the merger would result in a price increase of about 8% on
average.4 Still, the estimates of the efficiencies dwarfed the estimates of the deadweight
3 Comm'r of Competition v. Superior Propane Inc., [2000] C.C.T.D. No. 15, 7 C.P.R. (4th) 385 (Can.
Comp. Trib.) (original decision); [2001] 3 F.C. 185 (Fed. Ct. App.) (first Court of Appeal decision),
available at http://reports.fja.gc.ca/fc/2001/pub/v3/2001fc28500.html; [2002] C.C.T.D. No. 10, 18 C.P.R.
(4th) 417 (Can. Comp. Trib.) (re-determination); [2003] 3 F.C. 529 (Fed. Ct. App.) (second Court of
Appeal decision), available at http://reports.fja.gc.ca/fc/2003/pub/v3/2003fc31974.html.
4 Ross, Thomas W. and Ralph A. Winter, “The Efficiency Defense in Merger Law” 72 A NTITRUST LAW
J OURNAL (2005).
3

loss. Obviously, the precise comparison depends on the details of the merger, but it
should be no surprise when some estimates of efficiencies are much larger than the
estimate of deadweight loss.
Second, and perhaps related, parties often do a poor job of demonstrating
efficiencies. No matter how much we believe that efficiencies are a plausible outcome of
mergers, we cannot conclude that a merger will generate efficiencies simply because the
parties say it is so. Mere assertion is not proof or even, by itself, supporting evidence.
In our experience, one of the most commonly claimed efficiencies is overhead
expense. Often, little evidence is put forward to back up the assertion. The parties seem
to believe that it is self-evident that overhead expense is “fixed” and therefore can be
saved in a merger. This is not self-evident to me at all. In fact, there is quite a bit of
evidence to suggest that it is simply wrong. If so-called overhead expenses were in fact
fixed costs, then we would find that fixed costs are a smaller fraction of total costs for
large firms than for smaller ones. I am aware of no systematic evidence that such is the
case. It is certainly true that an organization has only one CEO, one CFO, one director of
marketing, and so on. The CEOs, CFOs, and marketing directors of large organizations
tend, however, to have higher compensation and larger staffs than their counterparts in
smaller organizations. If there are any savings from this sort of consolidation of central
office functions, they tend to be modest.
There are other types of efficiency claims that I find much more persuasive than
so-called fixed cost savings. One possible efficiency from mergers is knowledge transfer.
The acquirer knows something about operating the business that the acquiree does not (or
vice versa). One challenge that I imagine merging parties have is how to demonstrate
4

these sorts of efficiencies to the agencies. I see two principal barriers. First, the acquirer
might not know specifically how it is going to change the operations of the target. It
might merely know that it is more efficient. Second, even if the acquirer does know
specifically how to improve the operations of the target, it might be reluctant to share that
information. True, there are procedures for keeping information provided to the agencies
private. Still, I suspect that some companies might think it naïve to assume that the
confidentiality will be maintained.
With respect to the first of these barriers, evidence of productivity improvements
from previous acquisitions can be particularly useful in demonstrating that efficiencies
will be a plausible result of the merger. Although somewhat more speculative, evidence
of substantial differences in productivity between the firms can be evidence of the
possibility for efficiencies.
For it to make sense for parties to submit evidence of this sort, the agencies may
need to consider the burden of proof for a credible efficiency claim. As I have already
said, we cannot accept efficiency claims without evaluating them critically; but it is too
much to ask that parties demonstrate them beyond all doubt. Just as the agencies make
enforcement decisions based on assessments of the probability, not the certainty, of
anticompetitive effects, we must also do so with respect to the probability of efficiencies.
Again, I would like to thank the Commission for the opportunity to appear here
today, and I look forward to answering any questions.
5