Transcript Lect8MergerSimulation

Simulating Mergers
MSc Regulation and Competition
March 2005
Overview
•
•
•
•
This talk on ideas and methods
Instruction for the software
Practice with the software
Input into coursework
Merger simulation overview
•
•
•
•
•
•
Legal framework
Limitations of earlier approaches
What merger simulation is
Key assumptions: the economics
Implementation
Uses and limitations
Legal framework
“Balance of advantage” arguments admitted:
• US anti trust enforcement: application of rule
of reason. Merger guidelines on HHI
• EU Merger Regulation and task force
– “create or strengthen an dominant position”
• UK Enterprise Act 2002…“substantial
lessening of competition”
Standard approaches
• Define the market
• Work out market shares and entry
barriers
• Apply judgement on dominance and
likely effects
Difficulties
• Delineation of market not clear
(geographic or product)
• Differing degrees of substitution
• What would be the impact of merger
– pricing
– allowing for cost savings
Merger simulation
Used in arguing for or against
• A model of market
– typically calibrated by econometric analysis
• Making assumptions about behaviour
• Effects on prices and welfare estimated
• Alternative scenarios and assumptions
can be tested
Example cases
• Staples/Office Depot (FTC)
– FTC predict 7% increase
– 15% pass-through rate
• WorldCom/Sprint (FCC)
– Effects on long distance prices
• Volvo/Scandia (EU - example)
– Price rises in national markets
• Kimberley Clark (Company)
Behaviour assumptions
• Mainly Nash equilibria:
– Cournot (quantity setting)
– Bertrand (price setting)
Nash equilibria look at “unilateral effects”
• Also possibility of looking at collusion
effects
The Cournot model
• Other firms’ outputs are fixed
• Choose output to maximise profits
• Unique equilibrium:
– market share depends on relative marginal
cost
• Most relevant for:
– homogeneous products
– markets with important capacity decisions
Bertrand
• Mainly for differentiated products=>
Chamberlin model
• Own- and cross-elasticities relevant
• Simple formula:
(Pi - MCi)/Pi = -1/ Ei
• BUT Ei may not be constant
– choice of demand function matters (Crooke
et al)
Demand functions
Depends on how fast elasticity rises with price
• Extremes:
– constant elasticity
– linear
• Intermediate
– Almost Ideal Demand System (AIDS)
– Logit
– Probit
Qualifications
• No substitute for understanding the
market
• Apparent precision is spurious
• Miss out key features of reality e.g.
potential competition
Arguments for
• “Scientific” – assumptions explicit
“Merger Simulation is Science
• “Based on well-established economic theory
• E.g., econometric theory, theory of consumer demand, and
oligopoly theory
• Underlying assumptions are clearly laid out and can often be
tested
• Demand model can be tested
• Nash-Bertrand assumption can be tested by comparison to premerger margins
• Process and results can be replicated
• A third party following the same steps would reach the same
results
• Even though there are modeling choices along the way, these
choices are fully described “
• (extract from Leonard presentation)
Sources of software
•Canadian Competition Authority
– Cournot model
– http://csgb.ubc.ca/ccpp/simulation/
•Vanderbilt University
– Bertrand, linear demand
– http://mss.math.vanderbilt.edu/~pscrooke/MSS/linearmerger.html
– http://www.antitrust.org/simulation.html
– These give limited control over outputs
For this exercise we grew our own teaching
aids
Running the software
• Excel file with macros
• DiffProdStud1.xls or Coursework
handout
• cournot-102.xls
Further reading
• Numerous papers by Froeb and Werden
• Crooke Froeb and Tschantz on functional
forms
• Greg Leonard (Nera) In defense of merger
simulation
http://www.usdoj.gov/atr/public/workshops/docs/202615.pdf
• Mike Walker (Charles Rivers Associates)
Paper to Industrial Economics Study group
Nottingham, December 2005