Transcript Chapter 1: The Market - University of Minnesota

Lecture #11: Introduction to the New Empirical
Industrial Organization (NEIO)
modified from:
http://spirit.tau.ac.il/public/gandal/lecture11i.pdf
 What is the old empirical IO?
 The old empirical IO refers to studies that tried to draw
inferences about the relationship between the structure of an
industry (in particular, its concentration level) and its
profitability.
 Problems with those studies are that
 it is hard to measure profitability,
 industry structure may be endogenous, and
 there is little connection between empirical work and theory.
 The NEIO is in many ways a reaction to that empirical tradition.
NEIO works “in harmony” with (game) theory.
The Structure-Conduct-Performance (SCP)
Paradigm: The Old Empirical IO
 How does industry concentration affect
price-cost margins or other similar
measures?
 The “plain vanilla” SCP regression
regresses profitability on concentration.
 The Cournot model of non-cooperative
oligopolistic competition in homogeneous
product industries relates market structure
to performance.
 It can be shown that shared weighted
average markup in an industry is a
function of its Herfindahl index and the
elasticity of demand.
 p  mci 
i  p  si 
where


HHI    si2 
 i

HHI
 
 This implies:
ln(PCM) = 0 + 1 ln(HHI) + 2 ln() + 
(1)
where PCM = the share weighted average firm markup.
 Using cross section data, (1) could be estimated and a test for
Cournot would be 0 = 0, 1 = 1, and 2 = -1.
 In practice, regressions of the following form were performed:
ln(PCM) = 0 + 1 C4 + 
(2)
where C4 was the share of the four largest firms in the
industry.
 What are the problems with these regressions?
 Data: Price cost margins are not available, hence accounting
returns on assets (as well as census data on manufacturer
margins) were often used.
 Data: Additional RHS variables such as the elasticity of
demand are hard to measure for many industries.
 Data: Market definitions were problematic. (When using a
single industry more thought can be put into the issue.)
 Simultaneity: Both margins and concentration are
endogenous.
Demsetz (J. Law and Econ., 1973) noted that some firms have
a cost advantage, leading to a large share and high profits.
Hence, there could be a correlation between C4 and profits.
The New Empirical Industrial Organization
(NEIO) Paradigm: The New Empirical IO

The NEIO studies
 build on the econometric progress made by SCP paradigm,
 use economic theory,
 describe techniques for estimating the degree of
competitiveness in an industry,
 use bare bones prices and quantities, and do not rely on cost
or profit data,
 typically assume that the firms are behaving as if they are
Bertrand competition,
 use comparative statics of equilibria to draw inferences about
profits and costs, and
 focus mainly on a single industry in order to deal better with
product heterogeneity, institutional details, etc.
Bresnahan (Economic Letters, 1982):
The Oligopoly Solution Concept is Identified
This paper and others (e.g., A. Nevo, “Identification of the Oligopoly Solution Concept in a
Differentiated Products Industry,” Economics Letters, 1998, 391-395) show that the oligopoly
solution concept can be identified econometrically.
 Demand:
MC:
Q = 0 + 1 P + 2 Y
MC = 0 + 1 Q + 2 W
+
+
(1)
(2)
where Y and W are exogenous.
 FOC for a PC firm:
FOC for a monopoly:
P/Q)
MC = P
( MR is: P)
MC = P + Q/1 ( MR is: P + Q *
FOC for the oligopolistic firm: MC = (1 - ) P +  (P + Q/1)
where
 = 0 if the industry is competitive

Demand:
(1)
MC:
(2)
FOC:
Q
=
0 + 1 P + 2 Y + 
MC = 0 + 1 Q + 2 W + 
MC = (1 - ) P +  (P + Q/1)
 Substituting FOC into (2), the supply relationship is:
(1 - ) P +  (P + Q/1) = 0 + 1 Q + 2 W + 
 P =  (- Q/1 ) + 1 Q + 0 + 2 W + 
(3)
 Since both (1) and (3) have one endogenous variable and since
there is one excluded exogenous variable from each equation,
both equations are identified. But, are we estimating P = MC
(competitive) or MR = MC (monopoly)?
Rewrite (3) as:
P = (1 -  /1) Q + 0 + 2 W + 
We can obtain a coefficient for (1 -  /1) since (3) is identified
and we can get 1 by estimating (1).
 In the figure, E1 could
either be an equilibrium for
a monopolist with marginal
cost MCM, or for a perfectly
competitive industry with
cost MCC.
MCC
 Increase Y to shift the
demand curve out to D2
and both the monopolistic
and competitive equilibria
move to E2.
E2
E1
MCM
 Unless we know marginal
costs, we cannot
distinguish between
competitive or monopoly
(nor anything in between).
MR1 MR2
D1
D2
The solution to the identification problem
 MC:
MC = 0 + 1 Q + 2 W + 
(2)
 Let demand be:
Q = 0 + 1 P + 2 Y + 3 P Z + 4
Z+ 
where Z is another exogenous variable. The key is that Z enters
interactively with P, so that changes in P and Z both rotate and
vertically shift demand. (Note: Z might be the price of a substitute
good, which makes the interaction term intuitive.)
 Given the new demand, the marginal revenue for a monopolist, (P
+ Q * P/Q), is no longer P + Q/1. It is now: P + Q / (1 + 3
Z).
 Thus, the oligopolistic firm’s FOC becomes
MC = (1 - ) P +  [P + Q / (1 + 3 Z)]
 Substituting FOC into (2), the supply relationship becomes:
 Demand:
Q
=
0 + 1 P + 2 Y + 3 P Z + 4 Z + 
Supply Relation:
`
P
= -  [Q / (1 + 3 Z)] + 1 Q + 0 + 2 W +

 -  Q*
+ 1 Q + 0 + 2 W +

 Since the demand is still identified, we can estimate 1 and
 3.
Thus, we can identify both  and 1 in the supply relation.
 Note 1: This result can be generalized beyond linear functions.
Note 2: There are other assumptions that can generate
identification.
e.g.
Marginal cost that does not vary with quantity (1 =
 Graphically, an
exogenous change in
the price of the
substitute rotates the
demand curve around
E1.
 If there is perfect
competition, this will
have no effect on the
equilibrium price and
it will stay at the point
associated with E1.
 But, if there is
monopoly power, the
equilibrium will
change to E3.
MCC
E1
E2
MR1
MCM
MR2
D1
D2
Why do we care about the structural model?
 We want to estimate parameters or effects not directly
observed in the data (e.g., returns to scale, elasticity of
demand).
 We want to perform welfare analysis (e.g., measure
welfare gains due to entry or welfare losses due to
market power).
 We want to simulate changes in the equilibrium (e.g.,
impact of mergers).
 We want to compare relative predictive performance of
competing theories.