Industrial Organization
Central Concern: Market Power
Existence, Magnitude, Sources, Implications.
Microeconomics of Imperfectly Competitive Market.
Partial Equilibrium Analysis.
No income e¤ect.
With no externalities:
Pareto E¢ ciency () Social Surplus maximization
Social surplus: Consumer Surplus + Producer Surplus
E¢ ciency )
marginal willingness to pay = price paid for marginal unit
sold = marginal production cost.
Perfect competition can secure this (under eventually
non-increasing returns to scale).
Socially optimal output:
p = M C:
Market power:
price markup above marginal cost.
*
Market power measured by Lerner index (L).
One …rm:
L=
p
MC
p
Many …rms
n
X
M Ci
]
pi
i=1
= (weighted) average of price-cost margin
L =
p
si [ i
= "expected" market power faced by consumers.
Note that if marginal cost for each …rm i is constant
(= ci), then
L =
=
n
X
i=1
n
X
p
si [ i
ci
pi
]
pq
ciqi
si [ i i
]
piqi
i=1
n
X
si[ i ]; i = prof it; Ri = revenue of …rm i;
Ri
i=1
= (weighted) average pro…t rate in the industry.
=
Why concerned about market power?
Social ine¢ ciency.
Underproduction of goods relative to social optimality.
Loss of social surplus: deadweight loss.
Also, production ine¢ ciency (equimarginal cost principle
violated)
) industry cost of production not minimal.
"Chicago argument" : The ine¢ ciency is static. May
disappear in a dynamic context.
Technological progress requires Schumpeterian creative
destruction of old …rms by new innovators who then yield
market power
High price-cost margin compensating them for their investment in R&D and innovation
Soon dispossessed by a fresh generation of leaders.
At each point of time, snap shot may show high market
power.
The identity of market power wielders shifts radically over
time.
Consumers reap bene…ts of technological change.
Industries that are technologically sedate on the other
hand, exhibit none of these and consumers may actually
be worse o¤ there.
Note: Monopoly with perfect price discrimination ) ef f iciency:
So, perhaps the objection to market power is based on
"fairness"?
Industrial organization literature: 50s - early 70Dominated by the Structure-Conduct-Performance (SCP) paradigm.
Structure = market concentration (also, product structure, technology, costs, vertical integration structure).
Market concentration refers to how concentrated market
shares are.
*
Market concentration = n1 in symmetric homogenous good market where n = number of …rms.
*
More generally, concentration measured by Her…ndahl index (H)
H =
n
X
s2i
i=1
n
X
qi
si =
;Q =
qi;
Q
i=1
qi = output of …rm i; si = market share of …rm i:
Note that H re‡ects not only number of …rms but the
way market shares are distributed between …rms.
Conduct: behavior (pricing, R&D, investment, competitiveness, collusion, advertising...)
Performance: e¢ ciency of market (MARKET POWER,
pro…ts, innovation, product variety, distributional consequence)
SCP arose from antitrust and justice department’s concern over market power in post-war US.
*
Does market power exist? Where? How much?.
*
What structural features of markets explains market power? Empirical enquiry.
*
Academic hypotheses of SCP paradigm:
S
! C:
greater concentration leads to higher collusion,
more anti-competitive behavior
(entry barriers, predation etc.).
C
! P :
greater collusion, more anti-competitive behavior
leads to greater market power.
S
! P : more concentration leads to higher prices.
)
S
! P (by-pass Conduct):
* Structure is primitive, performance is the explained variable.
* One of the primary forms of this hypotheses:
Higher market concentration (H)
) Higher market power (L).
* Almost mindlessly, ran cross section regressions (across
industries) for over 20 years. Adding all kinds of variables.
* Measured L by pro…t rates.
* Net result: INCONCLUSIVE, AMBIGUOUS.
* DATA PROBLEMS. (pro…t rates often accounting profits).
* Endogeneity problem: S may be determined by C and
P:May lead to negative relationship between S and P:
Example: High pro…t rates may lead to high entry of new
…rms and, therefore, low concentration.
* Yielded no insight into conduct of …rms (C treated as
black box).
* Chicago School (Stigler etc.) - brought out these kind
of arguments in simple theoretical models.
* Existence of very high market power doubtful.
* High concentration - even monopoly - may not lead to
high market power.
(For example, Coase conjecture in durable goods markets).
* One problem with early theory in I.O.: lack of tools to
handle oligopolies.
* Early 1970s: non-cooperative game theory enters economic model building.
Explosion of research in IO theory from mid-70s.
* Directly analyzes Conduct instead of treating it as
a black box and treats S; C; P as being simultaneously
determined.
* Explicit analysis of strategic behavior in imperfectly
competitive markets (Oligopolies),
Game theoretic foundations of observed behavior and institutions in markets;
Incomplete information.
This course: Study of modern Industrial Organization
theory.
Monopoly.
Homogenous good market and simple pricing.
Cost function: C (q ):
Market demand function: D(p); Inverse demand: P (q ).
*
Market power:
1
L=
(pm)
where (p) is the price elasticity of demand at price p
and pm is the monopoly price:
Monopoly power in the sense of deviation from marginal
cost pricing is higher in markets with less elastic demand.
*
Deadweight loss of monopoly:
DW L =
Z qc
qm
[P (q )
C 0(q )]dq
where q c is the socially optimal output de…ned by
P (q ) = C 0(q c):
DWL also called welfare loss (e¢ ciency loss) due to monopoly
and the Harberger triangle.
*
DWL is actually an upper bound on e¢ ciency gain
possible through state intervention in a monopoly market
- the actual gains might be much smaller.
*
Unlike market power, DWL does not necessarily decrease with price elasticity of demand.
*
Empirical estimates of DWL: very small. Harberger
(1954): 0.1% of GDP. Controversial.
*
Policy measures to correct ine¢ ciency caused by
monopolies:
a) Subsidy (not tax): Output subsidy s per unit.
Optimal subsidy:
s=
q cP 0(q c)
Problem 1: Financing subsidy through taxes might cause
other distortions and welfare loss in the economy.
Problem 2: Setting subsidy optimally requires that regulator have information on cost function - often a private
information of the …rm. Firm has incentive to secure
higher subsidy by mis-reporting cost.
Problem 3: Regulatory capture.
b) Price regulation (price ceiling).
Problems 2 and 3 exist here too.
* Additional problem with price regulation: in natural
monopoly (increasing returns to scale), marginal cost pricing implies …rm earns loss. Need subsidy.
Way out: set price regulation such that price = average
cost.
(Called Rate of return regulation: "fair" rate of return on
capital).
Firm breaks even.
Not fully e¢ cient but reduces DWL.
Problem: Creates horrible incentives for cost reduction,
R&D, quality improvement etc.
Way out: Lagged price adjustment so that …rms can earn
positive rent for some time.
*
Current regulation literature: Asymmetric information between …rm and regulator is the most serious
problem in regulation of monopoly.
Simple example: Constant returns to scale; Constant unit
cost = c:
Regulator does not know c = cL or cH .
0 < cL < cH
Firm knows its unit cost.
Regulator asks …rm to report unit cost and sets a price
ceiling p(c) according to the reported unit cost.
Assume: monopoly price corresponding to unit cost cL >
cH :
If p(cL) 6= p(cH );…rm will report the value of unit cost
that leads to higher price ceiling.
Cannot secure both truthful reporting and e¢ ciency with
simple price regulation.
For example, e¢ ciency requires p(cH ) = cH ; p(cL) =
cL, but then monopolist will lie if actual cost is cL:
Need subsidy.
For example, a lump sum subsidy A to be paid if reported
unit cost is cL:
Set p(cH ) = cH ; p(cL) = cL:
A = (cH
cL)D(cH ):
In general, monopolist must be paid informational rent for
truthful reporting - lower his cost, higher the rent paid.
But such rents increase the incentive for regulatory capture and need to be …nanced externally.
* Broad economic view:
Except in the case of natural monopoly, best way to deal
with monopoly power
understand why there is no greater competition in the
market
- for example, remove barriers to entry and incentives for
predation of small new entrants by judicious review of the
conduct of monopoly …rm;
remove barriers to international trade so as to introduce
foreign competition etc.
* Other sources of welfare loss due to monopoly power:
1. Rent seeking behavior.
(Welfare loss may include DWL + monopoly pro…t; may
even exceed it).
2. Cost Distortion due to lack of yardstick competition.