TMP 38E050 Advanced Topics Economics of Competition and
Download
Report
Transcript TMP 38E050 Advanced Topics Economics of Competition and
4. Mergers and Acquisitions
• Mergers are usually categorized by closeness of markets that
firms operate in
• Horizontal merger
– Merging firms operate in same relevant market, firms are
directly competing
– Market shares in relevant markets change as result of
merger
• Vertical merger
– Merging firms operate at different stages of a production
or distribution chain
– Firms products belong to same relevant market do not
compete horizontally
– At least one firm can potentially be using the other firms'
products as inputs in its production
HKKK TMP 38E050
1
© Markku Stenborg 2005
• Conglomerate merger
4. Mergers and Acquisitions
– Mergers not belonging to those above
– Product extension
• Products of the firms not competing but firms use
close marketing channels or production processes
– Market extension
• Products are competing but relevant geographic
markets are separate
– Pure conglomerate mergers (none of those mentioned)
Effects of Merger
• Suppose duopoly which behaves competitively
• Assume firms have identical cost functions and constant
returns to scale prevail
– MC1 = AC1, there are no fixed costs
HKKK TMP 38E050
2
© Markku Stenborg 2005
• Profit maximization under perfect competition forces firms to
4. Mergers and Acquisitions
price at marginal cost: pc = MC1
• Case 1: Merger to monopoly and costs stay at original level
– Profit maximization rule (MC = MR) implies output Qm1
and price level pm1 so that deadweight loss DL1 takes
place
– DL = (Qc- Qm1)(pm1- pc)/2
– This is strict decrease welfare
– Also, merger means an income transfer from customers
to owners of newco.
– In this case, there would be reasons to block merger
– Merger needs to be blocked for its deadweight loss
creating effect, not because it means income
redistribution
HKKK TMP 38E050
3
© Markku Stenborg 2005
• Case 2: Merger involves synergies
4. Mergers and Acquisitions
– Assume cost savings occur through decrease in marginal
costs
– MC1 decreases to MC2
– Monopoly profit maximization implies price level pm2
which is lower than that without cost savings pm1
– Deadweight loss occurs, but it is smaller than that
without cost savings
– DL = (Qc- Qm2)(pm2- pc)/2
– Cost savings due to the decrease in MC
• Amount is (pc-MC2)Qm2
– Efficiency is increased due to cost savings and decreased
due to market power - deadweight loss
HKKK TMP 38E050
4
© Markku Stenborg 2005
• Case 2 illustrates typical situation in antitrust
4. Mergers and Acquisitions
– Many types of decisions and conduct by firms may be
harmful for welfare while increasing it in other ways
– From antitrust authority point of view, we face a trade-off
– To determine whether certain conduct or decisions to
merge are harmful on welfare, the authority should
compare gains and losses to welfare
• In US, this seems to be the case, efficiency defence
• In EU, efficiency gains are more of reason to block
merger, efficiency offense
• Difference partly due to legislation?
– Market dominance in EU
– Significant lessening of competition in US
HKKK TMP 38E050
5
© Markku Stenborg 2005
Incentives for horizontal mergers
4. Mergers and Acquisitions
• Salant, Switzer & Reynolds (QJE, 1982)
Merger in Cournot market
• Assume an industry structure characterized by:
– n identical firms (cost functions are identical)
– Cournot or capacity competition
– Constant returns to scale: C(qi) = C(q) = cq, c > 0
– Linear demand is assumed linear: p(Q) = a - bQ, a,b > 0
– No possibilities for entry
• Profit function of any firm is then
(
a
b
(
q
Q
)
)
q
c
q
, Q
q
i
i
i i
i
i
j
ij
• Firm i's Cournot-Nash equilibrium profit is
( a c) 2
cn
i
b(n 1) 2
HKKK TMP 38E050
6
© Markku Stenborg 2005
• Merger between any two firms: there is one firm less in the
4. Mergers and Acquisitions
industry than before
– n firm industry changes into n-1 firm industry
• Suppose m of n firms decide to merge (1 < m < n)
• m firms have incentive to merge if being part of merged
entity gives more profit than staying unmerged, that is, if
1
(a c) 2
1 m b(n m 1) 2
( a c) 2
b(n 1) 2
• that is if
2
2
(
n
1
)
(
m
1
)
(
n
m
1
)
2
2
2
n
m
1
m
n
m
0
• Define LHS = A
HKKK TMP 38E050
7
© Markku Stenborg 2005
Case 1: m=1
4. Mergers and Acquisitions
A2n1n2.
• Notice that only if n=2, merger is profitable
• Monopoly created
• Hence, only if in duopoly both firms merge we have the
merger being in all firms' interest
Case 2: m=2
A4n1n.
2
• Notice that only if n=3, merger is profitable
– This again means we have a monopoly being created
– Only if in triopoly all firms merge, merger is in all firms'
interest
HKKK TMP 38E050
8
© Markku Stenborg 2005
Case 3: m=5
4. Mergers and Acquisitions
2
A
1
0
n
1
9
n
.
•
•
•
If n=6, merger is profitable: monopoly created
But now even with n=7 merging is profitable
– Creation of a duopoly through merger is profitable
With n=8, merger is again unprofitable
More generally
• Notice that
A
/
n2
m
2
n
•
•
which is < 0
Thus, A is decreasing in n, number of firms in industry
More there are firms before merger, other things equal, more
difficult it is for merger to be profitable for merging firms
HKKK TMP 38E050
9
© Markku Stenborg 2005
• Notice also that
4. Mergers and Acquisitions
A
/
m
1
2
n
2
m
•
•
•
•
•
which is > 0
Thus, A is increasing in m, number of firms that decide to
merge
– More there are firms that take part in merger, other
things equal, easier it is for merger to be profitable for
merging firms
Irrespective of value of m or n, only if 80 % of firms in
industry takes part in merger, merger is profitable
Merger to monopoly is always in firms' interest
Typical Cournot model where nothing but number of firms
changes price level increases after merger
This follows from quantity competition since quantities are
strategic substitutes
HKKK TMP 38E050
10
© Markku Stenborg 2005
• Decrease in output by one firm is (partly) matched by an
Mergers
increase in4.
output
by rivaland
firm Acquisitions
• In Cournot model, once some firms merge, they decrease
their total output, as they act as single firm
• Firms not party to merger increase their output
• Under many parameter values, firms which mostly benefit
from merger are non-merging firms
– Business stealing effect
• Model says that mergers are not usually profitable
• Then we should not usually observe mergers, assuming that
firms are rationally behaving agents!
• Not a good description of the real world where mergers are
taking place in increasing numbers
• Model misses some essential aspects of the phenomenon
– Mergers occur endogenously, not exogenously
– Cost savings needs be incorporated
HKKK TMP 38E050
11
© Markku Stenborg 2005
• In Salant et al. one reason for mergers being unprofitable
4. Mergers
and Acquisitions
due to strategic
substitutes
– Decrease in production of some firms is matched by an
increase in production by the competitors
• One way to overcome this effect is to assume U-shaped
costs (strictly convex costs)
– Rivals have less incentive for expansion of production as
costs are increased
– Mergers are more probable than in Salant et al
Mergers in Bertrand Market
• In models above firms' strategies were quantities
• Deneckere & Davidson (RJE 1985) merger incentives under
price competition
• Prices are strategic complements
– Price increase by some firms is matched by price increase
of rival firms
– Reaction functions are upward sloping
HKKK TMP 38E050
12
© Markku Stenborg 2005
• In differentiated products Bertrand model firms engage in
4. Mergers and Acquisitions
price competition
• Price increase followed by merger is matched by price
increase of rivals
– Reaction of outsiders reinforces initial price increase that
results from merger
– Then merger of any size is beneficial for merging firms
– No business stealing effect
• Notice that this model predicts industries would usually
evolve into monopoly!
• This, luckily, is not really what happens in real world
• There seems to be forces which prevent monopolization
– These forces are not easily modelled and simple models
do not descibe real world phenomena in satisfying way
HKKK TMP 38E050
13
© Markku Stenborg 2005
• Notice that busines stealing effect is very much true in real
4. Mergers and Acquisitions
world
– Often, firms benefiting from mergers are non-merging
firms
– Thus, usually Cournot competition best describes real
world phenomena, this holds with merger theory as well
• In preceding models acquiring and target firms were not
differentiated
• Firms were ”black boxes”, mere MC-functions
– Only effect is reduction in number of (symmetric) firms
• In real world acquisitions, there usually is buying and selling
side in transaction
• Transaction creates a larger entity
• Seller sets price based on many factors
– Asset value of the firm
– Expected evolution of industry (expected profits)
HKKK TMP 38E050
14
© Markku Stenborg 2005
• Kamien & Zang (QJE 1990): In quantity competition, does
4. Mergers and Acquisitions
monopolization of industries take place when acquisition
process is endogenous?
• In quantity game, total industry profit increases with a
decreasing number of firms
• Any firm increases its profit as number of firms in industry
diminishes
– This follows from the nature of Cournot competition
• Seller knows that it would gain in profits if it would sell later
rather than sooner
• As a consequence of this, sellers want to ask more than
buyers want to pay
– Monopoly profit is maximum buyer can pay
• In Cournot model following can be showed: complete
monopolization of an industry is possible only if originally
there were only a few firms in industry
HKKK TMP 38E050
15
© Markku Stenborg 2005
Welfare effects of mergers
• Merger without cost savings reduces welfare; if merger
involves cost savings, we have trade-off
• Farrell & Shapiro (AER 1990) is most thorough model on
welfare implications of horizontal mergers
– Quantity competition and general demand structures
– Cost-savings are allowed
– Mergers without synergies increase price and hurt
consumers
– Cost saving is proportional to post-merger output
– Deadweight loss is proportional to output reduction
– If cost saving outweigh the deadweight loss, net welfare
effect of merger is positive
HKKK TMP 38E050
16
© Markku Stenborg 2005
Merger simulation
4. Mergers and Acquisitions
• Market definition is hard with differentiated goods and can
be misleading
– Market definition is {0,1} decision, good is ”in” or ”out”
– In reality goods belong to [0,1], they pose varying
degree of competitive pressure to each other
• Increase in market power is interesting, not market
definition
• Pure structural analysis of competitive effects can be
misleading
• Simulation uses economic models grounded in theory to
predict effect of mergers on prices in relevant markets
• Simulation allows direct measuring of changes in market
power
– Easier than measuring of market power
HKKK TMP 38E050
17
© Markku Stenborg 2005
• Simulation allows to evaluate likelihood of synergies
4. Mergers and Acquisitions
offsetting price increases
• Simulation requires estimation of demands
– Minimum: own and cross-price elasticities
• Merger simulation: the big picture
– Demand estimation
• Create demand models
• Get data and estimate demands
• Calibrate demand model(s)
– constant elasticity
– linear
– logit
– AIDS, etc
to produce pre-merger prices, quantities, and demand
elasticities
HKKK TMP 38E050
18
© Markku Stenborg 2005
• Calibrate model: set parameters so that it exactly
4. Mergers and Acquisitions
predicts pre-merger equilibrium
– Plugging pre-merger prices into model must yield
pre-merger shares
• Predict post-merger marginal costs
– Try to evaluate synergies
– Use demand model & post-merger costs to compute
post-merger prices
– Idea: if post-merger prices are well above pre-merger
level, transaction increases market power
• Measuring market power is hard
– Market power = L
– L = (p-c)/p [0, 1/e] so that eL [0, 1]
– has basically same info content as L
– Quality of market power measure depends on accuracy of
estimates of marginal costs and demand elasticity
HKKK TMP 38E050
19
© Markku Stenborg 2005
– Data and estimation problems lead to biased measure of
4. Mergers and Acquisitions
market power
• Why would measuring changes in market power be easier?
– Estimated price change reacts less to estimated MC or
demand, as we use same ”instrument” to measure pre
and post-merger market power
• Limitation of simulation: price increase predictions are
sensitive to functional form used for demand
– Functional form of demand determines magnitude of
price increases from merger
– Linear and logit demand yield smallest price increases
– Constant elasticity and AIDS demand typically yield price
increases that are at least several times larger than those
with linear or logit demand
HKKK TMP 38E050
20
© Markku Stenborg 2005
• Use calibrated models in manner that makes them
4. Mergers and Acquisitions
insensitive to functional form of demand
– Compute compensating marginal cost reductions (CMCR)
that exactly offsets price-increasing effects
– CMCRs do not depend on functional form of demand as
pre and post merger equilibrium prices and quantities are
precisely same
– If merger synergies appear likely to reduce merging
firms’ cost as much as CMCRs, merger is unlikely to harm
consumers
– If merger synergies clearly fall well short, significant
price increases are likely
• Visit http://antitrust.org/simulation.html
– Fool around with Linear Bertrand Merger
– If you have access to Mathematica, take a look at
SimMerger to get feeling of what simulation is about
HKKK TMP 38E050
21
© Markku Stenborg 2005