International Journal of Industrial Organization

23 (2005) 829 848

www.elsevier.com/locate/econbase

Consumer surplus vs. welfare standard in a political

economy model of merger control
Damien J. Neven a,*, Lars-Hendrik Roller b
a
b

Graduate Institute of International Studies, Geneva, Switzerland

Wissenschaftszentrum Berlin and Humboldt University, Germany
Available online 14 October 2005

Abstract
This paper considers the political economy environment that an antitrust agency is operating in and asks
under what circumstances a consumer surplus standard yields higher welfare than a welfare standard. In
particular, we address how institutional settingssuch as transparency and accountabilityinteract with
the choice of an appropriate standard. We consider a framework in which the antitrust agency can be
influenced by third parties (at a cost in terms of real resources) and in which the agency is imperfectly
monitored. A welfare comparison between the two standards reveals that neither standard dominates. The
consumer surplus standard is attractive relative to a welfare standard, when lobbying is efficient, when
accountability is low, where mergers are large and when a marginal increase in merger size is highly
profitable.
D 2005 Elsevier B.V. All rights reserved.
JEL classification: L40; H11
Keywords: Merger control; Political economy; Antitrust policy; Capture

1. Introduction
The purpose of this paper is to evaluate alternative standards that can be assigned to an
antitrust agency in charge of merger control. It is striking that some of the major antitrust
agencies appear to operate with objectives that differ from welfare maximization. In particular,
both the U.S. as well as EU merger control can be interpreted as maximizing consumer surplus
rather than aggregate welfare.1
* Corresponding author.
E-mail address: neven@hei.unige.ch (D.J. Neven).
1
See for instance Gellhorn and Kovacic (1994).
0167-7187/$ - see front matter D 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.ijindorg.2005.08.011

830

D.J. Neven, L.-H. Roller / Int. J. Ind. Organ. 23 (2005) 829848

In a world with no regulatory failures, excluding firms profits from the objectives assigned to
the antitrust authority is hard to justify.2 However, in a setting where institutional and political
economy considerations are taken into account, the welfare implications of various standards
that are given to the antitrust authority are more subtle. In particular, it may be welfare enhancing
to assign a standard that is not first-best.
This paper considers the political economy environment in which an antitrust agency operates
and asks under what circumstances a consumer surplus standard yields higher welfare than a
welfare standard.3 In particular, we address how institutional settingssuch as transparency and
accountabilityinteract with the choice of an appropriate standard. We model the political
environment that the antitrust agency is subject to through a common agency framework (a` la
Bernheim and Whinston, 1986).4 In this context, interested parties provide inducements to the
antitrust agency which are contingent on the outcome of the merger review.
More specifically, we consider a situation that can be characterized by a three stage process.
In the first stage, a merger is notified and the interested parties provide contingent bids to the
agency. In order for the merger to be potentially welfare enhancing we assume that there are
efficiency gains, which are known to the firms and the antitrust agency.5 We consider three
interested parties: consumers, the merging firms, and the (non-merging) competitors. Consumers
do not lobby the antitrust agency. This may arise for at least two reasons. First, consumers may
not be well informed about the consequences of proposed mergers (we will assume that
consumers cannot observe the characteristics of the merger) and accordingly may not be able to
formulate appropriate contingent bids. Second, consumers may face prohibitive transaction costs
in representing their interests. These costs could be associated with the traditional problems of
free-riding and collective action with numerous agents.
We further allow for an inefficient lobbying technology, which we associate with
transparency. With greater transparency influence activities have to take indirect routes which
are typically less efficient. For instance, influence takes place through indirect means like
expensive lunches or the promise of lucrative jobs in the private sector (the brevolving doorQ).
Whereas pure transfers do not entail any efficiency losses, indirect means of influencing the
agency typically involve some real resource cost.6
In the second stage, the antitrust agency decides whether or not to allow the proposed merger.
The agency observes the bids and the characteristics of the merger. Most importantly, the
antitrust agency is held accountable to the standard that has been assigned by law (i.e. either
2

There is a large literature on what the goals of antitrust should be, see for instance Harberber (1971) and Scherer
(1993). In this paper we concentrate on consumer surplus and net social welfare, as those are the objectives that have
received the most attention.
3
We will use the term welfare standard for what is commonly referred to as net social welfare standard.
4
The common agency framework has been used to model a number of other policy decisions. See for instance,
Grossman and Helpman (1994) or Rama and Tabellini (1998). Policy choices have also been analysed using models of
representative democracies, in which agents with different policy preferences come forward in the electoral process (see
Besley and Coate (1997)). Recently Besley and Coate (2000) have also proposed a model which encompasses both
approaches. Empirical evidence suggests that influence by interested parties is an essential feature of merger control (see
for instance Neven et al. (1994)). Political competition will also matter at the time when merger law are designed but may
be less important. Even though it would be desirable to consider both aspects, our model thus focuses on what is arguably
the more important aspect of political economy interactions in merger control.
5
We therefore do not focus on political economy issues surrounding the befficiency defenseQ debate, i.e. where firms
may have some private information about the efficiencies of the merger. For a paper that analyzes informational lobbying
see Lagerlof and Heidhues (2002).
6
See also Posner (1975).

D.J. Neven, L.-H. Roller / Int. J. Ind. Organ. 23 (2005) 829848

831

consumer surplus or welfare standard). The incentive for the agency to follow the standard is
introduced into the common agency framework through monitoring. We use the simplest
possible monitoring technology, in which a review of the agency takes place with some
exogenous probability. Once a review is launched, a penalty on the agency is imposed whenever
the decision of the agency is not in line with the assigned standard. We will assume that the
penalty is linear and increasing in the difference between the optimal outcome associated with
the assigned standard and the actual outcome. Under those circumstances we show that the
objective function of the agency is a weighted sum of the assigned standard and the bids by the
interested parties.
In the final stage, product market competition occurs and bids and profits are realized.7
Rather than specifying this stage explicitly, it suffices that some general monotonicity
assumptions on the reduced form profit functions, as well as consumer surplus, are met. In
equilibrium, firms provide perks anticipating the decision taken by the antitrust agency and the
profits that will accrue from the outcome of this decision.
We characterize the equilibrium under two alternative standardswelfare and consumer
surplus8for a range of possible mergers indexed by their efficiency level. In general, we find
that in terms of welfare neither standard is first-best. More specifically, we show that the
consumer surplus is attractive relative to a welfare standard, when lobbying is efficient, when
accountability is low, where mergers are large and when a marginal increase in merger size is
highly profitable. The findings suggest that a reform of the standard that is assigned to a
competition agency needs to consider the political economy in which the antitrust agency
operates. For instance, our model implies that it may not be appropriate for the EU to move
towards a welfare standard unless the transparency and the accountability of its procedures are
further improved.
Our model is closely related to the work of Laffont and Tirole (1991) and Besanko and
Spulber (1993). Laffont and Tirole (1991) also analyze the extent to which a regulatory agency
can be manipulated by its constituencies. Their model is however designed to capture the
essential features of the regulation of utilities rather than merger control. In their model, there is a
regulatory agency which has an imperfect signal about the characteristics of a firm which is
regulated by the government. The firm can provide perks to the agency in order to induce misreporting of its characteristics by the agency. In equilibrium, the government can prevent
collusion and offers a low powered incentive contract to the firm as well as a remuneration for
the agency which is contingent on its report. Besanko and Spulber (1993) consider a model of
merger control in which the characteristics of the merger cannot be ascertained by the antitrust
agency. In their model, the antitrust agency cannot be influenced by the affected parties. They
also evaluate alternative objective functions that the government might assign to the antitrust
agency. They find that the optimal rule should give more weight to consumer surplus than
profits.
We present the model in Section 2 and derive the equilibrium in Section 3. The comparative
statics with respect to the size of the merger and transparency is presented in Section 4. Section 5
concludes.
7

We assume that contingent perks can be enforced. As discussed by Laffont and Tirole (1991), such contracts should
not be thought of as being enforced by Courts. Rather, these contracts are quasi-enforceable and their implementation
rests on the parties willingness to abide by their promises. bRevolving doorsQ or stock market options could serve as
devices to implement contingent bids.
8
We focus on these polar cases (rather than consider a continuum of objective functions characterised by different
weights for profits) because intermediate cases are presumably hard to implement in practice.

832

D.J. Neven, L.-H. Roller / Int. J. Ind. Organ. 23 (2005) 829848

2. The model
We consider an industry with N firms. A merger in this industry is characterized by the
number of firms involved, M, and by the level of efficiency, e, which is achieved by the merged
entity. The efficiency e can be thought of as the reduction of the marginal cost accruing to the
merging firms, and is observable by all players. We therefore abstract from informational
lobbying in this paper.
The structure of the game is illustrated in Fig. 1. In stage 3, firms compete given the decision
that has been taken by the agency. Profits and bids are realized.
In stage 2, the agency decides on the merger. The decision of the authority is either to ban
(D = 0) or to allow (D = 1) the merger (no remedies are allowed). Let F = F m + F c be the sum of
the merging parties bids ( F m ) and the competing firms bids ( F c ), which are contingent on the
decision taken by the agency. We assume that a fraction of the bids (1
d) are wasted, so that
only dF(D) will be transferred to the agency. As discussed above, a large value of 1
d is
interpreted as indicative of high transparency.9 Denote the probability that the antitrust agency be
audited by c and the penalty by B(D).
Then the antitrust agencys expected utility u(D) is given by:
u D dF D
cB D:

In other words the agency maximizes the share of the bid which is effectively transferred less
the expected penalty. We consider two types of standards, namely consumer surplus (CS) and
welfare standards (W). Denote O(D) a {W(D),CS(D)} as the value of the standard as a function
of the decision taken by the agency. Furthermore denote O* is the maximum value of the
standard, i.e. O4 maxD O D. We will assume that the penalty is linearly increasing in the
difference between the optimal outcome and the actual outcome under a particular standard, i.e.
we assume that B = b(O*
O(D)). Substituting into the agencys utility function yields:
u D dF D
cbO4
O D:
Since O* does not depend on the decision D, the decision by the antitrust agency is given by:
max U D O D

Daf0;1g

d
F D:
cb

Hence, given this auditing process, the agency takes both the standard as well as the
contingent bids into account. The weight attached to these two objectives is determined by the
d
parameter au cb
, which is associated with the auditing process (accountability) and the lobbying
efficiency (transparency). As can be seen, the relative importance of the contingent bids in the
objective function of the agency decreases in transparency (low d) and in accountability (high b
and c).
For ease of reference, the objective function of the agency when it has been assigned a
welfare standard will be written as U 1. When the agency is endowed with a consumer surplus
standard, its objective function will be denoted U 2.
9
The waste associated with lobbying could also be affected by the number of agents belonging to any group with
common interests. The waste associated with lobbying by the merging firms (competitors) could thus be made dependent
on the number of merging firms (competitors), which would not affect any of the substantive results below.

D.J. Neven, L.-H. Roller / Int. J. Ind. Organ. 23 (2005) 829848

833

Firms choose bids (Fm, Fc)

Antitrust Agency either

blocks or allows the
Merger
Allow

Ban

Product Market Competition

with N-M+1 firms

Product Market Competition

with N firms

Fig. 1. Structure of the game.

In stage 1, firms provide contingent bids to the agency. The objective function of the firms are
the changes in their net profit, namely the change in their profit minus the bids that they provide
to the agency.
3. The equilibrium
We look for subgame perfect equilibria and restrict ourselves to truthful equilibria, i.e.
equilibria in which the firms bid for any policy reflects the marginal benefit that they obtain
from the implementation of this policy.10
We start by considering stage 3.
3.1. Product market competition in stage 3
Let DP *m denote the change in aggregate gross equilibrium profits of the merging firms and
DP*c denote the change in aggregate gross equilibrium profits of the competitors (that is the
profits before bids are deducted), if the merger is realized. Let DCS* denote the change in
consumer surplus. Note that if the merger is banned, the change in profits and consumer surplus
is simply zero. For notational convenience, we will drop the D as well as the d*T notation below,
such that we use P m for DP m*, P c for DP*,
c and CS for DCS*.
We do not explicitly specify the market game but assume that the following properties of the
profit functions as well as consumer surplus hold in equilibrium:11
A1 :
A2 :
A3 :
A4 :

BPm Pc
BPc
BPm
Be N0; Be b0;
Be
BCS
Be N0
BPm Pc CS
b0
BM
BPm Pc
N0
BM

N0
:

10
There are potentially many equilibria in common agency games. As discussed by Bernheim and Whinston (1986),
truthful equilibria are however attractive because they are coalition proof (and vice versa). They are also efficient in the
sense that the alternative selected in a truthful equilibrium maximises the sum of the payoffs of the players. Kirchsteiger and
Prat (2000) have explored another class of equilibria in common agency games, which they refer to as bnatural equilibriaQ. In
a natural equilibrium, each principal offers only one positive contribution to the common agent. In our model, firms and
competitors only offer one positive contribution so that in our context, truthful equilibria are also bnaturalQ.
11
We also assume that they are continuous and differentiable.

834

D.J. Neven, L.-H. Roller / Int. J. Ind. Organ. 23 (2005) 829848

m + c + CS

CS

Fig. 2. Efficiency, profits and welfare.

Hence, we assume that the profitability of a merger increases with the level of efficiency that
it can achieve. The profits of competitors fall with this level of efficiency but the industry profit
still rises. We also assume that as the efficiency gains increase, the equilibrium price falls so that
the consumer surplus increases. These assumptions (A1 and A2) imply that welfare increases in
line with the efficiency gain. It is straightforward but tedious to check that these properties
actually hold for a Cournot model with homogenous products. The last two assumptions (A3 and
A4) restrict our analysis to situations where larger mergers increase industry profits and reduce
welfare.12
Beyond our maintained assumptions about the profit function A1A4, we further assume
that the competitors (whose profit falls with efficiency) and the consumers (whose surplus
increases with efficiency) do not change for some identical value of the efficiency gain denoted
by eV. In other words, aeV N 0, s.t. P c(eV) = CS(eV) = 0. This property implies that competitors are
unaffected by the merger when the equilibrium does not change. This property holds for a
Cournot model with homogenous products (see Bond, 1996). Finally, we restrict attention to
the range of efficiency parameters which ensure that mergers are profitable (i.e. such that
P m N 0).
It is useful to note the impact of a merger on the interests of merging firms, competitors and
consumers. Fig. 2 illustrates these effects as a function of the efficiency achieved by the merger.
Note that consumers and the competing firms never have congruent interests. When efficiency is
such that the price increases after the merger, the interests of the merging firms and their
competitors are aligned. By contrast, when efficiency is large enough to guarantee that the price
falls, merging firms and consumers both benefit, while competitors loose.
For further reference, it is also useful to define the efficiency level which guarantees that the
change in total welfare is unaffected by the merger. Denote the change in welfare at a given
efficiency gain by S(e) = P c(e) + P m(e) + CS(e). Note that by A1 and A2, S(e) is increasing in e.
Define the efficiency level at which total welfare is unchanged by the merger as e such that
S (e ) = 0 (see also Fig. 2).
Note that the change in welfare at eV is positive, since P c(eV) = CS(eV) = 0 and P m(eV) N 0. The
last inequality holds, since at an efficiency level of eV, price and output are unaffected by the
12

We abstract from the fact that M can only take integer values.

D.J. Neven, L.-H. Roller / Int. J. Ind. Organ. 23 (2005) 829848

835

merger. Total revenues are thus unaffected but total costs fall in line with the efficiency gain so
that the change in profits has to be positive. Furthermore, given the monotonicity of welfare
(through A1 and A2), we have that e b eV.
Our assumptions with respect to the size of the merger (A3 and A4) can also be illustrated in
Fig. 2. A4 implies that the sum of firms profits is shifted upwards for larger mergers. By A3,
consumer surplus is shifted downwards and the efficiency level for which welfare is unaffected
(e) moves to the right.
3.2. Decisions by the antitrust authority in stage 2
In stage 2, the agency decides on the merger. In order to avoid unessential complexities, we
assume that if the agency is indifferent between allowing and prohibiting a merger, it will decide
to allow it. The equilibrium at this stage is straightforward: the agency simply compares the
levels of utility that it achieves under each outcome and selects the outcome that yields the
highest level.
Consider first the welfare standard. In this case, the agencys utility if it allows the merger is
given by: U 1(D = 1) = P m + P c + CS + aF(D = 1). Similarly, the utility if the agency blocks the
merger is given by U 1(D = 0) = aF(D = 0).
For the consumer surplus standard, the agencys utility if it allows the merger is given by:
U 2(D = 1) = CS + aF(D = 1), while the utility if the agency blocks the merger is given by
U 2(D = 0) = aF(D = 0). Therefore, the decision by the agency is characterized by,
D 1 if and only if Ui D 1 NUi D 0
where i = 1, 2 indicates the welfare and consumer standard, respectively.
3.3. Lobbying decisions at Stage 1
We derive equilibrium bids for both welfare and consumer surplus standards. Note that the
bids are contingent on the actions of the antitrust agency. We begin with the welfare standard.
3.3.1. Welfare standard
Assume that the agency is assigned the welfare standard. In order to solve for the equilibrium
bids, it is convenient to distinguish among various parameter regions with respect to the
efficiency level. As noted above, when e b eV, both the merging firms and their competitors
benefit from the merger and hence will never bid to influence the antitrust authority against the
merger. We therefore have that the sum of the bids F(D = 0) = 0.
We first consider the firms incentives to bid in favor of the merger for regions e V eV.
Consider the region where e V e V eV. This is the range of efficiency for which mergers do not
increase the consumer surplus but do not reduce welfare.
Lemma 1. Let e V e V e. Firms will not bid and the merger is allowed.
Proof. The agency will allow the merger without any bids, since U 1(D = 1) = P m +
P c + CS z U 1(D = 0) = 0. Therefore, in equilibrium, firms will not bid. 5
In this region, the efficiency of the merger is such that welfare does not fall. Hence, firms do
not have to bid in order to influence the antitrust agency, which allows the merger without any
inducement.

836

D.J. Neven, L.-H. Roller / Int. J. Ind. Organ. 23 (2005) 829848

Let us now focus on efficiency levels below e . In this region, the change in welfare is
negative and firms (which benefit from the merger) will have to provide incentives to the agency
if they want the merger to be allowed. Given the decision of the agency in stage 2, firms will
have to ensure that U 1(D = 1) z 0. Consider the highest amount that firms may have an incentive
to bid, i.e. their entire profit. The resulting value of the utility of the agency if the merger is
allowed is then given by S 1(e) = P m + P c + CS + aP m + aP c. Let e 1 be the efficiency level such
that S 1(e 1) = 0.
Lemma 2. For e b e1, firms will not bid and the merger is blocked. For e1 V e b e, firms bid such
that aF =
(Pm + Pc + CS) and the merger is allowed.
Proof. Note that S 1(e) is monotonically increasing and continuous in e by A1 and A2. Since
S 1(e ) N 0, we have that e 1 b e and that e 1 exists and is unique (assuming that S 1 (e) b 0 for some
possibly negative e). Let e b e 1 such that S 1(e) b 0, which implies that U 1(D = 1) b 0 for the
maximum bids. Therefore, the merger is blocked and it is optimal for firms not to bid. Let
e 1 b e b e , which implies that S 1(e) N 0. Any pair of bids F m(D = 1) and F c(D = 1) such that
U 1(D = 1) = aF + P m + P c + CS = 0 is an equilibrium.13 The merger is allowed. 5
Hence, whenever efficiency is insufficient to guarantee that the merger will increase
welfare, firms have to provide incentives to the agency in order to have the merger waved
through. However, the profit of the merging firms and its competitors increase in line with
the level of efficiency. There is thus a range of the efficiency parameter (below that which
guarantees no change in welfare) for which firms have sufficient profit to provide adequate
incentives to the antitrust agency. In this region, we consequently have two types of
inefficiencies, which occur simultaneously. The first one arises because the merger is pushed
through by the lobbying activity of firms, even though it reduces welfare. We therefore have
a type II error. In addition, there is bidding in equilibrium and this entails some waste.
It is worth noting at this point that the agency does not obtain any rent from the political
economy interactions because firms always provide just enough incentives to make the antitrust
agency indifferent between allowing and prohibiting the merger. It is indeed a standard feature
of equilibrium in common agency games that the agent obtains positive rents from the
interactions only if the principals have divergent interests (see for instance Grossman and
Helpman, 1994).
We now consider the region where the incentives of the firms are not aligned, i.e. when e is
above eV.
Lemma 3. For e N eV, the merger is allowed. In equilibrium the merging firms bid
Fm(D = 1) = max{0,
aPc
(Pm + Pc + CS)}.
Proof. Given that competitors are hurt by the merger, we must have that F c(D = 1) = 0, which
implies that the agency will allow the merger when U 1 (D = 1) = P m + P c + CS +
F m(D = 1) z U 1(D = 0) = aF c(D = 0). Note that by A1, we have that P m N
P c, which implies
that in equilibrium F m(D = 1) b P m, i.e. the merging firms are always able to push the merger
through. Assume that P m + P c + CS b
aP c so that competitors could have the merger

13

Note that there is a continuum of equilibria in this range. The comparative statics results that we discuss below are
however valid for any equilibrium in this continuum. Multiple equilibria also arise in other parameter ranges discussed
below but the comparative statics is never affected. The multiplicity of equilibria is thus unimportant in the present
context.

D.J. Neven, L.-H. Roller / Int. J. Ind. Organ. 23 (2005) 829848

837

prohibited in the absence of bid by the merging firms. F m(D = 1) =
aP c
(P m + P c + CS) is

then a best reply to F c(D = 0) =
aP c. F c(D = 0) =
aP c is a (weak) best reply to
F m(D = 1) =
aP c
(P m + P c + CS). In equilibrium, the maximum bid of the competing firms
is neutralized by the merging firms. Next, assume that P m + P c + CS N aP c. Then F m(D = 1) = 0
is an equilibrium since the maximum bid of the competing firms cannot block the merger.
Furthermore, any contingent bid by competitors F c(D = 0) is a (weak) best reply. 5
In this region, the level of efficiency is such that competitors are harmed by the merger and
would want to influence the antitrust agency to block it. We therefore have opposing interests on
the part of the firms. In principle, the competitors might loose more from the mergers than what
the merger entails in terms of welfare gains, i.e. it is possible that P m + P c + CS b
aP c. In this
case, where bcompetitors are badly hurtQ by a merger, they are able to compensate the agency for
the loss of welfare that would arise if the merger is prohibited. However, as the above lemma
shows, the merging firms are always able to neutralize the bids by the competitors, which
implies that the merger will always go through. The agency does not make an error but there is a
social cost associated with lobbying. When the maximum bid of the competitors could not
overturn the merger even if the merging firms do not bid (P m + P c + CS b
aP c), the merging
firms do not bid and there is no waste.
In sum, we find that relative to the adjacent parameter range (i.e. to the right of eV), the
emergence of opposing interests among firms does not change the outcome of the merger
decision, but may introduce lobbying activity, and hence some inefficiency. When there is no
interest which dominates, the merging firms, which can always trump the competitors, has to
lobby.
We now turn to the alternative standard.
3.3.2. Consumer surplus standard
Assume that the agency is assigned the consumer surplus standard. As before, we first focus
on the parameter region for which firms incentives are aligned (e V eV), such that F(D = 0) = 0.
We first consider the region such that e V e , i.e. where efficiency is not sufficient to guarantee
that the change in welfare is positive.
Lemma 4. Let e V e. Firms will not bid and the merger is blocked.
Proof. The agency will allow the merger iff U 2(D = 1) = CS + aF(D = 1) z 0. This cannot hold
since P c(e) + P m(e) + CS(e) V 0 for e V e . The merger is blocked and it is optimal for firms not to
bid. 5
In this region, both the change in consumer surplus and welfare are non-positive. Hence,
firms do not have sufficient resources to compensate the antitrust agency for the loss of
consumers surplus that a merger would entail. As a result, firms do not bid and the merger is
prohibited.
Let us now focus on efficiency level above e , such that e b e b eV. In this region, the change in
consumers surplus is still negative and firms might be able to provide enough incentive to the
agency in order to get the merger approved. Given the decision of the agency in stage 3, firms
will have to ensure that U 2(D = 1) z 0. Consider the highest amount that firms can bid, i.e. their
entire profit. The resulting value of the utility of the agency if the merger is allowed is then given
by S 2(e) = CS + aP m + aP c. Let e 2 be the efficiency level such that S 2(e 2) = 0. Note that S 2(e) is
continuous and monotonically increasing in e by A1 and A2. Since S 2(e) b 0 and S 2(eV) N 0 it
follows that e b e 2 b eV and that it is unique.

838

D.J. Neven, L.-H. Roller / Int. J. Ind. Organ. 23 (2005) 829848

Lemma 5. For e b e b e2, firms do not bid and the merger is blocked. For e2 V e b e V, firms bid
such that a F = -CS and the merger is allowed.
Proof. Let e b e b e 2. In this case, S 2(e) b 0, which implies that the merger is blocked even if firms
bid their entire profits. Hence, firms will not bid and the merger is blocked. Let e 2 V e b eV, which
implies that S 2(e) z 0. Any pair of bids F m(D = 1) and F c(D = 1)such that U 2(D = 1) = aF + CS = 0
is an equilibrium. The merger is allowed. 5
As efficiency increases beyond the level at which welfare is unchanged, profits increase and
the harm to consumers falls. There is a region (e b e b e 2) for which the profits are still
insufficient to provide adequate incentives to the antitrust agency and the merger is prohibited
even though it would increase welfare. In this region, there is a type I error but no social waste
associated with lobbying. Beyond this level of efficiency (e 2 V e b eV), the merger is allowed but
only because firms provide adequate inducement, which involve some social waste. As the
efficiency level approaches the level for which consumer surplus is unaffected, the bids and
hence the social waste converge to zero.
We now consider the region where the firms incentives are not aligned, i.e. such that e is
above eV.
Lemma 6. For e z eV, the merger is allowed. The merging firms bid Fm(D = 1) = max{0,

aPc
CS}.
Proof. See the proof of Lemma 3. 5
In this region, the level of efficiency is such that competitors are harmed by the merger and
would want to influence the antitrust agency to block the merger. In the case where bcompetitors
are badly hurtQ by a merger (CS b
aP c), they are able to compensate the agency for the loss in
consumer surplus that would arise if the merger is prohibited. In this case, the merging firms, which
can always trump the competitors, need to compensate for the maximum bid that competitors can
make. The agency makes no error in its decision but lobbying leads to social waste and some rent
for the antitrust agency. When the maximum bid of competitors cannot compensate for the change
in consumer surplus, the merging firms do not need to bid and there is no waste.
In sum, we find (as in the case of the welfare standard) that the merger is allowed despite
opposing interests. Under the consumer surplus standard, the merger is pushed through, to the
left of eV, by a joint action of the merging firms and their competitors. To the right, the merger
might be pushed through again by the action of the merging firms which compensate for that of
the competitors. When lobbying takes place, some inefficiency will also arise. Comparing the
condition in Lemma 6 with Lemma 3, we find that the condition for a bwastefulQ equilibrium to
exist is stricter for the welfare standard. This implies that there exist a parameter region for
which equilibrium under the welfare standard does not involve any bidding, while the
equilibrium under the consumer surplus standard does involve bidding (and therefore waste).
The reverse is not true. In addition, the bid of the merging firms under the consumer surplus
standard ( F m(D = 1) =
aP c
CS) is always higher than the bid under the welfare standard
( F m(D = 1) =
aP c
(P m + P c + CS)).
4. Welfare analysis
In this section, we compare welfare under the two standards. We begin by defining the change
in welfare from a merger by adding up the change in expected utility of all agents, i.e. firms

D.J. Neven, L.-H. Roller / Int. J. Ind. Organ. 23 (2005) 829848

839

(P m + P c
F), consumers (CS), and the antitrust agency (aF - cB), as well as the penalty (cB)
yielding,
W 4 Pm Pc CS
1
dF:

.
In order to provide a benchmark, we briefly consider the first best outcome, denoted by W
Under the first best, the social planner would not allow any bidding in order to set the waste to
= 0. For e z e , a merger is allowed so that
zero. For e b e , the merger will be banned, so that W
= P m + P c + CS.
W
The next two lemmas characterize the change in welfare stemming from a merger under the
two standards. Denote W 1 and W 2 as the change in equilibrium welfare under the welfare
standard and consumer surplus standard, respectively. For example, whenever the merger is
blocked in equilibrium (which also implies that no bidding takes place), we have that
W 1 = W 2 = 0. Given the results of Lemmata 3 and 6, we will assume that CS N
aP c, which
ensures that firms will not bid when e N eV under either standard.
Lemma 7. The welfare effect of a merger under the welfare standard is given by
0
1
0
if
ebe1


W1 @ Pm Pc CS 1 1
d
if
e1 Vebe A
d cb
Pm Pc CS
if
e Ve
Proof. The first and third statements follow directly from respectively Lemma 2 and Lemma 1.
Consider the second statement; from Lemma 2, it follows that in equilibrium firms bid such that
a( F m + F c) =
(P m + P c + CS). Substituting this into W* and rearranging yields the expression
in the lemma. 5
The lemma illustrates the difference between the equilibrium welfare under a welfare standard
. For relatively low efficiency levels (e b e 1) and for relatively high
(W 1) and the first best W
). However, for intermediate
efficiency mergers (e V e), the welfare standard is first-best (W 1 = W
levels of efficiency (e V e b e ), the welfare standard is not first-best. Recall that in this case, the
= 0, while in
first-best involves no bidding and that the merger be blocked, yielding W
equilibrium
bidding
occurs
and
the
merger
is
allowed,
yielding
W

P
1
m Pc CS



1 1
d
cb
.
As
a
result
W
b
0
(since
welfare
is
negative
for
e
b
e
).

1
d
The previous lemma further illustrates that there are two types of inefficiencies that occur
simultaneously in this equilibrium: (i) the merger is allowed even though it yields a negative
welfare, so that there is a type II error, and (ii) there is waste associated with the bids. The first
inefficiency is equal to P m + P c + CS, while the second inefficiency from bidding is given by the
factor 1 1
d
d cb which is larger than 1, unless the bidding technology is perfect (d = 1).
Lemma 8. The welfare effect of a merger under the consumer surplus standard is given by
0
1
0
if
ebe
2

W2 @ Pm Pc CS 1 1
d
if e2 VebeV A:
d cb
Pm Pc CS
if
eVVe
Proof. The first and third statements follow directly from respectively Lemmata 4 and 5.
Consider the second statement; from Lemma 5, it follows that in equilibrium firms bid such that
a( F m + F c) =
CS. Substituting this into W* and rearranging yields the expression in the
lemma. 5

840

D.J. Neven, L.-H. Roller / Int. J. Ind. Organ. 23 (2005) 829848

As the lemma shows, the consumer surplus standard does not achieve the first best for two
regions of efficiency. When e V e b e 2 a merger should be allowed, yet the agency will block the
merger under a consumer surplus standard. There is loss due to a type I error, amounting to
P m + P c + CS. However, there is no bidding, and hence no loss from the inefficient lobbying
technology. The second region where the consumer surplus is not first-best is when e 2 V e b eV. In
this case, mergers are allowed under a consumer surplus standard, which is first best. However,
there is a loss due to bidding, which can be seen as follows. Using the equilibrium bidding
condition from Lemma 5 once again, i.e. a( F m + F c) =
CS, we can equivalently write W 2 for
the region e 2 V e b e as P m + P c + CS
(1
d)( F m + F c). In other words welfare is reduced by the
waste from the bidding process.
The previous results are further illustrated in Fig. 3 which graphs the difference in equilibrium
welfare under the two rules, i.e. W 1
W 2. When efficiency gains are very small (i.e. e b e 1), the
merger is banned under both standards, no bids are put forward, the two standards are equivalent
= W 1 = W 2 = 0). Similarly, when eV b e, mergers are allowed and no
and yield the first best (i.e. W
bids needed under either standard, yielding the first-best. We now consider the more interesting
intermediate cases.
When e 1 V e b e, mergers are still prevented under the consumer surplus standard, no bids are
= W 2 = 0). By contrast, under the welfare standard,
put forward and the first best is achieved (W
firms bid and the merger is allowed. As discussed above this introduces two types of
inefficiencies: (i) there is a type II error, and (ii)
there is waste
 associated with the bids. We
therefore have that W1
W2 Pm Pc CS 1 1
d
cb
, which is negative. Since the
d
welfare loss is decreasing in the efficiency of the merger W 1
W 2 is monotonically increasing in
the range. At e = eP m + P c + CS = 0, and W 1
W 2 = 0 (see Fig. 3).
When e V e b e 2, under the welfare standard firms do not bid and the merger is allowed so that
= P m + P c + CS). Under the consumer surplus standard, the
the first best is achieved (W 1 = W
merger is blocked and we have a type I error. We thus have W 1
W 2 = P m + P c + CS, which is
monotonically increasing in e.
Finally, when e 2 = e b e, the first-best is achieved under the welfare standard, such that
W 1 = P m + P c + CS. Under the consumer surplus standard, there is no type I error but there is
waste associated with bidding. Using Lemma 8, we get that W1
W2
CS 1
d
d cb
1
dFm Fc , which is positive. Since CS is decreasing in e, W 1
W 2 is monotonically
decreasing in e in this region (see Fig. 3).
w1-w2

(1 )(Fm + Fc ) = CS
+CS
e1

e2

( + CS ) 1 +

1
b

Fig. 3. W 1
W 2.

1
b

b =1

D.J. Neven, L.-H. Roller / Int. J. Ind. Organ. 23 (2005) 829848

841

To complete the graph in Fig. 3, we need to show whether the function W 1
W 2 is continuous
at e = e 2. Note that approaching e 2 from the right, we have that the entire profit is bid, such that
a( F m + F c ) = a(P m + P c ) =
CS. Substituting

 this into W 1
W 2 , we can express
cb
1
W1
W2 Pm Pc CS
Pm Pc d cb . Recall that to the left of e 2 we have that
W 1
W 2 = P m + P c + CS. Therefore W 1
W 2 is continuous at e = e 2,when cb = 1, which
corresponds to a situation where the expected penalty is equal to the loss in welfare. Moreover,
when cb b 1, then W 1
W 2 has an upward discontinuity, while cb N 1 yields a downward
discontinuity (see Fig. 3).
Overall, it appears that neither standard dominates over the entire range of parameters, even
though one of the two standards is always first best for any given efficiency level. Consumer
surplus and welfare standards give rise to different types of costs depending on the efficiency
level: when e b e then the consumer surplus standard yields higher welfare, while the opposite
holds for e N e .
On the one hand, a number of relatively inefficient mergerswhich decrease welfareare
pushed through under a welfare standard. This result accords with intuition; when the agency is
held accountable to a welfare standard, but firms can influence the agency, one would expect the
outcome to be biased in favor of firms and against consumers. With a welfare standard, lobbying
activity thus lowers welfare.14
On the other hand, some relatively efficient mergers (which would increase welfare) are
prohibited under the consumer surplus standard. When the agency is held accountable to defend
consumers interest, firms interests tend to be compromised and mergers that enhance welfare
are prohibited. Firms lobbying activity will tend to compensate for the narrow objective of the
agency and thereby reduce the range of relatively efficient mergers that are prohibited. Without
lobbying, all mergers in between e and eV would be prohibited with a loss in welfare equal to
P m + P c + CS. Assuming that the accountability is relatively tough (i.e. cb z 1, which includes
cb = 1, where the expected penalty is equal to the welfare loss.), then the welfare loss with
lobbying is strictly smaller for the region between e 2 and eV (see Fig. 3). Under the consumer
surplus standard, lobbying thus strictly improves welfare. Below we will assume that cb z 1, in
other words we concentrate on monitoring where the expected penalty is at least as large as the
welfare loss.
A comparison between the two standards over the entire range of parameters can also be
undertaken. However, such a comparison depends on the distribution of the efficiencies of the
mergers that the agency will face.15 For instance, if mergers are likely to be very efficient, the
welfare standard dominates. At the opposite, when mergers are rather inefficient on average, a
consumer surplus standard will be preferred.
Rather than making explicit assumptions about the distribution of expected mergers and the
allocation of lobbying effort and derive an explicit comparison between the two standard, we
focus on comparative statics. In particular, the next subsection examines how the comparison
14

Note however that if all interests were equally represented and equally efficient in their lobbying, the welfare standard
would not give rise to any error. As discussed above, it is a standard feature of common agency games that when the
agent has a welfare standard and the principals bid their marginal benefit, the agent mimics the behaviour of an
independent agent which would simply maximise welfare (see for instance Grossman and Helpman, 1994). Hence, if
consumers had equally efficient access to the antitrust agency, all welfare enhancing mergers would be allowed in our
model and all welfare decreasing mergers would be prohibited. However, the first best would still not be obtained as a
potentially large amount of waste would occur because of lobbying.
15
Moreover, the investigation costs may vary across the two standards.

842

D.J. Neven, L.-H. Roller / Int. J. Ind. Organ. 23 (2005) 829848

between the two standards would be affected by a change in transparency (d), the accountability
of the agency (cb) and a change in the size of the mergers (M).
4.1. The trade-off between welfare and consumer surplus standards
We first investigate the trade-off between the two standards with respect to transparency and
accountability. In the context of our model, a more transparent lobbying process is thus
associated with a lower d. We characterize accountability through cb, which represents the
expected penalty of the antitrust agency if it does not follow the assigned standard. For instance,
when cb = 1, then the expected penalty is equal to the welfare loss. An increase in cb is therefore
interpreted as higher accountability.
The following proposition shows how the efficiency regions are affected, i.e. how e 1,e 2, e and
eV change with d and cb. Recall from the definitions of e 1, e 2, e and eV that the impact is given
d
through a (where au cb
). With respect to the incentives faced by the agency and the equilibrium
outcomes, a more accountable agency is thus isomorphic to a more transparent lobbying
environment. With respect to welfare consequences, the matter is however different (see below).
Proposition 1. The efficiency region over which the consumer surplus standard dominates
increases the less transparent the lobbying process and the less accountable the agency is. The
efficiency region over which the welfare standard dominates is unchanged.
Proof. By definition of e and eV and by A1 and the monotonicity of S 1(d ) and S 2(d ), we have
Be2
Be
Be V
1
that Be
Ba b0; Ba b0; Ba Ba 0. 5
Proposition 1 illustrates that a less transparent lobbying process increases the range of
parameter values for which mergers that decrease welfare are pushed through under a welfare
standard (i.e. in between e 1 and e ). This accords with intuition; as a higher share of profit can be
used to effectively influence the agency, mergers which require marginally more influence over
the agency can now be pushed through. The same argument applies to accountability. Whenever
the agency is less accountable it takes less perks to push a merger through.
By contrast, the range of parameters for which the welfare standard dominates is unaffected,
since neither e nor eV are affected by the transparency of the lobbying process or accountability.
The reason for this is that an infra-marginal amount of additional resources would be required for
firms to push through the efficient outcomes near e . A marginal increase in the efficiency of
lobbying does not change the outcome. Also, in the upper limit of the inefficiency region (near
eV) only a marginal amount of resources is necessary to push the merger through at this point and
more efficient lobbying does not affect the outcome.
Hence, in terms of the size of the efficiency ranges, the consumer surplus standard becomes
relatively more attractive as transparency decreases and as the antitrust agency becomes less
accountable.
The next proposition examines the difference in the level of welfare under the two standards.
Proposition 2. In the efficiency region where the welfare standard dominates (e N e ), W 1
W 2 is
increasing in the transparency of the lobbying process and the accountability of the antitrust
agency. In the efficiency region where the consumer surplus standard dominates (e b e ), the
effect of a change in transparency and accountability on W 1
W 2 is ambiguous.
Proof. Consider the region e 2 V e b eV. From Lemmata 7 and 8 we have W1
W2
CS 1
d
d cb.
Since CS does not change with d, we have that W 1
W 2 increases in cb and decreases in d.

D.J. Neven, L.-H. Roller / Int. J. Ind. Organ. 23 (2005) 829848

843

Consider now the region e V e b e 2. In this region (see Lemmata 7 and 8), we have
W 1
W 2 = P m + P c + CS, which is independent of cb and d. Since Be 2 / Ba b 0 by Proposition
1 and cb N 1, the first part of the proposition follows. For the second part of the proposition
consider the region e 1 V e b e . By Proposition 1, Be 1 / Ba b 0, so that we first consider e 1 + Be 1 /
Ba
b e . In

b e1
d
 this region, from Lemmata 7 and 8 we have W1
W2 Pm Pc CS
1 d cb , which is decreasing in cb and increasing in d. Since Be 1 / Ba b 0, there is a new
region where the consumer surplus standard is dominated, i.e. W 1
W 2 b 0. 5
The Proposition is illustrated in Fig. 4 and offers two useful insights. The first insight relates
to the parameter range for which the consumer surplus standard yields an inefficient outcome,
while the welfare standard is first-best (e N e ). Even though the efficiency range is unaffected by
the lobbying process, Proposition 2 illustrates that the type of inefficiency that occurs in this
range is not: as transparency is reduced (or the accountability of the antitrust agency is reduced),
the range for which a type I error occurs shrinks, and the range of parameter for which wasteful
lobbying occurs increases (i.e. e 2 moves left, see Fig. 4). In other words, a less transparent
lobbying environment increases the resources available to influence the antitrust agency, while
less accountability reduces the need to provide perks. Both effects imply that more mergers are
pushed through. As a result the range of parameters for which a type I error occurs shrinks.
Given our assumption that the process of accountability is tough enough (cb N 1), a type I error
always entails a higher efficiency loss than lobbying in this parameter range (see Fig. 4 again).
The waste associated with lobbying is only a fraction of the loss entailed by prohibiting the
merger which is the entire opportunity cost of the merger in terms of welfare. Hence, a reduction
of type I errors will reduce the cost associated with the consumer surplus standard in this area.
Consequently, there are two reasons as to why the cost associated with the consumer surplus
standard falls when the environment for lobbying becomes more favorable (i.e. a less transparent
and less accountable lobbying process). First, a more favorable lobbying environment implies
that fewer resources are wasted (lower accountability reduces the transfer of the agency and
hence the waste; lower transparency reduces the share of the transfer which is wasted). Second, a
more favorable lobbying environment reduces the range of parameters for which a type I error
occurs and this reduces the cost of the consumer surplus standard because a type I error is more
costly than the waste associated with successful lobbying. This result which will be useful for
further reference is collected in Lemma 9.
w1-w2

e1

e2

b = 1

Fig. 4. Comparative static with respect to the political economy.

844

D.J. Neven, L.-H. Roller / Int. J. Ind. Organ. 23 (2005) 829848

Lemma 9. In the efficiency region where the welfare standard dominates, the net cost of the
consumer surplus standard increases in e2, holding e and eV constant.
Proof. Follows from Proposition 2.

The second insight offered by Proposition 2 relates to the region where the consumer surplus
standard dominates (e b e ). As discussed above, both types of inefficiencies occur simultaneously
here under the welfare standard: there is a type II error and there is wasteful lobbying.
Interestingly, a more favorable lobbying environment (less transparency and less accountability)
have opposing effects on the two types of efficiencies. While the range of mergers for which a
type II error occurs is increased, the waste from lobbying is reduced. Less transparency reduces
the share of the transfer which is wasted while lower accountability reduces the amount of the
transfer which is required in equilibrium and hence the waste. As a result the impact of the
political economy is ambiguous. To put it differently, a favorable environment for lobbying is
not desirable because it will increase the scope of undesirable deals that firms can manage to
push through. On the other hand, it is desirable because it reduces the costs from lobbying. Note
also that it is always preferable to enhance transparency rather than accountability. Consider
possible changes in transparency and accountability which have the same effect on incentives.
The former increases welfare by more than the latter (because transparency reduces waste
directly whereas accountability reduces waste only because it induces lower transfer).
We next characterize the trade-off with respect to market structure.
Proposition 3. The larger the size of the merger, the larger the efficiency region over which the
consumer surplus standard dominates, and the larger the efficiency region over which the
welfare standard dominates. Moreover, when the consumer surplus standard is dominated, the
efficiency regions for which Type I errors occur shrinks relative to the efficiency region for which
waste occurs.
Proof. We first consider the region where the consumer surplus standard dominates. Note that
A3 can be written as BS / BM b 0. The definition of e then implies that Be / BM N 0, i.e e moves to
the right (see also Fig. 5). If BS 1 / BM N 0 (i.e. when merger size has a strong effect on industry
profits), then Be 1 / BM b 0, which implies that the region is getting larger. If BS 1 / BM b 0, then
Be 1 / BM N 0, i.e e 1 moves to the right. By A3 and A4, we have BS 1* / BM N BS / BM. Since BS 1 /
Be N BS / Be, it follows that Be / BM N Be 1 / BM, i.e. the function that has e 1 as a fixed point shifts
w1-w2

e1

e~
e2

b = 1

Fig. 5. Comparative static with respect to size of the merger.

D.J. Neven, L.-H. Roller / Int. J. Ind. Organ. 23 (2005) 829848

845

by less and is steeper than the function which has e as a fixed point, so that e moves faster to the
right than e 1 as M increases. We next consider the region where the consumer surplus standard is
dominated. We first show that this area increases. By A3 and A4 we have BCS / BM b BS / BM.
Since BCS / Be b BS / Be it follows that Be / BM b BeV / BM, i.e. the function that has eV as a fixed
point shifts more and is flatter than the function which has e as a fixed point, so that e moves less
to the right than eV as M increases. We now show the last statement of the proposition. By A3
and A4, we have BS 2 / BM N BCS / BM. Since BS 2 / Be N BCS / Be, it follows that BeV / BM N Be 2 /
BM, i.e. eV moves faster to the right than e 2 as M increases. 5
Proposition 3 indicates that in the presence of larger mergers, both standards perform
relatively worse, to the extent that the range of parameters for which either standard yields
inefficient outcome increases (see Fig. 5). In this sense, the larger the merger, the less desirable
are the outcomes.
Consider those mergers which reduce welfare but are pushed through under the welfare
standard. Larger mergers tend to reduce welfare for any level of the efficiency parameter, so that
in principle firms have to provide more inducement in order to push them through (i.e. e moves
rightsee Fig. 5). However, larger mergers also enhance the profits that are available to provide
inducement to the agency. Under the assumptions of our model,16 the latter effect always
dominates the former so that the range of parameters for which inefficient mergers are pushed
through increases.
Proposition 3 also illustrates that increasing the size of a merger does not necessarily lead to
more regulatory clearance. When increasing the size of mergers has a strong effect on the
1
industry profits, more precisely when BS
BM N0, then e 1 moves left (see the proof of proposition 3),
regulatory clearance will never be jeopardized by larger mergers. However, when the effect of
1
larger mergers on industry profits is smaller, more precisely BS
BM b0, then e 1 moves right but by
less than e. In this case, there will be a range of efficiency parameters for which regulatory
clearance will be jeopardized as the merger becomes larger.
Consider the range of efficiency parameters for which the consumer surplus standard yields
an inefficient outcome. As mergers become larger, both welfare and consumer surplus are
reduced (so that both e and eV shift right). Under the assumptions of our model, the effect on
consumer surplus is greater than the effect on welfare so the range of parameter values for which
the consumer surplus standard yields an inefficient outcome increases (eV moves further to the
right than e ). However, Proposition 3 also shows (see the proof) that eV moves further to the right
than e 2, which implies that the relative size of the regions for which a type I error occurs under
the consumer surplus standard tends to shrink relative to the region for which there is wasteful
lobbying. The reason is as follows: when the size of a merger increases, consumer surplus falls
and more inducement has to be provided to the agency. However, industry profits also increase
so that more inducement is available. As before, under the assumptions of our model, the latter
effect dominates the former and the range of parameter for which successful lobbying arises
increases. As a consequence, the range of parameter for which a type I error occurs will shrink
relative to the range for which successfulbut wastefullobbying arises.
Having considered how the ranges of efficiency parameters for which either standard is
inefficient change with merger size, the next proposition analyses how merger size changes the
relative costs and benefits of the two standards.
16

It follows from the fact that welfare is less affected by efficiency than industry profits and by the fact that welfare is an
average of profit (which increase with merger size) and consumer surplus (which decreases with merger size).

846

D.J. Neven, L.-H. Roller / Int. J. Ind. Organ. 23 (2005) 829848

Proposition 4. When the consumer surplus standard dominates, larger mergers increase the net
1
benefit of the consumer surplus standard, whenever BS
BM N0, i.e. when increasing merger size has
a strong effect on industry profits.
Proof. Consider the region where the consumer surplus standard dominates. By A3 and Lemma
2, we know that F m + F c increases in M. Using Lemmata 7 and 8, it follows that B(W 1
W 2) /
1
*
BM b 0 in this region. When BS
BM N0, e 1 moves left and the net benefit of the consumer surplus
standard increases. 5
The intuition behind this proposition is straightforward (see Fig. 5). As discussed above,
when increasing the size of mergers has a strong effect on industry profits, it will never
jeopardize regulatory clearance. In other words, the range of efficiency parameters for which the
mergers are pushed through with a given merger size includes the set of efficiency parameters for
which the mergers are pushed through with any lower merger size. In addition, for any value of
the efficiency parameter, a larger merger size implies that the type II error is more costly and that
a stronger inducement (hence more waste) has to be provided to the agency to push the merger
through. In those circumstances, the cost of the welfare standard is thus unambiguously greater.
When increasing the size of mergers has a weaker effect on industry profits, the matter is less
clear. In those circumstances e 1 moves right, and some larger mergers (with low efficiency) are
blocked, which is first best.
Proposition 5. When the welfare standard dominates, the effect of larger mergers on the net cost
of the consumer surplus standard is ambiguous. However, the consumer surplus standard is
more attractive with larger mergers whenever the lobbying process is less transparent and the
accountability of the antitrust agency is low.
Proof. Note that Be / BM and BeV / BM are unaffected by a. Further note that BS 2 / BM
is increasing in a which implies that B 2e 2 / BMBa is negative. The result follows then from
Lemma 9. 5
Proposition 5 confirms that the effect of merger size on the cost of the consumer surplus
standard is ambiguous. This should not come as a surprise given that increasing the merger size
both increases and shifts rightward the range of parameters for which the consumer surplus
standard yields an inefficient outcome.
The second part of the proposition uncovers a complementarity between merger size and the
political economy of lobbying: the more favorable the lobbying environment (i.e. less
transparent and less accountable), the more the region over which type I errors occur shrinks
when merger size increases. Hence, not only does a more favorable lobbying environment
reduce the cost of the consumer surplus standard by itself (see above), but it also makes the
consumer surplus standard more attractive when merger size increases.
Collecting the above results, we find that a favorable lobbying environment and large mergers
reinforce each other in reducing the cost of the consumer surplus standard in regions where it
yields an inefficient outcome. Where the welfare standard yields inefficient outcomes, large
mergers make matters worse independently of the efficiency of lobbying (at least when larger
mergers have a strong effect on industry profits).
5. Conclusion
This paper illustrates that the design of a standard that an antitrust agency is held accountable
to needs to consider the political economy in which the antitrust agency operates. In particular,

D.J. Neven, L.-H. Roller / Int. J. Ind. Organ. 23 (2005) 829848

847

we address how institutional settingssuch as transparency and accountabilityinteract with

the design questions of choosing an appropriate standard.
We find that neither a welfare standard nor a consumer surplus standard dominates. The
consumer surplus and welfare standards give rise to different types of inefficiencies: relatively
inefficient mergerswhich decrease welfareare pushed through under a welfare standard,
while relatively efficient mergers (which would increase welfare) are prohibited under the
consumer surplus standard. While lobbying activity is undesirable under a welfare standard, it
raises welfare under a consumer surplus standard.
The process of lobbyingas characterized by transparency of the lobbying process and
accountability of the antitrust agencyis shown to be important in terms of the relative
performance of the two standards. Both transparency and accountability make lobbying less
effective. Under a welfare standard this has two effects. On the one hand, it will reduce the scope
of undesirable deals that firms can manage to push through. On the other hand, transparency and
accountability impose a constraint that firms can only circumvent at a cost. By contrast,
transparency or accountability do not affect the scope of deals for which a consumer surplus
standard is inefficient. It only affects the balance between wrong decisions and waste in
lobbying. Higher transparency and stronger accountability actually shift the balance towards
wrong decisions because it reduces firms effectiveness in lobbying. Since wrong decisions are
socially more costly than lobbying, transparency and accountability are actually not desirable
under a consumer surplus standard.
We also find that the size of the proposed merger has a different impact on the performance of
the two standards. Under the welfare standard, there are two effects. On the one hand, larger
mergers, which have greater willingness to lobby the agency, manage to push more numerous
deals through, despite the fact that larger deals are also less desirable in terms of welfare (and
hence require more lobbying to be pushed through). On the other hand, the deals that are pushed
through may on average be less damaging so that the overall effect is ambiguous. Still, when
industry profits are strongly affected by the size of the merger, the former effect will dominate
and the welfare standard will perform relatively worse.
The matter is different for a consumer surplus standard. The range of deals for which inefficient
outcomes arise increases with larger mergers. But larger mergers also tend to shift the balance away
from wrong decisions and in favor of wasteful lobbying (because firms have more resources). As a
result, the performance of the consumer surplus standard is not unambiguously worse with larger
mergers. In addition, we observe that larger mergers will shift the balance away from decision
errors more firmly when transparency is low. Hence, low transparency and larger mergers are
circumstances that reinforce each other in making a consumer surplus standard more attractive.
In terms of policy, our results suggest that it may be more appropriate for an agency to move
towards a welfare standard when the transparency and accountability of its procedures are high.
The average size of notified mergers has been unusually large in the most recent merger wave
(see European Economy, 1999). In such an environment, our analysis implies that maintaining a
consumer surplus standard may be appropriate. Finally, it has been argued that a large proportion
of mergers actually fail to generate significant efficiencies (see Roller et al., 1999). In this case, a
consumer surplus standard would seem to be more appropriate.
Acknowledgements
We like to thank Johan Lagerlof and Steve Martin for detailed comments on a previous
version of the paper. We are also grateful for comments from Martin Hellwig, Konrad Stahl,

848

D.J. Neven, L.-H. Roller / Int. J. Ind. Organ. 23 (2005) 829848

Roland Strausz and Xavier Vives, as well as seminar participants at the Graduate Institute of
International Studies (Geneva), Carlos III University of Madrid, University of Mannheim and
the 9th WZB conference in Industrial Organization.
References
Besanko, D., Spulber, D., 1993. Contested mergers and equilibrium antitrust policy. The Journal of Law, Economics and
Organisation 9 (1), 1 29.
Bernheim, D., Whinston, M., 1986. Common agency. Econometrica 54 (4), 923 942.
Besley, T., Coate, S., 1997. An economic model of representative democracy. Quaterly Journal of Economics 101, 1 31.
Besley, T., Coate, S., 2000. Lobbying and Welfare in a Representative Democracy, mimeo, London School of
Economics.
Bond, E., 1996. Competition Policy in Customs Unions: a Natural Experiment using State Level Antitrust Enforcement,
mimeo, Penn State.
European Economy, 1999. Merger and Acquisitions, Supplement A, Economic Trends, p. 2.
Gellhorn, E., Kovacic, W., 1994. Antitrust Law and Economics in a Nutshell. West Publishing, St. Paul.
Grossman, G., Helpman, E., 1994. Protection for sale. American Economic Review 84 (4), 833 850.
Harberber, A.C., 1971 (September). Three basic postulates for applied welfare economics: an interpretive essay. Journal
of Economics Literature 9 (3), 785 797.
Kirchsteiger, G., Prat, A., 2000. Inefficient equilibria in lobbying. The Journal of Public Economics.
Laffont, J.-J., Tirole, J., 1991. The politics of government decision making: a theory of regulatory capture. Quarterly
Journal of Economics, 1089 1127.
Lagerlof, Johan, Heidhues, Paul, 2002. On the Desirability of an Efficiency Defense in Merger Control, mimeo.
Neven, D., Nuttal, R., Seabright, P., 1994. Merger in Daylight: the Economics and Politics of Merger Control in the EU.
CEPR, London.
Posner, R.A., 1975 (August). The social cost of monopoly and regulation. Journal of Political Economy 83 (4), 807 827.
Rama, M., Tabellini, G., 1998. Lobbying by capital and labor over trade and labor market policies. European Economic
Review 42, 1295 1316.
Roller, L.-H., Stenneck, J., Verboven, F., 1999. Efficiency gains from mergers. European Economy 5, 2001.
Scherer, F.M., 1993. Monopoly and Competition Policy, vol. I. Edward Elgar.