A R T I C L E S • A N D • F E A T U R E S
S P R I N G 1 9 9 6 · 2 3
T
he antitrust treatment of horizontal
mergers by the Justice
Department and the Federal
Trade Commission is one of
the most well developed and closely scrutinized
areas of antitrust law. The
enforcement agencies have extraordinary
experience reviewing mergers and the
merger bar is no less sophisticated. From
my perspective as an antitrust economist,
this sophistication permits merger
enforcement to be at the cutting edge
when it comes to incorporating economic
learning into competition policy.
The 1992 DOJ and FTC Merger
Guidelines have placed important attention
on the “unilateral effects” of a merger,
i.e., the tendency of a horizontal merger
to lead to higher prices simply by
virtue of the fact that the merger will
eliminate the direct competition between
the two merging firms, even if all other
firms in the market continue to compete
independently. Unilateral effects are contrasted
to “coordinated effects,” i.e., the
danger that the merger will lead to collusion
between the merged entity and its
remaining rivals.
This article discusses some of the
methods used by the Antitrust Division to
analyze unilateral effects in mergers
involving differentiated products. While
the methods outlined here do not replace
the standard steps of defining markets
and measuring market shares and concentration,
they can significantly supplement
those steps.
To place the topic in context, it is
instructive to compare mergers with
homogeneous products to those involving
differentiated products. For homogeneous
products, the traditional structural
approach of defining markets and measuring
market shares and market concentration
has deep roots, along with a rich
empirical tradition linking market structure
to performance. In these markets, it
is both natural and appropriate to count
up each firm’s unit or dollar sales, or
capacity, to measure market shares, and
to make inferences about a merger’s
effects based on market structure, including
the HHI of market concentration. The
danger of collusion is surely related to
market concentration (although quantifying
this relationship is difficult), and
economists’ primary model of non-cooperative
oligopolistic competition among
manufacturers of homogeneous goods
relates market structure to performance.1
Although economists continue to debate
the empirical relationship between market
structure and performance, there
exists a solid foundation for using market
structure prominently in evaluating horizontal
mergers involving homogeneous
goods.2
This traditional structural approach
towards merger policy, which dates back
to the 1960s but has been refined as just
described, is less attractive for differentiated
products. When products are highly
differentiated, concerns about coordinated
effects may be secondary to concerns
about unilateral effects. And, to assess
unilateral effects most accurately, it is
highly desirable to go beyond industry
concentration measures to look directly
at the extent of competition between the
merging brands. This is especially true if
competition is “localized,” i.e., if some
brands are especially close substitutes for
other brands due to their product characteristics
or image. To put this differently,
at the Antitrust Division we must concern
ourselves with the prospect that the
prices for one or more brands sold by the
merging parties will rise after the merger,
even if prices do not uniformly rise
throughout the relevant market. Such
concerns are not present, or are far less
pronounced, in markets for homogeneous
products.
The danger that a merger of Brands A
and B will cause anticompetitive price
increases for one or both of these brands
is greatest if the merging brands are
“close,” in the sense that many customers
using one brand consider the other brand
their second choice. Ultimately, unilateral
anticompetitive effects are based on the
following logic: As the price of Brand A
rises, some customers will shift from
Brand A to Brand B. Prior to the merger,
these customers would be lost to the firm
owning Brand A. After the merger, this
same firm owns Brand B and thus does
not lose these customers. As a result, the
price increase is more profitable to the
merged entity.
To explicate this logic, consider a
merger between Brands A and B. The
likely post-merger price increase for
Brand A is driven by two variables, each
of which we have a good chance of
observing with a proper investigation.
The first I call the Diversion Ratio from
Brand A to B. This Diversion Ratio is the
answer to the following question: If the
price of Brand A were to rise, what fraction
of the customers leaving Brand A
would switch to Brand B? The second
variable is the Gross Margin — the percentage
gap between price and incremental
cost — for Brand B.3
Roughly
speaking, a valuable index of the potential
anticompetitive unilateral effects is
obtained by multiplying the Diversion
Ratio by the Gross Margin. Any danger
of a unilateral price increase may be alleviated
by product repositioning, entry, or
efficiencies. Nonetheless, the Diversion
Ratio and the Gross Margin are the key
variables in the demand-side portion of
the analysis.
Mergers with Differentiated Products
by Carl Shapiro
Carl Shapiro is the Deputy Assistant
Attorney General in charge of economics
at the Antitrust Division, U.S. Department
of Justice. He is currently on leave from his
position as the Transamerica Professor of
Business Strategy at the Haas School of
Business and Professor of Economics in
the Economics Department at the
University of California at Berkeley. Dr.
Shapiro thanks Joseph Kattan, George
Rozanski, Steve Salop, Greg Werden, and
Robert Willig for valuable comments on
this paper.
A R T I C L E S • A N D • F E A T U R E S
2 4 · A N T I T R U S T
Monopolistic Competition
To a greater or lesser degree, virtually all
markets involve some element of product
differentiation.4
Even in a classic homogeneous-goods
market — such as the
market for an agricultural commodity or
for a specific chemical compound — producers
often attempt to differentiate
themselves based on product quality,
reliability, or customer service.
My emphasis here is on markets
where the brands are distinct in important
and long-lasting ways. It is helpful to
keep two categories in mind: (1) branded
consumer products, where each brand of
pens, bread, facial tissue, computer software,
or cereal, is distinct; and (2) physical
facilities that distribute or deliver
goods or services, such as supermarkets,
department stores, branch banks, or hospitals,
where the differentiation is based
on location. These two broad categories
correspond to the familiar tasks of product
and geographic market definition.
Unilateral Competitive Effects
The predominant approach taken by
economists studying markets with differentiated
products is to model the firms
as independently setting the prices of
each of their brands. As usual in economics,
we assume that each firm seeks
to maximize its own profits. This method
of analysis fits perfectly with the “unilateral
effects” portion of the Merger
Guidelines (§ 2.2).
It is fair to say that economic analysis
of differentiated-products mergers at the
Division typically focuses on unilateral
effects, unless there are structural factors
facilitating collusion following the merger
or there is a history of collusion in the
industry. This emphasis represents a significant
shift in a fairly short period of
time.
Economic analysis regarding unilateral
effects is more amenable to quantification
than is economic analysis of the
dangers of collusion. Ultimately, we are
trying to measure the added incentive to
raise price caused by the merger.
Employing a combination of game-theoretic
and econometric methods, we now
have the capability to estimate consumer
demand using industry data and, based
on these demand estimates, to derive specific
predictions regarding post-merger
prices.5
This contrasts somewhat with
the analysis of the dangers of collusion.
While we are fairly confident in listing
and analyzing factors that facilitate or
hinder collusion, including market structure,
there is no single accepted method
of quantifying the increased likelihood
of collusion attendant to a merger.
Our ability to predict unilateral effects
based on demand patterns is only as good
as the available data. Very often we must
make do with qualitative evidence or
rather rough estimates. However, even
when the data are limited, the theory of
monopolistic competition can provide
some very helpful predictions based on
premerger Gross Margins and market
shares, which often can be measured
using data routinely collected during the
second request process. These predictions
will need to be checked against the views
of industry participants, company documents,
and other information sources, but
they do constitute a vital part of merger
analysis.
A final caveat: the analysis below
applies when the firms independently set
uniform prices for their branded products.
To the extent that firms engage in
price negotiations on a customer-by-customer
basis or engage in other forms of
price discrimination, the analysis must
be modified or replaced with another
approach.
Estimating Post-Merger Prices:
Four Steps
Economists in the agencies and those
working for private parties largely agree
on what to look for to estimate unilateral
competitive effects, although they may
differ in specific implementation steps.
We seek to estimate the post-merger
Bertrand equilibrium in prices, accounting
for the new market structure in which
some brands are jointly owned that had
previously been independent and for the
new cost structure of the merged entity.6
The analytical steps in this exercise
can be described succinctly in a manner
consistent with the methodology of the
Merger Guidelines: If a significant proportion
of consumers considers the merging
firms’ products to be their first and
second choices (at premerger prices),
then the merged entity will have an incentive
to impose a nontrivial price increase
following the merger. Unless product
repositioning or entry would render such
a price increase unprofitable, and unless
synergies imply that the price increase
will not in fact raise profits, the merger
will injure consumers and be anticom-
petitive.
There are many valid ways to carry
out this analysis. As much as anything,
the method chosen depends upon the
available data. Generally, however, I find
it useful to structure my thinking about
unilateral effects in differentiated product
markets in terms of the following four
steps.7
Consider a merger between brands
A and B, and focus attention on the danger
that the price of Brand A will be elevated
after the merger:8
(1) Consider a price increase for
Brand A of, say, 10 percent. Try to measure
what fraction of the sales lost by
Brand A due to this price increase would
be captured by Brand B. I call this fraction
the Diversion Ratio (from Brand A to
Brand B).
(2) Based on premerger Gross Margins
and the estimated Diversion Ratio,
calculate the post-merger price increase,
assuming no synergies or rival supply
responses.
(3) Try to account for any likely and
timely changes in prices or product offerings
by nonmerging parties, including
product repositioning and entry.
(4) If there are credible and documented
synergies that lower marginal
costs, reduce the predicted post-merger
prices accordingly. This step can lead to
a predicted price decrease.
The first two steps capture the
demand-side analysis; the third step is
the supply-side analysis; the final step
accounts for possible efficiencies. If these
steps indicate that the merged entity
would find it optimal to impose a significant
price increase following the merger,
then the merger is very likely to be anti-
competitive.
The Diversion Ratio
The concept underlying the Diversion
Ratio is easily comprehensible: “If you
raise your price, what fraction of your
lost customers will turn to your rival
(now merger partner)?”
The Diversion Ratio is a close cousin
of the cross-elasticity of demand between
the merging brands.9
If the unit sales of
the merging brands are equal prior to the
merger, the Diversion Ratio from Brand
A R T I C L E S • A N D • F E A T U R E S
S P R I N G 1 9 9 6 · 2 5
A to Brand B is equal to the cross elasticity
of demand from Brand A to Brand
B, divided by the own-price elasticity of
demand for Brand A.10
Suppose, for
example, that Brand A has an own-price
elasticity of demand of 2.0 (i.e., a 1 percent
increase in the price of Brand A
causes unit sales to decline by 2 percent),
and that the cross-price elasticity of
demand from Brand A to Brand B is 0.5.
If the two brands’ premerger unit sales
are equal, then the Diversion Ratio from
Brand A to Brand B is (0.5)/(2.0) or 25
percent. Stated differently, one-quarter of
the unit sales lost by Brand A if its price
rises are captured by Brand B.
In some cases, the Diversion Ratio
from Brand A to Brand B will be closely
linked to Brand B’s market share. In particular,
if all sales lost by Brand A are
captured by other brands in the market,
and if all brands are “equally close” to
each other, then the Diversion Ratio from
Brand A to Brand B may be stated as
sB
/(1-sA
), where sA
and sB
are the brands’
respective market shares.11
In the more
realistic situation where some customers
substituting away from Brand A switch to
products outside the market entirely, this
Diversion Ratio will be proportionately
lower. For example, if 20 percent of the
customers lost by Brand A leave the market
entirely, the Diversion Ratio from
Brand A to Brand B will instead be
(0.8)sB
/(1-sA
).
Since the Diversion Ratio plays a crucial
role in this analysis, in differentiatedproduct
mergers the Antitrust Division
will invariably want to know the best
estimate of the Diversion Ratio based on
the available evidence.
Estimating Unilateral
Competitive Effects in Practice
Steps 1 and 2: Demand-Side Analysis. In
practice, the demand-side analysis
depends heavily on the availability of
data.12
Two recent merger cases at the
Antitrust Division that led to consent
decrees illustrate what can be done with
detailed data.
The merger of Interstate Bakeries
Corporation and Continental Baking
Company involved the first and third
largest bakers of fresh bread in the
United States.13
The Division concluded
that the deal would substantially lessen
competition in the production and sale
of white pan bread in five regional
markets. The proposed Final Judgment
orders Interstate and Continental to divest
certain white bread brands in each
geographic market. Our competitive concerns,
as well as the ultimate relief, were
greatly informed by our economic analysis
of the merger’s anticompetitive unilateral
effects.
The Kimberly-Clark acquisition of
Scott Paper Company posed a threat to
competition in the markets for baby
wipes and facial tissues.14
Again, our
analysis was significantly informed by
econometric analysis of the demand for
the various brands in these markets, from
which we were able to estimate likely
price increases following the acquisition.
In both of these merger investigations,
we had access to excellent data on prices
and unit sales derived from checkout
scanners at retail locations. These data
provided information regarding competition
both from other brands and from
private label products. In the bakeries
merger, we focused on the brands of premium
white bread sold by Continental
and Interstate. In the baby wipes market,
we focused on competition between
Kimberly-Clark’s Huggies brand of baby
wipes and Scott’s two brands of baby
wipes, Baby Fresh and Wash-a-Bye
Baby; together the merging parties controlled
over 50 percent of the sales in this
$500 million market. In the facial tissue
market, we estimated the cross-elasticity
between Kimberly-Clark’s Kleenex
brand and Scott’s Scotties brand; together,
the two parties controlled nearly 60
percent of the sales in this $1.34 billion
market.
In these two mergers we were able,
with considerable work and making various
assumptions about the structure of
demand, to estimate a complete model of
demand for the various brands. These
methods subsume the Diversion Ratio
concept I stressed above. The calibrated
model of consumer demand derived from
the supermarket scanner data was then
used, in a high-tech version of Steps 1
and 2 above, to predict the likely postmerger
price increases for the various
merging brands.15
In this prediction exercise, we
assumed that all firms in the industry set
prices independently after the merger to
maximize profits. The predictions of the
model at this point do not account for
product repositioning, entry, or synergies.
In all cases, however, this demand-side
analysis did take account of competition
from private-label products as well as
branded goods. In the bread merger, the
computer model predicted price increases
in the 5 to 15 percent range for
Continental’s and Interstate’s premium
white pan breads in the Los Angeles and
Chicago areas. In baby wipes, we estimated
that there would have been substantial
price increases following the
merger. Price increases were also predicted
in facial tissue, especially for the
smaller Scotties brand.
While econometricians dream about
this type of “high-tech” analysis, in reality
the data are rarely available to do this
type of full-blown simulation analysis
with assurance. If econometric estimation
of elasticities based on transactions
data is not possible, relevant consumer
survey data may still be available to
directly estimate the Diversion Ratio.
Survey data are not as good as actual
transactions data, but can still be reliable
if valid sampling procedures were used
and if the results are not overly sensitive
to the framing of the questions. If reliable
survey data are also unavailable, it still
may be possible to use company documents
and other qualitative information
regarding consumers’ brand preferences
to estimate Diversion Ratios.
The firm’s market shares can be very
helpful in estimating Diversion Ratios if
none of the brands in the market are especially
“close” to or “distant” from each
other. As noted above, Diversion Ratios
will be proportional to market shares in
this case. For example, consider such a
market in which Brand A has a 25 percent
share and Brand B has a 15 percent
share. Suppose also that very few customers
of Brand A would reduce their
overall purchases in the market if Brand
A were to raise its price; instead these
customers would by and large pick
among the other brands. In this case, the
Diversion Ratio from Brand A to Brand
B is 20 percent.16
The Diversion Ratio
will be lower, to the extent that some of
Brand A’s customers reduce their total
purchases in the market when the price of
Brand A rises. If half of the customers
dropping Brand A leave the market altogether,
or if customers switching to other
A R T I C L E S • A N D • F E A T U R E S
2 6 · A N T I T R U S T
brands in the market reduce their purchases
by half when switching away from
their favorite Brand, the Diversion Ratio
from Brand A to Brand B will only be 10
percent.
Under this analysis, if the merging
brands are similar in characteristics, or if
the merging brands have large shares
within a broader product category, the
Diversion Ratio is likely to be high. Note
in particular that the Diversion Ratio is
likely to be high for a brand that is merging
with a dominant brand: the large market
share of the dominant brand makes it
likely that customers switching away
from the smaller brand will divert to the
dominant brand rather than elsewhere.
On the other hand, if the merging brands
are usually sold to different types of consumers,
or through different channels, or
if consumers’ preferences are such that
they can easily substitute to a broad range
of products (e.g., gifts instead of premium
pens, or breakfast foods instead of readyto-eat
cereals), the Diversion Ratio is
likely to be lower, other things equal.
Merely asserting that there are numerous
products to which consumers could
substitute, and thus the Diversion Ratio
must be low, ignores the importance of
diverse consumer preferences and does
not replace this step of the analysis. Nor
is a merger immunized merely because
the merging brands are not next-closest
substitutes, as some parties claim, any
more than a merger is immunized merely
because the merged entity still faces
some post-merger competition.
To complete the demand-side analysis,
the estimated Diversion Ratio can be
used, along with premerger Gross
Margins and perhaps other industry measures,
to give a rough prediction of the
post-merger price increases for the merging
brands. The principle here is that
high Gross Margins and high Diversion
Ratios suggest large post-merger price
increases. The tricky part is that the actual
calculation of the post-merger price
increase depends upon the specific shape
assumed for the demand curve.
A simple formula can be derived for
the post-merger price increase if one is
willing to assume that consumer demand
functions exhibit constant elasticity over
the relevant range of prices. Very often
when economists estimate demand using
data, they employ such constant-elasticity
demand functions. With constant-elasticity
demand, and assuming that the two
merging brands are symmetric prior to
the merger, the merged entity’s profitmaximizing
percentage price increase is
mD/(1–m–D).17
Here m is the premerger
Gross Margin and D is the Diversion
Ratio between the two merging brands.
For example, suppose that the premerger
price is $100, and the cost per
unit is $60, so the premerger markup,
m, is 40 percent, not uncommon at all
for differentiated products. If we assume
a Diversion Ratio of 0.2 (i.e., 20 percent
of the sales lost when the price of
Brand A goes up are captured by
Brand B), then the optimal post-merger
price increase in percentage terms is
(0.4)*(0.2)/(1–0.4–0.2) = 0.2: a 20 percent
price increase would maximize
profits.
This formula must be used with great
caution because it relies on several strong
assumptions, as I have noted. To the
extent that the elasticity of demand for a
brand rises as the price of that brand rises,
Diversion Ratio Example
T
he importance of the Diversion Ratio and the Gross Margin can be seen with a few
numerical examples. Consider a situation in which Brands A and B each have premerger
prices of $100 and premerger unit sales of 1000. Suppose that each brand has
a marginal cost of $60, so the premerger Gross Margin is ($100–$60)/$100 or 40%. Prior
to the merger, Brand A makes profits (often, contribution to fixed costs) of $40 per unit times
1000 units, or $40,000. Consistent with premerger optimal pricing by Brand A, suppose that
a 10% price increase by Brand A would lead to a 25% reduction in sales, to 750 units. At this
higher price, Brand A could earn profits of $50 per unit times 750 units, or $37,500. Since
this is less than the $40,000 figure, prior to the merger Firm A did not find it profitable to set
the higher price.
How would a merger between Brands A and B change this calculation? Suppose that the
Diversion Ratio from Brand A to Brand B is 30%. In other words, of the 250 units lost by Brand
A due to the price increase, 30%, or 75 units, are diverted to Brand B. The merged entity would
take into account the additional profits earned by Brand B from these customers when considering
raising prices from $100 to $110. Assuming the price of Brand B also rises to $110,
these 75 diverted sales generate profits of $50 per unit, or $3750 in total. Adding these to the
$37,500 from the premerger calculation, the post-merger profits per brand at the elevated
prices are $41,250. The merger has made it profitable to raise prices 10% above premerger
levels. In this example, a 10% price increase will be profitable after the merger if and only if
the Diversion Ratio is at least 20%.*
Compare this example to one in which the premerger Gross Margin is smaller. Specifically,
suppose now that the marginal cost is $80, so the premerger Gross Margin is only
($100–$80)/$100 or 20%. The premerger profits on Brand A are now only $20 per unit or
$20,000. Now, consistent with optimal premerger pricing, suppose that a 10% price increase
would result in a 50% loss of sales, to 500 units. Prior to the merger, a 10% price increase
would reduce profits to $30 per unit times 500 units or $15,000, some $5000 less than could
be earned at a price of $100.
What about after the merger? Now a 30% Diversion Ratio would mean 30% of the 500
lost units, or 150 units, would be captured by Brand B; at a $30 profit margin per unit, this
adds $4500 in profits. But this $4500 of extra profit is still not enough to make up for the loss
of $5000 in profits on Brand A. With the smaller Gross Margin in this example, a larger
Diversion Ratio would be necessary to make a 10% price increase profitable.
*In these examples, I ask only whether a 10% price increase would generate more profits than premerger prices.
Below, I discuss how to calculate the optimal post-merger price increase. Fortunately, with linear demand, the
maximal profitable price increase is exactly twice the optimal price increase. So, if a 10% price increase is profitable,
it will be optimal to raise prices at least 5%. Throughout this article, I use 10% price increases purely for
illustrative purposes.
A R T I C L E S • A N D • F E A T U R E S
S P R I N G 1 9 9 6 · 2 7
the constant-elasticity-of-demand calculations
will overestimate the post-merger
price increases. This overestimation can
be significant, especially if the formula as
stated generates a large percentage price
increase.18
Certainly it is desirable to find
support for the constant-elasticity
assumption in documents, in the testimony
of industry participants, or in the data
before using this formula or variants on it.
If demand instead takes a linear form,
the elasticity rises as the price rises, making
the optimal post-merger price
increase smaller. In this case, an alternative
formula can be derived: the optimal
post-merger percentage price increase
with linear demand (again with premerger
symmetry between the two brands) is
given by mD/2(1–D). This formula is
quite different from the earlier one. Using
the same numerical example as above,
with a premerger Gross Margin of 40 percent
and a Diversion Ratio of 20 percent,
the post-merger price increase would be
“only” 5 percent.19
I am keenly aware that the two formulas
presented above give quite different
predictions. This observation does not
take away from my main point — that
Diversion Ratios and Gross Margins are
key variables to explore in a merger
investigation involving differentiated
products — but should serve as a warning.
The fact is, the profit-maximizing
post-merger price increase is sensitive to
consumers’ ability to substitute away
from the merging brands as prices rise.20
It is simply not possible, and one should
not expect, to fully predict price changes
on the basis of two numbers, the
Diversion Ratio and the Gross Margin,
alone. One way or another, however, the
Diversion Ratio (or its close cousin) and
the Gross Margin will be an integral part
of the analysis.
Finally, even if data are limited and
precise predictions of post-merger price
increases are difficult to make with confidence,
these formulas still are a useful
starting point in gaining a sense of the
likely magnitude of any post-merger
price increase, if full-scale demand estimation
is not possible.21
And the numerator
in both of these expressions, mD, i.e.,
the product of the Gross Margin and the
Diversion Ratio, is a very useful quick
guide to the likely anticompetitive dangers,
even if we cannot predict the
post-merger price increase confidently
based on these two variables alone.
In some cases, we can measure Gross
Margins quite well, but may be uncertain
about the Diversion Ratios. In this situation
we can use either of the two formulas
above to ask how large the Diversion
Ratio must be in order for the optimal
post-merger price to be at least 10 percent
(say) above premerger levels. Using the
linear demand formula just given, the
post-merger price increase will be at
least 10 percent if and only if the
Diversion Ratio is at least 0.2/(m+0.2). If
the premerger Gross Margin is 40 percent
(m=0.4), prices will go up at least
10 percent if the Diversion Ratio is at
least one-third. Again, the larger the
Gross Margins, the lower the Diversion
Ratio necessary to raise anticompetitive
concerns.
Steps 1 and 2 invariably will lead to
the interim prediction that prices will rise
after the merger, if indeed Brands A and
B compete with each other. After all, a
merger between rival brands does eliminate
competition between those brands,
which in and of itself leads to higher
prices.22
But we do not condemn all horizontal
mergers, of course. To begin with,
we recognize that this incentive to raise
price is very slight for some mergers,
even horizontal ones. Steps 1 and 2
should detect this, in the form of a very
small predicted price increase. But there
are two more important reasons why horizontal
mergers often are not anticompetitive.
First, the true price increase may
be far smaller than predicted by Steps 1
and 2, or negligible, because rivals may
respond to defeat any price increase.
Second, the merger may reduce costs.
Steps 3 and 4 are where these crucial considerations
come into play.
Step 3: Product Repositioning and
Entry. If the merging brands are “close”
in attributes, the Diversion Ratio is likely
to be high, and Step 2 will suggest a
significant price increase. In precisely
this situation, however, it may well pay
for a rival firm to reposition its brand, or
introduce a new brand, closer to the
merging brands. And this threat could
well deter the price increase in the first
place. Alternatively, a de novo entrant
could locate its brand near Brands A and
B if prices of these brands were above
competitive levels. Very often in differentiated-product
markets, brands enter
and exit with some regularity, and existing
products may be repositioned either
through design changes or revised marketing
strategies. As a general rule, the
“farther” a brand must be moved to compete
more effectively with the merging
brands, the less likely it is that such a
move would in fact occur in response to
a post-merger price increase. As noted
generally in the Merger Guidelines, the
greater the sunk costs associated with
product entry or repositioning, and the
longer such supply responses take, the
less likely they are to deter or defeat an
anticompetitive price increase.
Notwithstanding the foregoing, rivals’
responses do not necessarily reduce the
profitability of a post-merger price
increase. Game-theoretic analyses of
pricing competition with differentiated
products indicate that rivals will typically
find it optimal to raise their prices in
response to higher prices set by the merging
firms. Accounting for these accommodating
responses tends to increase, not
decrease, the predicted post-merger price
increase.
Merging parties in consumer-goods
industries may be tempted to argue that
brand name is unimportant, but they
should be cautious in doing so. Such
claims are not credible if the parties
themselves have made substantial investments
in brand equity, or if the deal price
itself reflects substantial brand equity.
Attempts to downplay the importance of
brand names are particularly problematic
if new brands historically have found it
difficult to gain a secure foothold in the
market.
In the recent bread merger the
Division concluded that brand names
were very important. The evidence also
showed that a brand that achieved success
in one region might not meet with acceptance
in another area. Furthermore, new
entry required significant sunk investments
in brand promotion, as well as
investments in a route delivery network.
For these reasons and more our concerns
were not alleviated by the prospect of
product repositioning or entry.
In the facial tissue and baby wipes
markets, we also found that entry takes
significant time and expense. The baby
wipes market had unique aspects that
warranted an especially close look at the
A R T I C L E S • A N D • F E A T U R E S
2 8 · A N T I T R U S T
dynamics of entry: the most likely
entrants appeared to be firms with wellestablished
brand names in related consumer
products, and entry is surely made
easier by the fact that there is relatively
rapid turnover in demand in this market,
since individual babies, and even parents
with two or more children, soon outgrow
the need for wipes. Nonetheless, entry
into this market has proven difficult, and
the prospect of entry did not alleviate our
concerns.
In both cases, as in all mergers, the
Division analyzed whether a price
increase, or a reduction in quality, would
remain profitable, after accounting for
rivals’ supply-side responses. The history
of brand entry, exit, and positioning,
and the associated costs, is relevant for
this portion of the analysis.
Step 4: Synergies. If a post-merger
price increase is profitable, even after
accounting for rivals’ responses as well
as consumer substitution, the merger is
likely to be anticompetitive. Consumers
likely will be harmed by the combination,
unless it truly offers substantial efficiencies
that lower incremental costs.
Reductions in incremental costs can offset
the incentive to raise price, since the
merged entity, like any firm, will have an
incentive to set a lower price, the lower
are its incremental costs.
To be relevant, the cost savings must
truly stem from synergies specific to the
merger. If one firm alone can achieve
lower costs by expanding its scale of
operations, that should occur through
competition, not merger. Furthermore, if
they are to benefit consumers, the synergies
typically must lower incremental
costs.23
Market Definition and
Market Shares
In my experience the main battlefield in
litigated merger cases is market definition.
However, any attempt to make a
sharp distinction between products “in”
and “out” of the market can be misleading
if there is no clear break in the chain
of substitutes: if products “in” the market
are but distant substitutes for the merging
products, their significance may be overstated
by inclusion to the full extent
that their market share would suggest;
and if products “out” of the market have
significant cross-elasticity with the merging
products, their competitive significance
may well be understated by their
exclusion.
Since we need to define markets to
identify the lines of commerce affected
by the merger, it may be tempting for
counsel to argue for a very broad market
if there is no clear break in the chain of
substitutes. This may be inconsistent,
however, with a careful application of the
Merger Guidelines, which can lead to a
market boundary between very similar
products. In the pens case, testifying on
behalf of Gillette, I agreed with the
Division that one could define a market
boundary based on price, even though
pens were sold at a continuum of prices.
Judge Lamberth accepted this principle in
his opinion.24
In the case of geographic
markets the mere fact that there are gas
stations or supermarkets all over Los
Angeles does not necessarily imply that
the geographic markets for these products
are as broad as the entire L.A. area.
Nor is there a single market for all food,
despite the fact that it is difficult to draw
boundaries between food groups.
On the other hand, an anticompetitive
merger cannot be disguised by arguing
for a very narrow market so as to quarantine
the merging parties in separate
“markets.” Suppose that Brands A and B
propose to merge, but Brand X is situated
between them. Suppose further that a
merger between Brands A and X would
lead to at least a 5 percent price increase,
and likewise for a merger between
Brands B and X. Defense counsel might
be tempted to argue that A and X form a
market, and that B and X form a market,
but that A and B are not in the same market.
Still, if the Diversion Ratio between
A and B is significant (albeit smaller than
between A and X or between B and X),
the merger of A and B could well harm
consumers. In this case, it is appropriate
for the Division to use a price increase
of more than 5 percent in defining the
market, as suggested in the Merger
Guidelines.
Once markets are defined, using the
Guidelines approach there remains the
issue of how to use the firms’ market
shares in differentiated-products markets.
As the Guidelines point out, market share
numbers must be interpreted in conjunction
with evidence about the proximity
of the merging brands. If the merging
brands are close together, the Diversion
Ratio is likely to be high, and any given
level of market concentration is more
troubling. The reverse is true if the brands
are distant.
If all the brands in the market are
roughly equidistant from each other, then
the market shares of various brands will
be proportional to their Diversion Ratios,
making emphasis on the two brands’
market shares very appropriate. The
Guidelines look to both the sum of the
market shares of the merging brands (to
see if this sum exceeds 35 percent in evaluating
unilateral effects) and to the product
of the market shares (which will
reflect the number of consumers who
regard the merging brands as their first
and second choices, and determines the
increase in the HHI).
As illustrated above, if the brands are
equidistant from one another, information
about market shares can be combined
with a measure of the overall market
elasticity of demand to estimate the
Diversion Ratio between the merging
brands. This estimated Diversion Ratio
can then be combined with information
about premerger margins to give at least
a rough estimate of the profit-maximizing
post-merger price increase.
In arguing for a broad market, a common
tactic is to calculate a “critical elasticity”
of demand for a group of products
being considered as a market,25
and then
argue that the true elasticity is above this
critical level, making a 5 percent price
increase unprofitable. This method must
be used with great caution in the context
of differentiated products, to avoid at
least two pitfalls. First, there is no reason
to restrict attention to a uniform price
increase of 5 percent for the purposes of
market definition if a single firm controlling
the entire product category would
find it optimal to increase the prices of
different brands by different amounts.
Second, care must be taken to ensure that
the claimed “market” elasticity is consistent
with information about each brand’s
own elasticity of demand and the crosselasticities
of demand among the products
in the category. Remember, the
“market” elasticity will be lower than the
individual brand elasticities of demand,
and significantly so if the Diversion
Ratios are large. If each brand sells at a
high markup, this is strong evidence of a
A R T I C L E S • A N D • F E A T U R E S
S P R I N G 1 9 9 6 · 2 9
low price elasticity for each brand, which
is inconsistent with a high “market” elasticity
of demand. If the premerger
markups are large and the Diversion
Ratios among the brands are large, claims
of a large “market” elasticity of demand
are not credible.
Unilateral Effects in the Courts
In addition to the growing use of the
approach described here by Antitrust
Division economists, courts have also
incorporated this type of analysis into
their reasoning. For example, in the pens
case, Gillette, Parker Pen, and other firms
such as Cross, Schaeffer, and Mont
Blanc, offered fountain pens, roller ball
pens, ball point pens, and mechanical
pencils at a continuum of prices. I employed
the methods sketched out here
to develop my testimony on behalf of
Gillette, although we lacked data on
prices and quantities for premium pens to
econometrically estimate the Diversion
Ratio. This same general methodology
was employed by George Rozanski in his
analysis for the Antitrust Division.
Although Gillette and the Division agreed
it was appropriate to focus on unilateral
effects, we disagreed about how to interpret
the evidence of substitution between
fountain pens and roller ball and ball
point pens, and on the difficulty of product
repositioning.
Similarly, in the cereals merger
between Kraft and Nabisco,26
economic
experts for both sides, relying on supermarket
scanner data and survey evidence,
spent considerable time estimating elasticities
of demand for the purposes of
evaluating unilateral effects. Judge
Wood’s opinion discussed unilateral
effects at great length, emphasizing the
econometric estimates of cross-price elasticity
between the key merging brands,
Grape Nuts and Shredded Wheat, for
evaluating possible anticompetitive unilateral
effects.
Conclusion
Mergers in markets with differentiated
products may seem a confusing area, and
the case law provides less guidance than
one might like regarding how to define
markets to include “reasonable substitutes.”
The Merger Guidelines, and the
consensus view among economists of
how to analyze competition in differentiated-product
industries, together provide
a consistent, valid, and reliable way
of evaluating proposed horizontal mergers
involving differentiated products.
Central to the analysis are the Diversion
Ratios between the two merging brands,
which measure the fraction of customers
of each brand that consider the merging
brand their second choice. In cases
where detailed price and quantity data
are available, Diversion Ratios can be
calculated based on estimated elasticities
of demand. Alternatively, the
Diversion Ratios can be estimated based
on whatever pieces of evidence are available,
including more qualitative infor-
mation.
If a significant proportion of consumers
regard the merging brands as their
first and second choices, the Diversion
Ratios will be high, and the merger will
indeed create an incentive to raise price,
particularly if premerger Gross Margins
are large. This incentive can be undercut
by rivals’ product repositioning, by entry,
or by credible synergies.●
1
I am referring here to the Cournot model of quantity competition, which
dates back to 1838. See Joseph Farrell & Carl Shapiro, Horizontal
Mergers: An Equilibrium Analysis, 80 AM. ECON. REV. 107 (1990) (applying
this model to mergers); Asset Ownership and Market Structure in
Oligopoly, 21 RAND J. ECON. 275 (1990) (exploring the welfare properties
of this model, with applications to mergers and other transfers of assets
among rivals).
2
In assessing mergers among suppliers of homogeneous products, the main
amendment to a structural approach urged by economics is that it is important
to consider the elasticity of demand facing today’s suppliers in the
aggregate. If entry is easy, or if other distinct products offer good substitutes,
the elasticity facing today’s suppliers of the homogeneous good will
be high. And, if today’s suppliers indeed face highly elastic demand,
anticompetitive effects are less likely to be significant, for any given
market structure. In fact, a strong case can be made that the danger of anticompetitive
effects in these markets can be gauged by the ratio of the HHI
to the market elasticity of demand. One of the many strong points of the
Merger Guidelines is that their “hypothetical monopolist” approach to
market definition explicitly incorporates the aggregate elasticity of
demand into the market definition exercise.
3
It is important to measure incremental cost properly in calculating the premerger
margin. For example, if a firm is facing capacity constraints, one
must include some capital costs in the measure of incremental cost. Also,
the time frame and scale over which incremental costs are measured
should be commensurate with the unilateral effects being studied.
4
Economists have long realized that firms selling differentiated products
have some “market power” in a technical economic sense, although typically
not nearly enough to rise to the level of “monopoly” power. In the
1930s Joan Robinson and Edward Chamberlin developed the theory of
“monopolistic competition” to describe markets in which each firm has a
distinct product, but competes with several or many other firms. JOAN
ROBINSON, THE ECONOMICS OF IMPERFECT COMPETITION (1933); EDWARD
CHAMBERLIN, THE THEORY OF MONOPOLISTIC COMPETITION (1933). By
the 1990s economists have made great progress in understanding
competition with differentiated products, offering a sound foundation for
merger enforcement in such industries.
5
Let me emphasize that the reliability of this analysis depends very much
on the data and other evidence upon which it is based. Ideally, the analyst
should conduct a sensitivity analysis to make sure the results are not overly
sensitive to specific simplifying assumptions that must be made.
6
Bertrand Equilibrium refers to the pattern of prices that prevails if each
firm sets the price of its brands taking as given the prices of the other
brands. This is also known as a “Nash Equilibrium” in prices. Nash
Equilibrium is the dominant method by which economists model “noncooperative,”
i.e., non-collusive, behavior among rivals. Bertrand equilibrium
is not the only possible equilibrium concept, but, absent clear evidence
to the contrary, it is a very useful workhorse.
7
I do not want to leave the impression that the actual analysis precisely
tracks the four steps outlined below, rather these four steps form a conceptual
road map.
8
This analysis will then have to be repeated for Brand B to see if its price
will rise after the merger. If the merged entity has an incentive to raise
either price significantly, the merger may be anticompetitive. Furthermore,
if the firms own multiple brands, the four steps outlined here will need to
be repeated for each individual brand owned by either firm. With multiple
brands, not only must the analysis be repeated, but we must keep track
of diversion to all brands owned by the merger partner when considering
the repricing of any given brand.
9
In practice, economists often estimate the merging brands’ own- and
cross-price elasticities, from which the Diversion Ratio can be calculated.
I find the Diversion Ratio more intuitive and easier to work with, so I
frame the discussion in terms of the Diversion Ratio.
What’s New on the
Internet from the
Antitrust Section
http://www.abanet.org
Have you visited the Antitrust
Section on the Internet lately?
The Section’s site currently contains:
☛ a list of Section committee descriptions
with a committee application
and membership form;
☛ a list of Section publications and
descriptions with an order form;
☛ a calendar of upcoming Section
programs; and
☛ a form for comments and questions
that can be sent to the Section’s
e-mail address:
antitrust@attmail.com.
Visit us at the 44th Annual Antitrust
Spring Meeting! Stop by the Internet
display at the Antitrust Expo and see
what the Section’s home page has to
offer. Speakers’ papers from the 1996
Spring Meeting will be posted and
future developments are planned for the
site. We welcome your suggestions.◆
A R T I C L E S • A N D • F E A T U R E S
3 0 · A N T I T R U S T
10
More generally, the formula is DAB
= (EAB
/EA
)(xB
/xA
), where DAB
is the
Diversion Ratio from Brand A to Brand B, EA
is the own-price elasticity
of Brand A, and xA
and xB
are the unit sales of Brands A and B respectively.
Robert Willig, Merger Analysis, Industrial Organization Theory,
and Merger Guidelines, BROOKINGS PAPERS ON ECONOMIC ACTIVITY:
MICROECONOMICS 281 (1991), also notes that the ratio of the cross-price
to the own-price elasticity measures the share of the marginal sales of one
brand that will divert to another in response to a price increase.
11
In other words, among a group of brands that are all equally “close” or
“distant” substitutes, Diversion Ratios are proportional to market shares.
This is the essence of the “logit” model of demand. See Willig, supra note
10, for a discussion of the role of market shares in differentiated product
markets using the logit model. For further detail on using the logit model
to analyze mergers, see Gregory Werden & Luke Froeb, The Effects of
Mergers in Differentiated Products Industries: Logit Demand and Merger
Policy, 10 J.L., ECON. & ORG. 407 (1994).
12
Of course the effects of a merger cannot be predicted based on data analysis
alone. To be reliable, any data analysis must pass a reality check based
on business documents and the testimony of industry participants.
13
United States v. Interstate Bakeries Corp., Civil Action No. 95C-4194
(N.D. Ill. filed July 20, 1995)
14
United States v. Kimberly-Clark Corp., Civil Action No. 3:95 CV 3055-P
(N.D. Tex. filed Dec. 12, 1995).
15
The elasticities obtained from the econometric estimation of demand
were introduced into an industry model of the pricing of all brands to estimate
the unilateral effects of the merger. Some of the modeling methods
employed are described further in Gregory Werden & Luke Froeb,
Simulation as an Alternative to Structural Merger Policy in Differentiated
Products Industries (Economic Analysis Group Discussion Paper EAG
95-2 Sept. 1995).
16
This figure is obtained by dividing Brand B’s market share of 15% into
the combined share of all the brands to which customers of Brand A turn,
which is 75%. This is an application of the sB
/(1-sA
) formula provided
above.
17
Symmetry means that the two brands have equal unit sales and Gross
Margins prior to the merger and that the Diversion Ratio from A to B is
the same as that from B to A. This formula also assumes that each merging
firm sells a single brand prior to the merger. The analysis is more
involved and the formulas much more complex if the brands are not symmetric
or if the merging firms sell multiple brands prior to the merger.
18
In fact, the constant elasticity assumption becomes logically untenable as
the sum of the Gross Margin and the Diversion Ratio approaches or
exceeds unity. The dangers associated with the constant elasticity assumption
are greater, the larger are m and D.
19
I believe that the percentage price increases predicted by the logit model
tend to be comparable to those of the linear model, although I am unaware
of any comparably simple formula in the two-firm symmetric logit model.
20
The magnitude of the optimal price increase depends not only on the premerger
Gross Margin and Diversion Ratio, but also upon how quickly
elasticity itself rises as price rises above the premerger level. In principle,
this information can be obtained from the data or from other evidence of
how consumers would respond to price changes. Incorporating such
information is an essential “reality check” if simple formulas like the
ones displayed here are to be used.
21
Although it is difficult to generalize, and the Antitrust Division certainly
is not wedded to any single approach, there are some sound reasons to use
formulas based on linear or logit demand systems for these purposes in
preference over constant-elasticity systems. This is what the Division has
done in recent investigations. As I noted above, however, econometricians
often estimate constant-elasticity of demand systems empirically. All I can
say with confidence is that the demand system used to predict post-merger
price increases should, as much as possible, conform to the qualitative
as well as quantitative evidence that is available.
22
Oligopoly theory generally predicts that horizontal mergers will lead to at
least marginally higher prices if they generate no efficiencies, although the
tendency towards higher prices can be thwarted by product repositioning
or entry. This tendency is perhaps clearest in the case of differentiated
products and pricing (Bertrand) competition, where rivals typically will
choose to raise prices if the merging parties do so. See Raymond
Deneckere & Carl Davidson, Incentives to Form Coalitions with Bertrand
Competition, 16 RAND J. ECON. 473 (1985). However, the same result
applies with quantity (Cournot) competition, even though rivals
typically increase output as the merging firms restrict output. Even
with these responses, horizontal mergers under quantity competition
lead to higher prices unless they generate synergies. This is the “No
Synergies Theorem” proven in Farrell & Shapiro, Horizontal Mergers,
supra note 1.
23
The issue of how to measure incremental costs remains, however. This
involves questions of time and questions of scale, at least. The longer the
time frame over which we are looking, the more costs tend to be variable,
or incremental, rather than fixed. And some categories of costs may be
incremental with respect to large customers, but not small ones. In the
extreme case of one customer, such as the Department of Defense for some
weapons systems, cost savings in virtually any category may be passed
along to the customer, at least to some degree.
24
United States v. Gillette Co., 828 F. Supp. 78 (D.D.C. 1993).
25
The “critical elasticity” is the minimum elasticity for which a 5% price
increase would be unprofitable.
26
New York v. Kraft General Foods, Inc., 1995-1 Trade Cas. (CCH)
¶ 70,911 (S.D.N.Y.1995).
http://www.abanet.org
http://www.abanet.org
http://www.abanet.org
http://www.abanet.org
http://www.abanet.org
http://www.abanet.org
http://www.abanet.org
http://www.abanet.org
http://www.abanet.org
http://www.abanet.org
http://www.abanet.org
http://www.abanet.org
http://www.abanet.org
http://www.abanet.org
http://www.abanet.org
http://www.abanet.org
http://www.abanet.org
http://www.abanet.org
http://www.abanet.org
http://www.abanet.org
http://www.abanet.org
http://www.abanet.org
http://www.abanet.org
http://www.abanet.org
http://www.abanet.org
http://www.abanet.org