CourseBrick Spring2012 Zhou

Firms and Markets
(COR1-GB.1303.00/0P)

Spring 2012

Professor Jidong Zhou

Table of Contents

SYLLABUS AND OUTLINE
Syllabus 1
Correspondence between Topics and Textbook 6
Outline and Calendar 8

GROUP PRESENTATIONS
Outline of Topics and Directions for Presenting 10

PROBLEM SETS
Problem Set 1 (Math Review) 20
Problem Set 2 (Demand, Supply and Market Equilibrium) 21
Problem Set 3 (Pricing) 23
Problem Set 4 (Games) 25
Problem Set 5 (Information) 28

PRACTICE
Practice Questions (pre Mid-Term) 31
Practice Questions (post Mid-Term) 36

SUPPLEMENTARY LECTURE NOTES
Introduction 39
Calculus Review 42
Demand and Supply: Buyers, Seller and Markets 51
Utility 56
Demand 63
Logarithms 71
Economic Profit and Costs 72
Perfect Competition 78
Pricing 83
Advanced Pricing 88
Market Power 96
Auctions 100
Game Theory 105
Price Competition 115
Competition and Cooperation 124
Commitment, Entry and Exit 133
Imperfect Information 137
Externalities 145
Networks and Standards 148
Firms and Markets
Office: KMC 7-72
Telephone: (212) 998-0589
Email: jzhou@stern.nyu.edu

Syllabus 1

Syllabus

Course Description

The goal of this course is to give you some insight into how markets work. The first part of
the course starts with the study of decision making by consumers and firms and concludes
with a fundamental result in economics: a set of conditions under which markets function
efficiently. In the second part of the course, we focus on situations when, for one reason or
another, markets dont work efficiently. We will emphasize the importance of strategic
behavior, as modeled by game theory.

Microeconomics (as the topic of this course is frequently referred to) is an important
component of an MBA program. First, microeconomics focuses on specific dimensions of
optimal firm decision making, such as pricing and entry and exit. Second, the formal
economics perspective on business plays an important role in other areas of MBA study,
such as finance or marketing. Finally, by studying public policy towards market failures,
microeconomics highlights important factors conditioning firm strategy.

Some of the key concepts we will introduce include economic incentives, marginal analysis,
opportunity cost (which costs matter), market efficiency (what does it mean for a market to
work), strategic behavior (how to predict and respond to your rivals decisions), and
asymmetric information (what happens when others know something you do not). Our
experience with students in prior years is that much of this is intuitive. But much is not, and
our hope is that the combination of theoretical structure and practical examples will be
useful in the years to come. It will not make you a success on its own, but it might give you
an edge a few times when it matters.

Prerequisites

You are expected to be comfortable with basic algebra and calculus, including systems of
equations, logarithms and NPV calculations, and derivatives.

Firms and Markets
Office: KMC 7-72
Telephone: (212) 998-0589

Syllabus 2

Course Materials

Lecture notes. They review the theory relevant to most classes. In a few pages, they outline
and explain the conceptual issues for the day, define terms, give examples, and (where it
makes sense) work through numerical problems. They are intended to complement the
lectures rather than substitute for them.

Textbook. There is no required textbook for this course. However, if you want to have a
reference text, I recommend Michael Bayes Managerial Economics and Business Strategy
(McGraw-Hill, 6
th
or 5
th
edition), which is available in the bookstore or online. There is also
a study guide to accompany the text. Some students indicate that they find this book
helpful.

Slides. I will post the slides on Blackboard after each class, but keep in mind that there is
much more to the class than what you see in the slides. I recommend that you take notes
during the class as a supplement to the slides.

Additional materials. I will post additional materials on Blackboard, such as some useful
materials for group presentation, and other newspaper articles or research papers.

Deliverables and Grades

The various deliverables in the course are designed to develop different skills:

Class participation. It is important to integrate what you learn and be able to express it
effectively. Moreover, there is a great deal of collective insight and experience in the class
and we all benefit from sharing it. But the quality of your contributions is more important
than the quantity.

Individual problem sets. Problem sets emphasize quantitative applications of the principles and
tools developed in class. They are due at the start of class. They will not be graded, but will
be marked with a check (and possibly a plus or minus). Most of the problems are
quantitative; some require a qualitative answer and for these there may be no definitive right
or wrong, it is understanding the issues that is key. You should also note that the problem
sets are the best preparation for the exams. Dates and deadlines for all assignments can be
found in the detailed course outline.

Group presentation. Groups of about 4 or 5 students will be asked to make a presentation to
the class on a topic selected from a list of relevant themes. The goal is to apply economic
principles to real-world situations and to hone your communication skills. The content is
more qualitative than assignments.

Syllabus 3

Mid-term exam. Generally three or four problems, similar to the practice mid-term exams
that I will distribute in due course.

Final exam. Generally four or five problems, similar to the practice exam that I will
distribute toward the end of the class.

Your grade for the course will be based on your contributions to all of these deliverables, weighted as
follows:

Group presentation 30%
Mid-term exam 30% (only if it helps you!)
Final exam 40% (or 70%)

Your performance in terms of class participation and in problem sets will serve as a
tiebreaker if you are on the border between two grades.

The mid-term only counts if it increases your grade. This means that if you do better in the
final than the mid-term, then the mid-term will not count. The reason for doing this is that
the course will move fast and I am sensitive to the fact that some students may need some
time to familiarize themselves with what economics is about. Ultimately I care about what
you learn by the end of the course - the grading scheme is intended to be consistent with
that concern.

Final grades will follow the Schools guideline for core courses: no more than 35% of the class will
receive A or A. This guideline was instituted in response to student concerns that different sections
of a course might be graded by different standards.

Exams and re-grading

Requests for a make-up exam must be made in writing (email) at the earliest instance. Since the mid-
term is redemptive, there will be no make-up for the mid-term. You are responsible for checking the
midterm exam dates and avoid any conflict with other commitments. During an in-class exam, you
are not allowed to consult class notes, books, or any other material. However, you may consult one
page of notes (a standard-size sheet of paper written on one side). Questions about grading must be
made in writing and no more than a week after the exams are returned.

Honor Code

The Stern community believes that honesty and integrity are necessary for rewarding academic and
professional experiences. These qualities form the basis for the strong trust among members of the
academic community (students, faculty, and administrators) that is essential for excellence in
education. The Honor Code requires that each student act with integrity in all academic activities and
endeavor to hold his or her peers to the same standard.

Syllabus 4

In this course, you may discuss assignments with anyone, but any written work submitted for a grade
should be your own. On exams, you may bring in and consult one piece of paper with anything on it
you like, but your answers should be entirely your own work.

Students with disabilities

If you have a qualified disability and will require academic accommodation during this
course, please contact the Moses Center for Students with Disabilities (CSD, 998-4980) and
provide me with a letter from them verifying your registration and outlining the
accommodations they recommend. If you will need to take an exam at the CSD, you must
submit a completed Exam Accommodations Form to them at least one week prior to the
scheduled exam time to be guaranteed accommodation.

Getting help

I would like each of you to learn and gain as much as you can from this course. If you are
stuck, or have any difficulty with the material, dont hesitate to ask for assistance. Please send
me an email (jzhou@stern.nyu.edu), and I try my best to respond promptly. My regular
office hours are 4-6pm every Saturday, but if this does not work for you, feel free to email
me and set up an alternate time.

You can also get help from the teaching assistant, Esther Judelson. You can email her at
(erj229@stern.nyu.edu) with questions, or set up a mutually convenient time to meet.

All announcements regarding the course will be made on Blackboard. Besides administrative
issues, I may post clarifications on the class material (arising out of our discussion in class or
following from a fellow students questions). You are responsible for checking Blackboard
for announcements on a regular basis (i.e. at least a couple of times a week).

Syllabus 5

A preliminary list of topics to be covered:

Demand, supply and market equilibrium. Demand and supply. Market forces. Market
equilibrium. Shifts in demand and supply.

Demand. Utility function. Indifference curve. Budget constraint. Demand function.
Consumer surplus. Demand elasticity. Demand estimation. Risk aversion.

Economic cost analysis. Opportunity costs. Sunk costs. Marginal costs. Economic costs
and cash flows.

Perfect competition and market equilibrium. Short run equilibrium. Long run
equilibrium. Comparative statics. Welfare analysis: consumer surplus, producer surplus, and
economic efficiency.

Basic monopoly pricing. Profit maximization. Marginal revenue. Marginal cost. Elasticity
rule.

Advanced pricing. Price discrimination. Market segmentation. Two-part tariff. Quantity
discount. Versioning and bundling. Dynamic pricing.

Market power and policy. Economies of scale and economies of scope. Market power.
Public policy towards mergers.

Game theory. Strategies and payoffs. Simultaneous-move games and normal-form Games.
Sequential-move games and extensive-form games. Dominant and dominated strategies. Best
responses. Nash equilibrium. Backward induction. The prisoners dilemma, the coordination
game and other important games.

Price competition. Bertrand competition. How to avoid the Bertrand trap: cost leader,
cooperative pricing, limiting capacity, product differentiation, price matching.

Competition and cooperation. Cooperation in business. Cooperative pricing. Repeated
games. Trigger strategies. Cartels. Tacit collusion. Factors that make cooperation easier.

Commitment. Credibility. The value of a credible commitment. First mover advantages.
Preemption. Product proliferation. Entry and Exit.

Asymmetric information. Hidden actions. Moral hazard and the agency problem.
Incentive design. Hidden types. Adverse selection and the lemons problem. Screening and
signaling. Auctions.

Externalities and network effects. Positive and negative externalities. Coase theorem.
Network effects. Expectations and critical mass. Strategic compatibility decisions. Two-sided
markets

6

Where in Baye are the Topics that We Cover in Class?
[Baye 6
th
and 5
th
Editions]

Firms and Markets does not follow any of the existing textbooks closely (and that is why we
have developed and updates notes to supplement the lectures). However, many have found
it useful to have additional supplementary materials and sources for examples and practice
questions, though many have found that they can manage the course very well without much
reference to textbooks.

In addition to differing in coverage, Bayes presentation is in a somewhat different order and
style from ours.

This is a guide to the textbook. It is designed to help you find the bits in Baye that are
relevant for the topics covered in class. When you read Baye, use the lectures as a guide to
what is important.

Loosely, corresponding to our topics you will find the relevant material from Baye (5
th
and
6
th
Edition) as follows:

Supply and Demand:
36-64 (5
th
Edition)
36-65 (6
th
Edition)

Consumer Demand:
36-44, 74-95, 117-135 (5
th
Edition)
36-46, 74-95, 117-135 (6
th
Edition)

Economic Costs:
45-52, 177-191 (5
th
Edition)
46-52, 177-190 (6
th
Edition)

Competitive Markets:
267-280 (5
th
Edition)
266-279 (6
th
Edition)

Basic Pricing:
236-256 (very good background), 280-296, 397-399, 509-518 (5
th
Edition)
235-255, 278-294, 397-399, 510-520 (6
th
Edition)

Advanced Pricing:
404-417 (5
th
Edition)
404-417 (6
th
Edition)

7

Market Power:
280-284, 509-518 (5
th
Edition)
278-283, 510-520 (6
th
Edition)

Game Theory:
353-365, 378-386 (5
th
Edition)
353-365, 378-386 (6
th
Edition)

Pricing Games:
315-338 (but particularly 336-338) (5
th
Edition)
315-338 (but particularly 336-338) (6
th
Edition)

Cooperation:
330-332, 339-340, 365-377 (5
th
Edition)
330-332, 339-340, 365-377 (6
th
Edition)

Commitment:
378-386, 484-486, 491-494 (5
th
Edition)
378-386, 485-488, 492-495 (6
th
Edition)

Asymmetric Information:
449-455, 220-228 (5
th
Edition)
450-456, 219-227 (6
th
Edition)

Auctions:
455-466 (5
th
Edition)
456-466 (6
th
Edition)

Externalities:
518-526 (5
th
Edition)
520-528 (6
th
Edition)

8

Firms and Markets (COR1-GB. 1303.00/0P): Outline and Calendar

Date Topic
Problem Set
Presentation
Feb 11
Introduction.
Demand, supply and market equilibrium:
demand curve, supply curve, equilibrium
Demand I: utility function, indifference curve,
budget constraint, demand function, consumer
surplus

Feb 18
Demand II: demand elasticity, demand
estimation, risk aversion
Economic cost analysis: opportunity cost, sunk
cost, marginal cost, supply curve
PS#1 (Math Review) due at
the start of class

Feb 25
Perfect competition and market equilibrium:
short run equilibrium, long run equilibrium,
comparative statics, welfare analysis
Basic pricing: profit maximization, marginal
revenue and marginal cost, elasticity rule

Mar 3
Advanced pricing: price discrimination, market
segmentation, two-part tariff, quantity discount,
versioning, bundling, dynamic pricing
PS#2 (Demand, Supply,
and Market Equilibrium)
due at the start of class

Mar 10
Practice midterm exam
Market power and policy: economies of scale
and economies of scope, market power, public
policy towards mergers
PS#3 (Pricing) due at the
start of class

Mar 17

No Class Spring Break
Firms and Markets
Office: KMC 7-72
Telephone: (212) 998-0589

9

Mar 24
Midterm Exam (in class): same format as the
practice midterm (3 or 4 problems, largely
quantitative). You may consult one page of notes
(a standard-size piece of paper, single sided, with
anything on it you like) and a calculator
Game theory I: strategies and payoffs,
simultaneous-move games, normal-form games,
dominant and dominated strategies, best
responses, Nash equilibrium

Mar 31
Game theory II: the prisoners dilemma and
other important games, mixed strategies,
sequential-move games, extensive-form games,
backward induction
Price competition: Bertrand competition, how
to avoid the Bertrand trap: cost leader,
cooperative pricing, product differentiation,
limiting capacity, price matching

Apr 7
Competition and cooperation: cooperation in
business, cooperative pricing, repeated games,
trigger strategies, cartels, tacit collusion

PS#4 (Games) due at the
start of class
Presentation: Broadway
theaters; Google; GE; AT&T
Apr 14
Commitment: credibility, the value of a credible
commitment, first mover advantages,
preemption, product proliferation, entry and exit

Presentation: Real estate
brokerage; Mutual fund;
Virgin; Predatory pricing
Apr 21
Asymmetric Information: hidden actions, moral
hazard and the agency problem, incentive design,
hidden types, adverse selection and the lemon
problem, screening and signaling, auctions

Apr 28
Practice final exam
Externalities and network effects: positive and
negative externalities, Coase theorem, network
effects, expectation and critical mass,
compatibility decisions, two-sided markets
PS#5 (Information) due at
the start of class
Presentation: eBay; Health
reform; Microfinance
May 5
Final Exam (in class): same format as the
practice midterm (4 or 5 problems, largely
quantitative). You may consult one page of notes
(a standard-size piece of paper, single sided, with
anything on it you like) and a calculator

Group Presentation 10

Group Presentations

Each study group will give a presentation to the class from one of the topics listed below. You
should view your presentation as an opportunity to exhibit some creativity, to hone your
research and presentation skills, and to attack a real business economics issue.

An ideal presentation should last approximately 10-12 minutes. A time limit of 12 minutes will
be enforced ruthlessly! This means that you will not be allowed to keep talking after 12
minutes, leaving 3-5 minutes of questions. You should bear the following points in mind:

Informativeness: How much did we learn from the presentation?
Analysis: Did we gain novel insights into the topic? Were lessons from the course
applied effectively? Was evidence used effectively to support the argument?
Style: Was the presentation clear and compelling? Were the slides effective?

Above all, keep your classmates interested.

If you use Powerpoint, you should bring your presentation to class on a diskette, CD, or USB
pendrive. In addition, bring a hard copy of your slides to class and send me an email
with your presentation (before the day of the presentation) so that we have a back up in
case any problem occurs.

You can learn a great deal from observing and thinking critically about your classmates
presentations. What worked? What did not? Why? Were the points well argued? How could
they be argued better?

You will be asked to evaluate each others presentations, using the form that appears below.
This will encourage you to think critically about the presentations (what did you take away
from the presentation?). It will also generate more feedback for the presenters and contribute
to the grading process. The grading will be as follows:

Average of classmates assessment 50%
My assessment 50%

Firms and Markets


Presentation Topics and Sources

Choose your topic from the list below (by the end of February 19
th
). Topics will be allocated
on a first-come, first-served basis. To request a topic, send an email to the teaching assistant,
including your first, second, and third choices.

Each topic comes with a date (see the course outline) and a series of questions. Feel free to
modify some of the questions if you think that the modifications would lead to a more
interesting presentation. The material posted on Blackboard is intended to get you started.

Further potential sources of information and data include:
Bobst Library
NYU Virtual Business Library
Proquest (newspaper and magazine articles)
Lexis-Nexis (news articles, financial filings, legal decisions, etc.)
Investext (analyst reports)
Global Market Information Database
Hoovers Online (company profiles)
ABI/Inform
Market Research Monitor
DOJ Antitrust
FTC
FCC
EU Competition

(1) Price discrimination in Broadway theaters
Describe the various ways in which Broadway theatres segment their markets. Would you
encourage most Broadway theatres to offer more seat quality and pricing divisions than they
currently do? How would you solve the problem of scalpers? (Is it a problem?)

(2) Google and antitrust
The EC and the U.S. Federal Trade Commission are each currently investigating Google for
possible antitrust violations. Why is this happening? What are the grounds for the finding of
an antitrust violation in this case? What is Google's likely defense?

(3) GE and Honeywell
What was the strategic logic for GE to purchase Honeywell? Do you find it compelling? What
was the European Commissions logic in blocking the merger? Do you think it made the right
decision? What should Jack Welch have done?

(4) AT&T and T-Mobile
AT&T and T-Mobile are two of the largest providers of wireless communication services in
the United States. In March 2011 AT&T proposed to acquire T-Mobile, but in September
2011 the Department of Justice sued to block this takeover. What were the key concerns of
antitrust and regulatory agencies of the U.S. government about this proposed acquisition?
What were the key points that AT&T used to defend the acquisition? How did the other two


major operators Verizon and Sprint respond? In what ways could the acquisition lead to a
reduction or enhancement of national economic efficiency?

(5) Residential Real Estate Brokerage
The standard commission for brokerage services for house sales is 5-6 percent in the United
States, and it appears to be resilient in the face of changing demand as well as new technologies
and business models. What is the industrys structure? What is the nature of competition in
the industry? What shortcomings of economic efficiency appear to exist, if any, and why? Is
the relatively steady commission rate an indication of economic inefficiency? What do you
think will happen in the next five to ten years?

(6) Product differentiation in the mutual fund industry
Give an idea about the number of funds, the dispersion of fees charged to investors, and the
financial performance of different funds in the mutual fund industry. What is special about
these figures? Do these findings still hold when attention is restricted to apparently more
homogenous funds, such as the S&P 500 index funds? What are the possible explanations for
these findings? Do you think entry into the industry should be restricted?

(7) Virgin and British Airways fuel surcharge
In the summer of 2007, British Airways and Virgin were found to be colluding on fuel
surcharges. British Airways were fined $546m while, Virgin were not fined at all as part of a
relatively recent leniency program for whistleblowers. Discuss the case and more broadly assess
the success, impact and design of such leniency programs.

(8) Predatory Pricing: Spirit vs. Northwest
What is predatory pricing, why is it a concern for antitrust authorities, and how can it be
distinguished from competitive behavior? What is distinct about the airline industry? Illustrate
the theoretical arguments by describing and analyzing the antitrust battle betweenSpirit and
Northwest.

(9) eBay
What are the reasons for eBays success? How did eBay address the challenges of online
trading when traders are anonymous? Why did amazon.com and yahoo.com fail to achieve the
same measure of success in online auctions? Do specialized auction sites and direct online
selling threaten e-Bays dominance?

(10) Healthcare reform
There is widespread agreement that the healthcare system in the U.S. is excessively costly and
that the problems of cost and of excessive utilization are getting worse. What are the causes of
these problems? What are some of the solutions that have been suggested?

(11) Microfinance meets the market
Give a brief description of credit markets in poor and developing countries, highlighting the
role of moral hazard and adverse selection. Describe the evolution of the microfinance
industry, highlighting its success and failures and explaining how or why it is better equipped
than traditional moneylenders in resolving the usual problems in credit markets. Discuss the
future of microfinance and, in particular, whether for-profit models will play a successful role


in this industry? You may focus entirely on microfinance in poor and developing countries or,
if you wish, you may want to focus on the lessons from the microfinance experience that are of
relevance to the recent and novel social lending business models that have appeared or have
the potential to appear in the US.


Presentation Tips

This guide is designed to remind you of some basic skills to enhance your presentation to the
class. It is not concerned about preparation of the content of your presentation, which
obviously should be well researched, introduced succinctly, organized logically, with clear
transitions and conclusions.

You must be audible: When you practice, have a group member listen to you in a large
room and have that person alert you if you cannot be heard easily.

Make eye contact, or the semblance of it: When you rehearse your presentation, practice
scanning the room, even if you are looking just over the heads of the audience, or at peoples
eyebrows. Take in the whole room.

The norm at Stern tends to be that presenters use PowerPoint slides as their notes. Thats OK
as long as you dont turn your back on the audience. To deal with this potential problem,
some people use note cards to keep them facing front, on track, and as a back up in case of
technology failure.

Do not read your presentation. Reading is the best way to lose your audience. Speak
naturally so that you dont sound as if you have memorized your talk.

Many of us speak too fast. If you are one of them, build into your presentation pauses and
reminders to speak more slowly. Some people find it helpful to include at selected points in
their notes big red marks saying BREATHE! or PAUSE or SCAN AUDIENCE or
SLOW DOWN.

If English is not your first language and you are worried people will have difficulty
understanding you, or if you speak with a strong American regionalism, speak more
slowly and deliberately distinctly at the beginning of your presentation to allow the
audience to get used to your speech patterns.

Time yourself. Cut appropriately. Include the crucial information. Save the interesting but
non-crucial material for the question and answer period if you have extra time. Remember that
less is more: if you add too many details, your audience often loses track of your main points.

Plant yourself: Many people have trouble standing still when they give a talk. It helps some
people to think of their feet rooted in the ground, as if they could not move. It gives them
strength and keeps them literally more grounded. For some people, the choice of shoes is
important as well.

Since you are presenting in a group format, try to present a united front when you make your
presentation. For example, using we instead of I makes it sound as if you worked
together. Showing that you are listening when others talk is good, too.

Practice, preferably in front of other people who can give you honest and useful feedback.


Finally, try to enjoy yourself.

[These comments were prepared by Stephanie Nickerson, an NYU teaching consultant who has worked
extensively with Stern faculty and students.]

A few further and related points arising from presentations in past years that might be helpful,
though no doubt to a great extent these comments represent our own peccadilloes and so
ought to be treated with more than the usual caveat lector.

What is the presentation about?
This might sound trivial, but we have sat through numerous presentations (and countless
research seminars) where it was such hard work trying to figure out what the point that the
speaker was trying to make that no one in the audience bothered. Details are important,
arguments can be subtle, but at least the question that you are trying to address or the point of
the presentation should be crystal clear.

What should the audience take away?
People (or maybe we are just speaking for ourselves) have limited concentration spans and
abilities to retain information. If there is just one thing that you want them to take away from
the presentation, what is it? Make sure you have communicated it. This need not be done as
formally as a takeaway slide, but at the same time you as a presenter should be clear about
what it is that you want your audience to come away with. (This might be a specific
recommendation, or set of recommendations, or some other key pointsit depends on x and
on y).

Shaping your presentation
Allied to the two points above, once you are clear about what the presentation is about and
what you want your audience to go away thinking, then this should help to have a clear
structure for the presentation. This does not have to involve a formal roadmap for the
presentation slide though it couldbut you should be clear about the structure of the talk.

What to put on the slides
Many presentations we have seen have suffered a tendency to put a lot of information on the
slides. This has a number of drawbacks for presentation:
Often it results in the presenter reading the slide and keeps them from engaging with
the audience;
It also leads to the audience reading the slide and not paying attention or listening to
the speaker.
If there are a great number of facts of which you think the audience should be aware, you
might consider using handouts (though this is certainly not required for a good grade for
presentations in this course).

Given these drawbacks, it is perhaps surprising that friends and students who have worked in
consulting firms and elsewhere tell us that densely-packed slides with a great deal of text, tables
and graphs are far from uncommon. At first blush, this might appear something of a puzzle.
However, in many cases slides serve two very different purposes (perhaps unfortunately). First


and obviously, as presentation aids. Moreover, frequently slides also serve in effect as reports
on the projects as a whole (the client often gets no written report but a big slide deck as the
sole deliverable).

As another way of addressing this issue, you might think of slides (as viewed by your audience)
as valuable real estate. You dont want to put too little on a slide (youre wasting valuable
space); but you also dont want to put too much on a slide (youll overload your audience; they
wont absorb everything; and youve wasted time, effort, and space).

Excessive use of technique
Dont try to impress us with the excessive use of economics tools (e.g., abstract supply and
demand charts, or graphs of marginal somethings) unless they directly help you tell your story. Again,
your slides are valuable real estate that shouldnt be wasted!

Pitfalls from preparation
Having spent a great deal of time and effort researching your topic, there are two common
pitfalls:
Overestimating the audience: You may have spent hours trawling through the Internet
researching the topic, but few of the audience are likely to have done the same. Try to
have some empathy for them.
The sunk cost fallacy in presentation: After learning a great deal of facts and
interesting detail, theres a great temptation to try to communicate all of it. Almost
always, though, less is more, and trying to list every fact can get in the way of
communicating the main points that you want the audience to take away. The desire to
show that you have researched a topic and know a great deal about it may lead you to
present a lot of facts. Try instead to show that you have thought about it by using the
facts selectively and appropriately! In any case, your knowledge or (lack of it) is likely
to be made clear in the way you deal with questions and discussion.

Specific Issues When Presenting an Economic Analysis
Economics is about evaluating the effect of a trade-off. For example, in considering a
price reduction a monopolist has to trade-off lost revenue on units that would be sold
at the original price versus the extra revenue from any extra units sold due to the rpice
decrease. Be very clear about what you think the key economic trade-offs are in what
you are presenting. To force yourself to be clear, have a slide called Key Economic
Trade-off if you can.

Dont spend too much time on the details of the industry. Dont be afraid to simplify
things.

Relate the facts to the underlying economic trade-off

Use the facts together with an appropriate framework to generate a recommendation
on how to resolve the trade-off.

If there are caveats in interpreting you results, better to be upfront about them.


Summarizing data or analysis in a picture or diagram tends to be more effective than
using a whole bunch of dot points

Written by the Stern Economics Department.
Copyright 2012 NYU Stern School of Business.

18

(note: 1 means extremely bad ----- 7 means extremely good; 4 means average)

Group:______________________________________

Topic:______________________________________
1. Presentation Format

i. Good time management

ii. Clarity of presentation

iii. Logical structure of presentation

1---2---3---4---5---6---7+

1---2---3---4---5---6---7+

1---2---3---4---5---6---7+
2. Analysis

i. Research: found appropriate background information

ii Used data and information selectively and appropriately

iii. Used appropriate course concepts in the analysis

iv. Effective synthesis of concepts and data/information

1---2---3---4---5---6---7+

1---2---3---4---5---6---7+

1---2---3---4---5---6---7+

1---2---3---4---5---6---7+

3. Recommendations/Conclusions

i. Sound recommendations or conclusions

ii. Supported findings with hard evidence

iii. Dealt well with questions

1---2---3---4---5---6---7+

1---2---3---4---5---6---7+

1---2---3---4---5---6---7+
4. Overall Presentation

1---2---3---4---5---6---7+

Firms and Markets Presentation Evaluation

19

Group:_____________________ Topic:_____________________

In one or two sentences what did you understand as the bottom line in
the presentation?

Comments:

Problem Sets 20

Problem Set 1 (Math Review)

(1) Linear equations
Solve the following system of linear equations:
x = 10 2 y
y = x + 4

(2) Quadratic equations
a) Knowing that p>0, determine the solution to the equation: (1+p)
2
= 36
b) Solve the equation:

2
1
25
y

(3) Logarithms
Solve the following equations (log without subscript is natural logarithm to the base e):
a) log
10
x = 4.
b) log x = 4
c) log (x) log(5) = log (0.75)
d) log (2y) = 3

(4) Net Present Value
a) Suppose that the annual interest rate is 10%. What is the value of a perpetuity
that pays $30 every year from the beginning of next year?
b) Suppose that the annual interest rate is 5%. What is the value of an annuity
that pays $100 every year from the beginning of next year?

(5) Derivatives
Compute the derivatives of the following functions:
a) 2 y
3
+ 4 y
b) (x + 3) (2 x)

(6) Derivatives of derivatives
Compute the second derivatives of the functions in the previous exercise.

(7) Optimization
Find out the maximum of this function: (100-q)q

2012 Stern School of Business.

Problem Sets 21

Problem Set 2 (Demand, Supply, and Market Equilibrium)

(1) Demand and Supply basics
For each of the following, use a supply and demand diagram to deduce the impact of the event
on the stated market in terms of both price and quantity.
(a) Event: Cold freeze in Florida. Market: Orange juice.
(b) Event: Digital cameras. Market: Film
(c) Event: The FDA announces that aspartame may cause cancer. Market: Saccharin.
[Comment: aspartame and saccharin are low-calorie sweeteners.]
(d) Event: An import tariff is imposed on steel. Market: aluminum.

(2) Internet-enabled mobile phones

Prices and quantities demanded for an Internet-enabled mobile phone have been observed to
be as follows:

P Q
$100 600
$105 590
$110 575
$115 550
$120 510

(a) Calculate the approximate elasticity of demand when price is $105.
(b) Is the demand elasticity constant at all prices?
(c) If the monthly subscription fee for Internet access using the mobile phone falls
from $10 to $2, what would you expect to happen to the quantity of mobile phones
demanded at any given price?

(3) Wireless demand.
AT&T Wireless estimates that demand for its service has an elasticity of 1.5.
In 2001, 18mm customers generated revenues of approximately $ 12.5 billion at an effective
price of about 10 cents a minute. If elasticity is constant, what would revenues be at a price of
8 cents? 9 cents? 11 cents?

(4) Jareds restaurant

After working for years as a chef, your cousin Jared is thinking about opening his own
restaurant and has asked you to help with his financial analysis.

Basic expenses. To open a restaurant, he will have to quit his current job (which pays
$46k a year) and invest his savings of $200k (currently earning 6% annually) in
equipment. He estimates he will have to spend $4k during the first year to maintain the

Problem Sets 22

equipment and preserve its market value of $200k. He would not pay rent: his father
has gaven him title to a small building, whose current tenant is a shopkeeper. The
tenant estimates that she nets $3000 a month from sales, over and above the $2500 a
month she pays in rent.
Other expenses. He anticipates spending $50k a year for food, $40k for extra help, and
$14k for utilities and other supplies during the first year of operations.

Jareds question to you: How much revenue would he have to generate to make his first year
more profitable than staying in his current job?

5) Impact of a price ceiling in the instant oatmeal market

The demand and supply curves for instant oatmeal are as follows:

Q
d
=10-0.5 P
d
where Q
d
is the quantity of instant oatmeal packets (in million units) demanded
when the price consumers pay is P
d
.

Q
s
= -2 +P
s
, when P
s
is greater than or equal to 2
0 , when P
s
< 2
where Q
s
is the quantity of instant oatmeal (in million units) supplied when the price producers
receive is P
s
.

Suppose the Government imposes a price ceiling of $6 in the market for oatmeal.

a) What are the equilibrium price and quantity in the market without a price ceiling?
b) What is the size of the shortage in the market with the price ceiling? What is the
producer surplus?
c) What is the consumer surplus, assuming the good is purchased by consumers with the
highest willingness to pay? What is the net economic benefit? What is the deadweight
loss?


Problem Sets 23

Problem Set 3 (Pricing)

(1) MelCos Xamoff

The global pharmaceuticals giant, MelCo, has had great success with Xamoff, and over-the-
counter medicine that reduces exam-related anxiety. A patent currently protects Xamoff from
competition, although rumors persist that similar products are in development. Two years ago,
MelCo sold 25 million units for a price of $10 for a package of ten. Last year it raised the price
to $11, and sales fell to 22 million units.
a) Estimate the elasticity of demand for this product.
b) MelCo is considering a further price increase to $12. Estimate the impact on the
quantity sold. Do you expect revenue to increase or decrease?
c) A financial analyst estimates the cost of production at $2 a package. What price
maximizes MelCos profit?
d) When the patent expires in 2010, what do you expect to happen to demand? What
about the elasticity?

(2) EasyTax

EasyTax markets both a deluxe and a standard version of its software. The deluxe version
contains additional features that are likely to appeal to sophisticated users. The marginal costs
of producing and distributing the two versions are virtually equal, and equal to zero. The
market is equally divided into two types of users, sophisticated and unsophisticated (and
assume one user of each type). The maximum willingness to pay for both types of users for the
standard and deluxe versions are given in the following table:
Standard Deluxe
Unsophisticated 20 20
Sophisticated 35 100

a) Assuming that the company producing EasyTax can distinguish between sophisticated
and unsophisticated users (because, for instance, the sophisticated users are registered
accountants), what are the optimal prices to be charged to each category of users?
What is the companys profit?
b) Assume now that the company cannot distinguish between the two types of user. What
are the optimal prices for the two versions? What is the total profit?

Problem Sets 24

(3) Sals satellite company

Sals satellite company broadcasts TV to subscribers in LA and NY. Demand functions are
Q
NY
=50-(1/3)P
NY

Q
LA
=80-(2/3)P
LA
where Q is in thousands of subscriptions per year and P is the subscription price per year.
The cost of providing Q units of service is given by
TC=1000+30Q,
where Q= Q
NY
+ Q
LA
.

a) What are the profit-maximizing prices and quantities for the NY and LA markets?
b) As a consequence of a new satellite that the Pentagon developed, subscribers in LA are
now able to get the NY broadcast and vice versa so Sal can charge only a single price.
What is the profit-maximizing single price that he should charge?
c) In which situation is Sal better off? In terms of consumers surplus which situation do
people in LA prefer and which do people in NY prefer? Why?
d) In the situation with a single price, what are the MRs in each market, at the profit
maximizing price? Are they equal? Now consider the situation with different prices for
NY and LA. What is the MR in each markets at the profit maximizing prices?

(4) CD pure bundling

Suppose a music industry producer just finished producing two compilation CDs of music
from the 70s. One disk is pop music, the other disk is disco music. Suppose also that he
estimates the market to be evenly segmented into two types of customers. He estimates
valuations for each type of customer as follows.
Disk 1 Disk 2
Type A 10 2
Type B 9 3

The marginal cost of the disks is zero.
(a) If the producer wants to sell the disks individually, what prices should he set?
(b) Can the producer do better with bundling? Explain.
(c) Would it be better to mix the content of the disks and obtain two pop/disco disks of
music from the 70s?

Suppose now that the estimated evaluations are instead:
Disk 1 Disk 2
Type A 10 3
Type B 9 2

(d) Does pure bundling make sense in this case? Explain the difference between the two
demand profiles.


Problem Sets 25

Problems 4 (Games)

(1) Simultaneous move game
Use the following game to answer the questions below:

L C R
125 250 100
T 100 300 200
0 500 400
M 250 500 750
-100 300 350
B 0 400 -100

(a) Find each players dominant strategy, if it exists
(b) Find the Nash equilibrium

(2) Intel and AMD

The following normal form game depicts the pricing rivalry between Intel and AMD for
their newest top-of-the-line chip. Each company considers five possible price levels: $449,
$399, $349, $299 and $249. Monthly profits in million dollars are:

449 399 349 299 249
50 70 55 45 30
449 200 180 150 120 95
35 45 50 40 25
399 230 210 160 130 110
15 25 30 35 20
349 210 190 180 140 120
5 15 20 25 15
299 180 170 150 120 100
-5 10 15 20 10
249 120 110 105 95 90

(a) Find each firms best response to each possible strategy of the other.
(b) What is (are) the Nash equilibrium (equilibria) of the game?
(c) What combination of strategies generates the highest total profit? Why might the two
firms not be able to come to an agreement on this outcome?
(d) What aspects do you think are left out in this game? (This is an open question; please
limit your answer to a paragraph or two.)

AMD
Player 2
Player 1
Intel

Problem Sets 26

(3) Lowest price guaranteed

Two firms compete in prices. Each firm can set a high price or a low price. Profits as a
function of prices are given by the payoff matrix is:
Firm 2
Low High

Low 70,70 120,0
Firm 1
High 0,120 100,100

(a) Are there dominant strategies? What is the Nash Equilibrium of the game (if any)?
Suppose now that each firm offers its customers a lowest price guarantee: if firm 2 offers a
price lower than firm 1, then firm 1s customers are entitled to buy from firm 1 at the same
price, even if firm 1 had initially set a high price; and similarly for firm 2s customers if firm 1
offers a price lower than firm 2. (Lowest price guarantees are common in many markets,
including sporting goods, books, house wares, cellular phones, electronics, luggage and travel
accessories, toys, tires, eyewear, prescription drugs, etc.)

(b) How do lowest price guarantees change the game?
(c) What is the equilibrium of the new game? What lessons can we learn?

(4) A Bargaining Game

Two players, PI and PII, have to divide 23 M&M's: 12 regular ones (R's) and 11 peanut-coated
ones (P's). PI likes both types equally but PII likes only the P's (she will have no use for any
R's she gets). Thus, u
1
(r,p) = r + p and u
2
(r,p) = p. Moreover, both players' preferences are
known to both and the bargaining will be restricted to the allocations of M&M's, e.g., no
monetary side payments or other modifications involved.
(a) PI Divides and PII Chooses: PI divides the M&M's into two piles in any way he
chooses. Then PII selects one of the piles, leaving the other one to PI. What will be
the outcome of this game?
(b) PII Divides and PI Chooses: The same as in Problem 1, but with the order reversed.
What will be the outcome?

(5) The Centipede Game

There are two players Harry and Sally. They each start with 1 dollar in front of them. They
alternate in saying either Stop or Continue, starting with Harry. When a player says
Continue, 1 dollar is taken from his / her pile, and 2 dollars are put in the opponents pile.
As soon as any player says Stop, play is terminated and each player receives the money
currently in her pile. Alternatively, play stops if both players reach 100 dollars.
Can you predict the equilibrium outcome of this game?

To help you think about this problem, the game tree is given below. This is a famous problem
in game theory, and because of what the tree looks like, is called the Centipede game. As we

Problem Sets 27

100
100
have done in class, try to think of what your opponent would do, and work backwards. Think
of yourself as Harry. Ask yourself at each step what Sally would do. What does that mean for
what you should do?


Problem Sets 28

Problem Set 5 (Information)

(1) The Giant Corporation

You are a small firm in a market which is dominated by Giant Corp. which has 75% of the
market. Your current challenge: to decide whether to expand your capacity. If you dont, you
expect to earn $2m, as you did last year. Giant will earn $30m. If you do expand, your profits
depend on whether Giant responds aggressively by cutting its price or passively by maintaining
price at its current level. The estimated possible payoffs if you expand are:
If Giant responds aggressively, you will lose $2m and the giant earns $10m.
If Giant responds passively, your profits increase to $4m and the giant earns $20m.
Use a game tree to study the strategic interactions between you and Giant. Should you expand
capacity?

(2) 1982 16 K classic pen

Mary, a friend of mine at a West Coast business school wants to buy a present for her fathers
birthday. In particular, she knows that hes been wanting to get hold of a particular pen for his
collection. She know that hes particularly keen to get hold of the 1982 16 K classic. Even
though the list price in collectors manuals is only $250, Mary is willing to pay $500 for the pen.

My friend has been unable to find the pen at the local dealers and decided to try a new website,
www.whatiwanttopay.com. At this site people post descriptions of the items that they would
like to buy and how much they are willing to pay for them. Mary is aware that some sellers
might try to fool her by sending the 12 K Classic and it would be hard for her to work out
which shed received ahead of the birthday. Thus even if she had the opportunity to return it,
she would value the 12 K at only $100. Genuine sellers would meet Marys offer so long as it
was at least as high as the list price, whereas as fraudulent sellers would send her the 12 K
version if she posted any price greater than $100.

From speaking to friends, she estimates that out of all potential sellers 25% are genuine sellers
and the remainder fraudulent. She determines therefore to post a price of
$200(=0.75*100+0.25*500).

(a) She told you about this, and you suggest that shed be better off not posting. Why?
(b) Suppose that she assessed that 80% of potential sellers are genuine. Should she post to
this site? At what price?
(c) Suppose that the proportion of potential sellers who are genuine is x. What is the
smallest value of x at which Mary should post an offer at the site? At what price?

Problem Sets 29

(3) Pricing as a Signal

For a HQ Item (50%) For a LQ Item (50%)

Buy Not
HP 1, 1 0, 0
LP -1, 3 0, 0

This is a game between a seller (S) and a buyer (B). S sells an item which may be of high quality
(HQ) or low quality (LQ). Knowing the quality of the item, S can ask for a high price (HP) or a
low price (LP). B, hearing the price but not knowing the quality, has to decide to buy or not. The
resulting payoffs to S and B, are given above. Overall, there is a 50% probability that the item to
be sold is HQ, and a 50% probability that it is LQ.
(a) Suppose S charges a high price if its item is of high quality but otherwise a low price.
How should B react? Given Bs reaction, will S stick to its pricing policy? Can price
signal quality in this example? Why?
(b) Now let us modify the above game slightly. Suppose B can discover the items quality
with a probability 70% before she makes her purchase decision (or 70% of the buyers
in the market know the real quality of the item). Can price signal quality now (when the
buyer does not know the quality)?

(4) Natives and tourists

Consider a game between a consumer and firm. The consumer has two strategies, to buy or
not. The firm can produce either a low-quality product or a high-quality product. Customers
can not tell the quality until after they have bought it. Payoffs are as follows:

Firm

Consumer
Strategy Low-quality High quality
Dont Buy 0,0 0,-10
Buy -10,10 1,1

a) Find any dominant and dominated strategies. What is the Nash equilibrium of this
game? Are there any outcomes that would be preferred by both the firm and the
consumer?
Buy Not
HP 3, -2 0, 0
LP 1, 0.5 0, 0

Problem Sets 30

b) Now suppose that there are many potential customers and that consumers buy often.
Suppose that the consumer tells the firm, Ill buy your product and continue to buy it
if it is of good quality. But if it turns out to be shoddy, Ill tell all my friends never to
purchase anything from you again. Under what circumstances would such a
statement make a difference?
c) There is a common perception that locals get better service in restaurants than tourists.
Discuss briefly in the light of the analysis above.

(5) Vertical relations and vertical integration

Firm D is a supplier of Firm G, a large car manufacturer. For each design cycle, firm D makes
an investment of x dollars in design of the new model. Then firms D and G negotiate on a
price for G to pay D for its design work. Let V(x) = x
0.5
be the value generated by firm Ds
design efforts. In the past, it has always been the case that negotiations led to an equal split of
the value at stake, that is, each firm gets 0.5V(x).
a) What is the socially efficient level of investment?
b) What is the equilibrium level of investment?
c) Explain the difference between a) and b)
d) How might the social optimum be achieved?


Practice Questions 31

Practice Questions I

1. Demand for French wine
What will be the effect of the following events on the demand for French wine and the quantity
consumed? Be sure to distinguish between shifts of the demand curve and movements along the
curve. As appropriate, indicate whether you expect the impact will be mostly on price or
quantity; and distinguish between the short-run and the long-run impact. Use a graph if
necessary.
a) A decrease in the price of French wine.
b) A new study linking longevity with moderate amounts of red wine.
c) An increase in the price of Californian wine.
d) A severe drought in the wine growing regions of France.
e) A subsidy to French vineyards.

2. Costs, Elasticity
The market for cola flavored soft drink in Peru is for all intents and purposes a monopoly.
Inca Kola has an overwhelming market share.

The demand for Inca Kola in Peru for each year is given by

Q = 200 4 P.

The Inca Kola has a single plant which it built for 150 million in 1996. Each year it has a
loan repayment of 6 million to pay off the loan it took out to finance the plant. The plant
itself is fairly specialized and has zero scrap value, but shelving and office equipment is
worth 10 million. Each year Inca enters into a bulk electricity contract with Peru Electricity
in which they pay 2 million upfront for all the electricity they want. Their marginal cost
(which is mostly labor and maintenance) is 2 per unit.

(a) At the beginning of each production year, what components of Inca Kolas fixed costs
are sunk?

(b) How does you answer to (a) change if we are in the middle of a production year?

(c) Draw a graph of the marginal cost curve.

(d) Derive the marginal revenue function for Inca Cola

(e) Draw a diagram with Price on the vertical axis and Quantity on the horizontal axis that
shows Incas Demand curve, Marginal Revenue curve and Marginal Cost curve.

(f) Derive the optimal price and quantity for Inca Cola.


(g) On the diagram you drew in part (e) indicate the optimal price and quantity (you do not
need to have done part (f) to do this.

Remember elasticity? Answer the following:

(h) What is the elasticity of demand when demand is equal to 40?

(i) What is the elasticity of demand when demand is equal to 10?

(j) Why do we use elasticity measures rather than just the slope of a demand curve to talk
about how quantity responds to price changes?

3. Costs, Demand and Supply
Provide responses that are both concise and precise to the following short answer questions. Where
appropriate, the use of diagrams is encouraged.

a. Explain how accounting profit and economic profit are different.

b. When does a cost become sunk?

c. Using a demand and supply framework, and explaining each step in your logic, address
the following:
1. Describe the likely effects of a drop in the cost of employing skilled
craftsmen on the market for fashionable shoes.

2. How might an increase in the cost of car insurance affect the US
market for new cars?

3. Global warming seems likely to increase migration from the south
of the USA to the northern states. It also seems likely to increase
the need for insulation in the construction of houses in the
northern states. Taking these assertions as given, how would the
housing market in North Dakota be affected by global warming?

4. Mayor Bloomberg wants to make all the taxis in New York hybrid
cars by 2010. What effect will this have on the market for used cars
in New York?

4. Buttons/badges
As election day approaches in the mayoral race, the demand for buttons is hitting
unprecedented heights. Three months before election day, your friend John, who currently has
a small factory producing badges, is considering expanding production. Previously, this very
competitive industry was relatively stable and demand and supply were estimated at:
Demand: P Q 2 100


Supply: P Q 2

(Units are 100,000s of buttons and price in cents.)

John estimates that over the past months, monthly demand has risen to a higher level:

P Q 2 140 .

He estimates demand will remain at this level until the election, and then fall back down to the
old level. John calculates that by buying a new machine for $5,000 he would be able to
produce 40,000 more buttons a month at a marginal cost of 30c. (Assume that there are no
other additional costs of production.)

a) What was the old price of buttons? How many buttons were being produced at this
price?

b) What is the new price of buttons? How many buttons are produced at this price?

c) Should John buy the machine? In answering this question, notice that if John went
ahead and bought the machine his additional production would be tiny in comparison
with the overall size of the market and would have negligible effect on price. Also, in
no more than a couple of sentences mention any drawbacks in this analysis.

5. Water Filters

Assume that the market for water filters in the US has the following yearly demand function:
Q=5,000,000-100,000P

Suppose that the current price for a filter is $25.
(a) Compute the price elasticity of the demand at the current price.
(b) Compute the consumer surplus at the current price.

6. NY Mets pricing
Suppose the average demand for tickets at a NY Mets game is given by Q = 120,000 (1
P/140), where Q is number of tickets and P is price in dollars.
a) Assuming that marginal cost is zero, determine the optimal ticket price.
Suppose now the Mets stadium has a capacity of 55,000 seats.
b) What is the optimal price?
c) How much should the Mets be willing to pay for a 10,000 seat expansion of their
stadium?


7. East Village Voice
The East Village Voice, a popular newspaper appealing to the young Manhattan population,
shows the following P&L statement. Total production costs (including typesetting, printing
and overhead) are $24m a year. Average circulation, at the current price of $0.50, is 200,000
(for the purpose of this problem, assume that the newspaper is published 300 times a year). In
the past, an experiment was made to increase the price to $0.60; circulation dropped by 30%.

a) Determine the value of the demand elasticity at the current price level.
b) Assuming that printing is 25% of total production cost and making any other necessary
assumption, determine the value of marginal cost.
c) Determine the value of marginal revenue at the current price level.
d) Based on the values obtained in (a), (b), and (c), indicate whether price should be
increased, decreased, or kept constant.
e) Determine the optimal price level.

8. NYU Professional school
NYUs professional development school estimates that for the average potential enrollee the
demand for courses that they take is given by q=190-15p where q is the number of courses
taken and p is the price per course. Suppose that the cost of administering and providing the
course is a constant marginal cost of 10 per student.

a) What is the profit maximizing price for a course?
b) What is the elasticity at this price?
c) Suppose that NYU is considering introducing a fixed registration fee in addition to the
per course charge (i.e., a two-part tariff). Would you recommend such a change in the
way that NYU charges its students? Would the price per course change? What fixed fee
would you charge?

9. SpokenWord
Your software company has just completed the first version of SpokenWord, a voice-activated
word processor. As marketing manager, you have to decide on the pricing of the new
software. You commissioned a study to determine the potential demand for SpokenWord.
From this study, you know that there are essentially two market segments of equal size,
professionals (one million) and students (two million). Professionals would be willing to pay
up to $400 and students up to $100 for the full version of the software. A substantially scaled-
down version of the software would be worth $50 to students and worthless to professionals.
It is equally costly to sell any version. In fact, other than the initial development costs,
production costs are zero.

(a) What are the optimal prices for each version of the software?

Suppose that, instead of the scaled-down version, the firm sells an intermediate version that is
valued at $200 by professionals and $75 by students.


(b) What are the optimal prices for each version of the software? Is the firm better off
by selling the intermediate version instead of the scaled-down version?

10. Equality and night-club pricing Barmy Brussels Battles Bars
We read with interest that there is discussion of new legislation being introduced throughout
the European Union aims to end differential pricing between males and females, this being
viewed as a discriminatory practice. One of the leading firms in the industry, BANCA, the
European-wide Bars and Nightclubs Associated, which has considerable market power in this
industry, has protested arguing that such legislation would a have severely detrimental effects
on the profitability and viability of its business. Moreover, BANCA argues that its customers
would suffer through the introduction of such legislation.

In no more than a paragraph for each point, discuss
(a) How one might try to assess the extent to which customers might lose or benefit from
the current arrangement;
(b) What might account for the current pricing arrangements; and,
(c) Whether there are any particular features of this industry which one might take into
consideration.

11. Market shares. The following tables summarize 2001 market shares for investment
banking and credit cards (data courtesy Mike Mayos March 2002 analyst report on Citigroup
for Prudential Financial):

(a) Compute the HHI for each of these markets.

(b) What would the DOJ merger guidelines suggest about a merger between Goldman
Sachs and Merrill Lynch?

(c) What would the DOJ merger guidelines suggest about a merger between Banc One and
American Express?

Investment Banking (Fees)
Goldman Sachs 13.7%
Morgan Stanley 11.2%
Merrill Lynch 11.0%
Citigroup (SSB) 10.4%
CSFB 9.5%
JP Morgan 6.9%
UBS Warburg 4.5%
Lehman Brothers 4.4%
Deutsche Bank 3.4%
Dresdner Kleinwort 3.4%
Credit Cards (Receivables)
Citigroup 18.0%
MBNA America 13.5%
First USA/Bank One 12.3%
American Express 9.3%
Discover 8.7%
Chase 7.4%
Providian 5.9%
Capital One 5.8%
Bank of America 5.1%
Household Bank 2.9%


Practice Questions II

1. Specialized computer chip
You are currently the only producer of a specialized computer chip used in computer animation,
but are concerned that your success might attract competitors. You are now considering a major
investment in capacity that would reduce your marginal cost below that of any potential rival. The
impact of the investment on your strategic position is summarized below:
If you do not make the investment, you expect to make $20m if competitors enter, and
$50m if they do not. Your potential competitors, on the other hand, expect to make
$10m if they enter, nothing if they do not.
If you do make the investment, you expect to make $10m net of the investment if
competitors enter, and $40m (also net) if they do not. Your potential competitors, on the
other hand, expect to lose $2m if they enter, nothing if they do not.
Use a game tree to study this game and come up with a recommendation on whether to
make the investment or not. Comment specifically on this advice from a high-paid
consultant: Investment lowers your payoffs when there is entry and when there is not. Its a
mistake.

2. Biological MicroProcessor
Its 2016 and the biological microprocessor, for so long a mere promise of science fiction, has
finally become a commercial possibility. Intel and archrival AMD ponder the pros and cons of
moving ahead with a massive research effort to switch from conventional to biological
microprocessors.
Currently, Intel is making an annual profit of $3.2bn, whereas AMDs profit is $500m.
Analysts expect that, if Intel goes alone with the biological program, their profits will drop to
$2.8bn, whereas AMDs will drop to $400m. If however AMD moves alone, then AMD will
increase its net profits up to $1.2bn, whereas Intel will drop to $1.8bn. Finally, if both firms
initiate their own biological microprocessor programs, then expected profits are $1.6bn for
Intel and $100m for AMD.
a) Describe the game played between Intel and AMD if the two firms have to make their
decisions simultaneously. Determine the equilibrium of this game.
b) Suppose that Intel has the possibility of credibly committing to a course of action
before AMD observes Intels choice and makes its own choice. Describe this new
game and determine its equilibrium.
c) What is the value of commitment here?

3. Repeated Bertrand

Suppose that in an industry there are 3 firms competing in prices. Their marginal cost is the
same, constant and equal to 100. The demand is given by Q = 1000 P.
a) What is the outcome of this game if the market exists just for one period?


b) Suppose instead that these firms are on this market for the long run, and they play the
Bertrand game infinitely many times. At every period there is the same demand, and
firms can reset their prices. Suppose that the interest rate is r = 5%. Can cooperation
be sustained?

4. On campus corporate presentations: Sympathy for the Devil
Many students complain about the high number of corporate presentations taking place both
on and off-campus. In this question, we seek to explore this phenomenon.
Suppose that Stern is playing a game with other business schools (to be specific, think
of Columbia). Both schools are concerned with ensuring that their students find gainful
employment at the end of their courses of study. Specifically, suppose that they assign a value
of 100 from doing so. Suppose that if neither or both schools allow corporate presentations,
then recruiters are equally likely to hire from either school; however, if only one school allows
presentations then its students are likely to be hired with 75% probability, whereas the other
schools students chances of getting hired fall to 25%. Allowing corporate presentations takes
time, which could be used for additional classes or recreational activities to which the school
assigns a value 10.

a) Describe the elements of this game. What would you expect to be the outcome?
b) A limitation of the analysis in (a) is that this game is played out by Stern and Columbia
every year. Suppose that each has an annual interest rate of 5%. It is possible to
achieve a better outcome than in a). Explain how and illustrate your argument formally
(that is, if you argue for a different equilibrium then prove that this is indeed an
equilibrium).
c) Can you think of any factors that might make it difficult to achieve this better
outcome? That is, are there any limitations or anything missing from this description
of the game?
d) In the remainder of the question, we highlight a different aspect of on-campus
presentations. Specifically suppose now that recruiters prefer to speak to full rooms
rather than empty ones. If speaking to empty rooms, in a pique of annoyance the
recruiter will not hire from the school (more likely they will not return the following
year). Suppose moreover that there are ten genuinely interested students, who will
come to the presentation and each of whom values the presence of a happy recruiter at
15, but an unhappy one at 0, but for the other 90 students there is an opportunity cost
of 1 in coming to the presentation whether the recruiter is happy or not (they would
rather spend their time reviewing Firms and Markets). What would be the result? Can
the school do anything, which would raise overall welfare?

5. Suits R Us
A new breed of investment firm is capitalising on the boom in litigiousness and taking a piece
of the profits. [These firms] underwrite lawsuits in exchange for a share of proceeds.
(Financial Times, Dec 3, 2001.)

One example is Suits R Us. Potential plaintiffs come to them with cases. For each one, SRU
decides whether or not to take it; if so, it offers a standard percentage rate contract. Given


SRUs experience with many previous cases, the firm has a better idea than most plaintiffs of
the odds that the case will win in court.

(a) Suppose you have a potential case in hand (in which you would be the plaintiff). What
considerations should you take into account when deciding whether to apply to Suits R
Us?

(b) Name two other situations that lead to similar problems.

6. Buffett and GE
Paraphrased from the Wall Street Journal (November 13, 2002):

GE is apparently trying to sell Employers Re, its huge reinsurance operation. Reinsurers
provide insurance to insurers, allowing insurers to spread their risks or cap their potential
losses. The top 3 players in this business are Munich Re, Swiss Re, and General Cologne,
which is controlled by Warren Buffetts Berkshire Hathaway. GE CEO Jeffrey Immelt stated
that Employers Re was a business he wanted to reposition for value translation:
consolidate, sell, shrink or fix. Why? Employers Re was a big money maker after its 1984
acquisition by GE, but has become a trouble spot of late. In October, GE said that Employers
Re expects a 2002 loss of $350 million to $450 million, and break-even results in 2003.

Selling the business has proved difficult. A public offering was derailed by a sour IPO market
and steep losses in the unit. Observers also cite concerns about the adequacy of the units
claims reserves and Employers Res own use of reinsurance. A large reason for concern about
reserves is mounting asbestos litigation, which is adding billions of dollars to the insurance
industrys already huge exposure. Concerns about reinsurance revolve around Employers
purchase of financial reinsurance, which is known for smoothing earnings as well as
spreading risk. Under its typically complex terms, the purchaser often pledges future
investment income to the reinsurer selling the coverage. Berkshires National Indemnity unit is
one of the biggest providers of financial reinsurance.

With Munich Re, Swiss Re, and some other potential bidders in financial distress, Berkshire
may emerge as the only bidder. Berkshire would likely wind down much of Employers Res
operations, using its own claims-handling expertise to try to save money along the way, and
squeeze bigger profits out of the parts it keeps. Analysts suggest that Mr. Buffett isnt likely to
do the deal unless it looks like a no-lose situation. Any deal, therefore, is likely to include
limitations on how soon Berkshire would pay losses on the policies it acquires or some other
sort of cap on payments.

(a) Why might GE want to sell Employers Re, even though in the current market its likely
to be forced to accept poor terms?
(b) Why might Employers RE be worth more to Buffett than to GE?
(c) What problems do you see to completing a deal? What solutions would you
recommend to resolve them?


Lecture Notes 39

Introduction to Firms and Markets

The goal of this course is to give you some insight into how markets work. The course is
structured in two parts. In the first part of the course, we study decision making by consumers
and firms. We explore the fundamentals of demand and identify categories of costs that firms
must consider when taking critical business decisions like pricing, exit or market entry. We
study how supply and demand determines prices in efficient markets. We learn about market
power and how the interplay between cost and demand fundamentals determines profit-
maximizing decisions for firms. The second part of the course focuses on situations where, for
one reason or another, markets dont work efficiently. Here, we emphasize the importance of
strategic behavior, as modeled by game theory. Strategic interactions between firms in markets
can be represented as games and we learn to predict the outcomes of such games and analyze
how best firms can respond to their rivals strategies. We discuss the basics of competition
(how do firms compete on price?) and cooperation (how do firms collude?), asymmetric
information (what happens when sellers know more than buyers?) and market design
(auctions).

Microeconomics (as the broad topic of this course is often called) is an important component
of an MBA program. First, microeconomics focuses on specific dimensions of optimal firm
decision making, such as pricing, entry and exit. Second, the formal economics perspective on
business plays an important role in other fields of study, such as finance or marketing. Finally,
by studying public policy towards market failures, microeconomics highlights important factors
conditioning firm strategy.

Our experience with students in prior years is that much of what we learn in this course is
intuitive. But much is not, and our hope is that the combination of theoretical structure and
practical examples will be useful to you in the years to come. It will not make you a success on
its own, but it might give you an edge a few times when it matters!

Business can be viewed from many perspectives, including those of sociology and psychology.
Economists tackle the subject by adopting a formal, explicit analysis of decision-makers that is
based upon a number of assumptions. First, we assume that individuals make rational
decisions to maximize some objective given a range of feasible choices. For example, firms
might maximize profit or consumers might maximize utility (a catch-all to include whatever
they care about). Secondly, we generally assume that a firm is a single decision-maker with a
clear objective. Third, we assume that most economic interactions take place in markets,
where buyers and sellers interact through price. All of these assumptions are questionable, and
perhaps even wrong, but they bring some clarity to our analysis that we feel is useful.

Our task is therefore to determine how individuals define their goals and examine the actions
they take to attain those goals, and then analyze the outcomes that result from their actions.
We focus on how buyers and sellers interact through markets, how firms adopt strategies to
Firms and Markets
Lecture Notes

Lecture Notes 40

interact with each other, how competitive forces bring resources to their most profitable use,
how these same forces can make it difficult to sustain above-average profits, and how
governments set rules to limit the power and influence of individual firms. The trick in all of
this analysis is to simplify (eliminate nonessential complications) without losing the flavor of
the situation being examined. Its a delicate balance, and one that makes the subject of this
course as much art as science.

Markets

Markets are incredibly numerous and varied. The New York Stock Exchange is one
prominent local market, but so are the farmers market at Union Square and the markets for
pharmaceuticals, cable TV programs, and consulting and legal services. Markets can be highly
regulated, such as the US aerospace procurement market, or as informal as buyers and sellers
haggling over cheap sunglasses on a street corner. Our formal definition of a market is a
collection of buyers and sellers who interact in transactions for a product or a set of products.

Although we tend to focus on a market, in practice its not easy to decide where one market
begins and another ends, either geographically or in the range of products. Are Balduccis and
the Union Square farmers in the same market? It depends on the issue. This is more than a
theoretical nicety. The Staples/Office Depot anti-trust case, which effectively ended their
proposed 1996 merger, hinged on precisely this issue. Was WalMart part of the same market
(in which case the combined firm would have a small market share) or not (in which case the
market share in many areas was substantial)? The courts ruled not, agreeing with the FTC
that the merged company would have an unacceptably high market share in many locations.

Markets (once we agree on their boundaries) differ along many dimensions: the number of
buyers and sellers, the degree of differentiation among products, the transparency of market
transactions, and so on. On one end of the spectrum, we consider perfectly competitive
markets, in which many buyers and sellers of a uniform product make transactions in
transparent markets. In a perfectly competitive market the large numbers of buyers and sellers
means that no individual firm has any meaningful control (market power) over the price. Its
an extreme case, but illustrates the impact of competitive forces. Managers abhor such
markets, since competition makes it very difficult to make money. Such products are referred
to disparagingly as commodities, since commodity markets are often good examples (wheat,
basic steel, photocopy paper).

At another end of the spectrum, markets with one seller are referred to as monopolies.
Monopolies are said to have market power, generally restricting output and setting prices above
what wed see under perfect competition. Famous near-monopolies over the years include
Standard Oil at the turn of the century, Deutsche Telecomm, and DeBeers (a monopoly
distributor of diamonds).

While some markets lie at the extremes, the majority are located somewhere in between. We
refer to markets with a few sellers as oligopolies. These markets are much more common.
Theyre also more interesting and challenging to study, since some oligopolies are extremely
competitive, while others behave much like monopolies. Hence a central issue for us will be to
understand the features of a market that make it more or less competitive.

Lecture Notes 41

Firms

The large business firm is one of the distinguishing features of the modern world. For the last
century or more, many products have been made and sold by large organizations. The MO
course is devoted, in part, to how such large collections of people work. We operate at a more
macro level, treating each firm as a single object maximizing (say) profits. More on this
shortly. Like markets, firms are diverse along many dimensions, including geographical span
(think of Coca-Cola versus the Campus Eatery), in span and scope of product offerings
(General Electric versus a shoe store), and in degree of vertical integration. Over the course of
this class, we will learn about many distinguishing characteristics that lead to differentiation and
competition among firms.

In focusing on firms, we make an important assumption about firm behavior: firms and
managers maximize profits or shareholder value (we think of them as the same thing). Does
this make sense? On the face of it, how can the 100,000 employees of Citigroup be thought of
as working together to maximize anything? Why would the managers of General Electric work
for the good of shareholders? How can the shareholders know whether the managers are
doing this or not? Although we own a few shares of many firms through mutual funds, we do
not have the time or resources to monitor managements actions, other than what we read in
the New York Times or Wall Street Journal. And even though Boards of Directors are ostensibly
looking out for our interests, they are rarely independent from the CEO and perhaps other top
operatives in the company.

How, then, might we defend profit maximization? We might argue, first, that Darwinian
competition will weed out the firms that dont. Or we might argue that managers do this to
maximize their own reputations, and that competition leads managers with good reputations to
be paid better. Perhaps the most compelling argument is based on capital markets: firms that
do not maximize value will be taken over and their management replaced. None of these
arguments is watertight, but they give us some hope that the assumption is, at least, a
reasonable approximation.

It Depends

There is an old joke that people love to tell about economists. Harry Truman, the joke goes,
once asked for a one-armed economist. His advisors asked: Why one arm? Because,
Truman remarked, economists always answer questions with: on the one hand it's this, on the
other hand it's that. For once, I'd like to have an economist who will give me one straight
answer.

We need to tell you now: If you're looking for the one-armed economist, this isn't the course
for you. Economics is not a set of answers. It is, instead, a framework for thinking about
questions systematically. It will not give you the answer, but it will allow you to ask the right
questions.

The answer to most questions in this class is, then, It depends. After taking the course, you
will be able to say what it is they depend on.

Lecture Notes 42

Calculus Review

In microeconomics, calculus gives you a sharper understanding of (among other things)
elasticity and pricing. You can learn microeconomics without calculus, but the logic is
invariably less clear. The same is true in some other fields. In finance, for example, calculus is
the natural tool for understanding portfolio choice (the minimum variance portfolio) and bond
duration (the sensitivity of price to yield). What follows is a short, (relatively) non-technical
review of the aspects of calculus needed for the course.

Functions

In economics, we often use relations between two variables: demand depends on price, cost
depends on quantity produced, and so on. We call these relations "functions." Formally, a
function f assigns a (single) value y to each possible value of x. We write it this way: y = f(x).
In a spreadsheet program, you might imagine setting up a table with a grid of values for x. The
function would then be a formula that computes y for each value of x.

Perhaps the easiest way to think about a function is to draw it: put x on the horizontal axis and
plot the values of y associated with each x on the vertical axis. Some examples are given in
Exhibit 1. We'll generally be interested in functions that are "continuous" [they don't have
"jumps," as in (b)] and "smooth" [they don't have "kinks," as in (c)].

Exhibit 1: Examples of Functions

Firms and Markets
Lecture Notes
y = f(x)
x
(a)
(b)
(c)

Lecture Notes 43

Slopes and Derivatives

The "slope" of a function is a measure of how steep it is: the ratio of the change in y to the
change in x. For a straight line, we can find the slope by choosing two points and computing
the ratio of the change in y to the change in x. For most functions, though, the slope is
different at every point. Take Example (a) in Exhibit 1, for example. The slope is initially
positive as the function increases, then turns negative as the function slopes down, then turns
positive again at the end.

Suppose that f(x) is something like profits that we want to maximize. We are looking at a
function like the one below, trying to find the value of x (that we will call x*) where f(x) is the
biggest. It might be a whole number, x*=6 or x*=7. It might also be a messy number, like
x*=6.2223453443. (Using discrete methods, you can check at 1, 2, 3, 4, . but not at more
messy numbers.) But, if we have an easy way of calculating the slope of the line, we have an
easy way of finding x*: its the point at which the slope is equal to zero. (Think of f(x) as being
a kind of hill that you are climbing, starting at x=0; the slope goes up for a while, then it is zero
once you reach the top of the hill, then it slopes down for a while. Sloping up is a positive
slope, sloping down is a negative slope.)

y=f(x)

x* x

What is a slope? Its an easy concept for a straight line, since a straight line has a constant
slope. On a straight line, start anywhere and then increase x (by any amount). The slope is the
amount y goes up, divided by the amount x went up:

Ay/Ax =[f(x+Ax)-f(x)]/Ax

Lecture Notes 44

Lets take an example: y = 3x + 1. And we graph the line:

Exhibit 2: y= 3x+1
y y=3x+1

1 2 x

For example if you are initially at x=0, then y=1. If you move to x=2, then y=7 at that point.
So the slope is 3
0 2
1 7
=
=
A
A
x
y
. But I could also have chosen two points very close to each
other: For example if you are initially at x=1, then y=4. If you move to x=1.05, then y=4.15 at
that point. So the slope is 3
1 05 . 1
4 15 . 4
=
=
A
A
x
y
, the same slope. Now suppose that you have a
curvy line: for example, y=x
2
. Then the slope is different at every point.

Exhibit 2: y= x
2

1

1 x

Now, to get a measure of slope that is as accurate as possible, you need to choose two points
that are incredibly close together: Lets say we want the slope at x=1. Lets look at x=1 (where
1
4
7

Lecture Notes 45

y=1) and x=1.0001 (where y=1.00020001); the rough measure of slope that we get is
0001 . 2
1 0001 . 1
1 00020001 . 1
=

=
A
A
x
y
. The closer you pick the second value of x to 1, the closer you
get to a slope of exactly 2.

The "derivative"
dx
dy
of a function f(x) gives us its slope at each point x if the function is
continuous and smooth. Formally, we say that the derivative is
x
y
A
A
as Ax gets "incredibly
small." We write the derivative as f '(x) or
dx
dy
or
( )
dx
x f d ) (
. In the latter case, the d's are
intended to be suggestive of small changes, analogous to Ay/Ax but with the understanding
that we are talking about small changes in x.

There are formulae you can just use for the derivatives that we are going to use. If they're new
to you, don't try to make sense of them. Take them as facts to be memorized and put to work.
(Exhibit 4 tells you how we got the formulae, if youre interested. And Exhibit 5 gives you
more formulae you can use, in future classes.)

0
) 10 (
=
dx
d
If y=10, thats a flat straight line, and the slope is zero.

5
) 5 (
=
dx
x d
If y=5x, when you take the derivative, you just drop the x.

x
dx
dx
2
2
= If y=x
2
, when you take the derivative, you bring the 2 down in front.
Notice that this is what we found in Exhibit 3, looking at the curve y=x
2
: when x=0, the
derivative is 2 times 0, which is zero; that is, the line is flat. When x=1, the derivative is 2 times
1, which is 2.

So if you have a line that is y= 3x + 1, you take the derivative of each bit in turn, and you get
3
) 1 3 (
=
+
dx
x d
, which is what we found in Exhibit 2.
If you have a line that is y=5x
2
+ 4x + 7, the derivative is 4 10 + = x
dx
dy
.

Examples. Find the derivatives of the following functions [answers in brackets]:
- 2x + 27 [2]
- 2x
2
+ 3x +27 [4x+3]
- 2x
2
+ 3x 14 [4x+3]
- (x2)(2x+7) [4x+3]

Lecture Notes 46

Technical point. If the function has jumps or kinks, the notion of slope simply doesn't make
sense at those points. Take Example (b) in Exhibit 1. What is the slope where the function
jumps? We could find the slope just to the left of the jump, or just to the right, but not at the
point where it jumps. Similarly, it's not clear what the slope is at the kink in Example (c). We
say, in these cases, that the derivative doesn't exist. We'll make sure that most of our functions
are continuous and smooth so that this hardly ever happens.

Finding the Maximum of a Function

Now to our purpose. We'll often want to know the value x that produces the maximum value
of a function f(x) for x between (two numbers) a and b. We do this by setting the derivative
f'(x) equal to zero and solving for x. Why does this work? You can see in Example (a) of
Exhibit 1 that a function is flat (has zero slope) at a maximum. We will use this insight.

Example. Find the maximum of this function: Profits (Q) = -Q
2
+ 9Q 4, with respect to Q.
(What does that mean? That means Profits are a function of Q, and I want to know the slope
as Q changes; Q is like the x term.)
The derivative is 9 2
) 4 9 (
2
+ =
+ +
Q
dQ
Q Q d
. So, at the optimal Q*, the derivative is zero,
and so that means -2Q* + 9 = 0, which I solve to get Q*=4.5. Profits are maximized by
producing exactly 4 (say we are measuring quantity in tons, this is four-and-a-half tons).

Profits

4.5 Q

-4

Try a couple: Find the maximum of
- -5x
2
+ 2x + 11 [f '(x) = -10x+2 = 0, x = 5]
- function (a) in Exhibit 1 [there are two points with zero slope, but only the first is a
maximum]
- Profits = - 2Q
2
+ 5Q + 12

Lecture Notes 47

Technical aside. Does this always work? If we set the derivative equal to zero, do we always
get a maximum? The answer is no. Here are some of the things that can go wrong:
(a) The point might be a minimum, rather than a maximum. For example, in Example (a) of
Exhibit 1 the function has both a maximum and a mimimum. Both have
derivatives/slopes of zero.
(b) The maximum could be at one of the endpoints, a or b. There's no way to tell without
comparing your answer to f(a) and f(b).
(c) There may be more than one "local maximum"
(d) The slope might be zero without being either a maximum or a minimum: for example, the
function might increase for a while, flatten out (with slope of zero), then start increasing
again. An example is the function f(x) = x
3
at the point x=0. [Draw this to make sure you
understand the point.]
All of these things can happen in principle, but our job is to make sure they don't happen in
this class. (If you want to be extra careful, there are ways to check for each of these problems.
One is the co-called second-order condition referred to in our notes.)

Maximizing Profit

Here's a common application. Suppose a firm faces a demand for its product of q = 10 2p
(q and p are quantity and price, respectively). The cost of production is 2 per unit. What is the
firm's profit? What level of output produces the greatest profit?

Answer. Profit is revenue (pq) minus cost (2q). The trick is to first express it in terms of
quantity. We need to use the demand curve to eliminate price from revenue: p = (10-q)/2 so
pq = [(10-q)/2]q. Profit (expressed as a function of q) is therefore

Profit(q) = [(10q)/2]q 2q = 5q q
2
/2 2q.

To find the quantity associated with maximum profit, we set the derivative equal to zero:

Profit'(q) = 3 q = 0,
So q = 3. What's the price? Look at the demand curve: if q = 3, then p satisfies 3 = 102p,
so p = 7/2. [Please check!!]

Further Reading
There are lots of good calculus books. One good review is
- Bernard Zandy, Cliffs Quick Review Calculus, Cliffs Notes, 1993

Lecture Notes 48

Exhibit 4: Rules for Differentiation

Function f(x)
Derivative f'(x)
Comments
Rules for Combinations of Functions
g(x) + h(x)
g'(x) + h'(x)

ag(x) + bh(x)
ag'(x) + bh'(x)
a and b are numbers (constants)
g(x)h(x)
g(x)h'(x) + g' (x)h(x)

g(x)/h(x)
[g' (x)h(x) g(x)h'(x)]/[h(x)]
2

g[h(x)]
g'[h(x)]h'(x)
"chain rule"
Rules for Specific Functions
a 0 a is a number
ax + b a a and b are numbers
ax
n
anx
n-1
a is a number, n an integer
ae
bx
abe
bx

a log x a/x a is a number,
log means "natural log"

Exhibit 5: Where do the formulae come from?
When we talk about the slope of a curvy line at a particular point, what do we mean? The
simplest way to think about it is to relate it to the slope of a straight line. At any given point,
say x=1, there is a line that is parallel to the curve at just that point; in other words, it has the
same slope as the line at that point. (We call that line the tangent.) So when we talk about
slope, we mean the slope of the tangent line.

y= x
2

1

1 x

Lecture Notes 49

One geometry way of getting the tangent line at x=1 is to draw the line between x=1 and
another point on the curve, say x=2; then draw the line between x=1 and x=1.5; then, between
x=1 and x=1.2, and keep getting closer. The closer you get, the closer you get to the tangent.

1

1 x

Now it is easy to find the slope of the line that goes between x
0
=1 and another point, x
1
; the
slope is simply
0 1
0 1
x x
y y
. So, if your two points are x

0
=1 (which means y
0
=1) and x
1
=1.2
(which means y
1
=1.2
2
=1.44), then the slope of the line between those two points is
2 . 2
1 2 . 1
1 44 . 1
0 1
0 1
=
x x
y y
.

That means that as I choose a point x
1
that is closer and closer to x
0
, I will get a slope that is
closer and closer to the slope of my tangent. In math terms:
0 1
0 1
at
0 1
0
lim
x x
y y
dx
dy
x x
x

=

and we use this to calculate the derivative.

So, for example, lets take the function y=f(x)= 3x+1. We want to find the derivative of this
function at some point x
0
. It is:

Lecture Notes 50

( ) ( )
( )
3
3
lim
1 3 1 3
lim lim
0 1
0 1
0 1
0 1
0 1
0 1
at
0 1
0 1 0 1
0
=
+ +
=

x x
x x
x x
x x
x x
y y
dx
dy
x x
x x x x
x

And so we find that the slope is 3, no matter what specific point x
0
we were interested in.

Lets take another example, the function y=f(x)=x
2
. We want to find the derivative of this
function at some point x
0
. It is:
0
0 1
0 1
0 1 0 1
0 1
2
0
2
1
0 1
0 1
at
2
) ( lim
) )( (
lim
) ( ) (
lim lim
0 1
0 1
0 1 0 1
0
x
x x
x x
x x x x
x x
x x
x x
y y
dx
dy
x x
x x
x x x x
x
=
+ =
+
=

(factorizing)
Because x
1
gradually becomes x
0
. So, if I want to know the derivative at point x
0
=1, I plug that
in, and I get 2.

Net Present Value (NPV)
When the interest rate is r, the net present value of a cash flow X
0
, X
1
, X
2
, , X
n
is

In particular, if n= and X
0
= X
1
= =X, then

Thus,

Lecture Notes 51

Supply and Demand: Buyers, Sellers, and Markets

There are many situations in which you might want to ask: What is a reasonable price for a
product X? You might be an equity analyst trying to value a firm that produces X. Or you
could be a consultant trying to evaluate the benefits for a client of entering the market for X.
Or you could be a venture capital investor trying to assess the prospects of a startup planning
to produce X. A glib answer is to look at the market's current price, but there are times that's
not enough. It's not enough, for example, in a market that does not yet exist. And its not
enough when you think the current price is likely to change. When production at Canada's
Diavik diamond mine comes online in 2003, what is the likely impact on the price of
diamonds? What is the impact on airline fares of an increase in the price of jet fuel?

The supply and demand diagram is the basic tool of analysis for addressing such questions. In
this lecture, we describe the tool and explain how to use it. A critical distinction is the
difference between movements along curves and factors that cause one or both of the curves
to shift. Unless you've had economics before, this will sound a little mysterious, but it should
be less so by the end of class. In future lectures, we go into greater depth on the supply and
demand curves themselves.

Buyers: Demand

You might think of demand as summarizing buyers willingness to buy a product at various
price points. As the price decreases, generally more people were willing to buy more units. A
demand curve represents graphically the relationship between quantity desired and price: the
number of units people are willing to buy (the demand) at a given price. By tradition, we
graph this relation with price on the vertical axis and quantity on the horizontal access. This
can be confusing, but its hard to break a century-old habit among economists. The demand
curve is normally downward-sloping, meaning that more people are willing to buy a good if the
price is lower. Similarly, if someone is willing to buy a product at one price, he or she is
generally willing to buy it at a lower price, as well.

The location and shape of the demand curve depend on peoples tastes. Hence the demand
for cod fish is greater in Portugal than it is in Japan, and the demand for rice is greater in Japan
than Portugal. The demand for wireless phone service is relatively sensitive to price (the
demand curve is flat) but the demand for basic service is not (steep), since people are less
willing to forgo basic service to save money for other uses. We often represent the demand
curve as a straight line, but that's only a graphical convenience.

The quantity of a product people are willing to buy depends, of course, on more than price. It
can depend on the prices of other products, the incomes of buyers, the age distribution of the
population, and lots of other things. We represent all of these other factors in a demand
Firms and Markets
Lecture Notes

Lecture Notes 52

function, in which quantity demanded is a mathematical function of as many of these other
variables as we need. (Pay close attention to whats coming!) The question is how to picture
changes in these variables in a diagram with price and quantity on the axes. We do so by
shifting the demand curve up or down when these other variables change. The demand curve
gives us the relationship between price and quantity for given values of the other variables. If
any of these other variables changes, the demand curve shifts.

Now to some examples. Prices of other goods often affect the quantity demanded. For
example, if the price of Coke rises, we might expect the quantity of Pepsi demanded to
increase at any given price. The demand curve for Pepsi shifts up and to the right. We say that
Coke and Pepsi are substitutes since an increase in the other's price leads to an increase in
demand. If the dependence goes the other way, we call the products complements. For
example, if the price of gas goes up, we would expect the demand for SUVs to decline.

Population and income also affect demand. For example, the demand for undergraduate
education is increasing due to a demographic bulge of people of college age. We could
represent this as a shift up/right of the demand curve for education. Income also affects
demand for many products, with more people willing to buy higher-price goods as income
increases. In the UK, for example, the demand for relatively dclass vacations in Southern
Europe has declined as people have become prosperous enough to afford vacations in Florida
and the Caribbean. We can represent this as a shift down/left in the demand curve for
European vacations. Advertising can also have an impact on demand, by raising the visibility
and perceived quality of products.

Sellers: Supply

The supply curve represents the quantity sellers are willing to sell at different prices for given
values of other variables. The supply curve is generally upward-sloping since companies will
be willing to produce and sell more of a good as prices go up. For example, crude oil
producers supply more oil at higher prices, since a higher price makes more wells profitable.
Markets differ in how steep the curve is. A high price of electricity has elicited little immediate
increase in supply in California (the supply curve is steep, at least in the short run). But if there
were a sudden increase in the demand for basic blue jeans, we would expect producers to meet
it with little increase in price (supply curve is flat).

Like demand, supply is affected by factors other than price and we represent changes in these
factors by shifts in the supply curve. One such factor is input prices. If the price of jet fuel
rises, we would expect it to increase the price at which airlines are willing to sell tickets (the
supply curve shifts up). Technology also affects the supply curve. If improvements in
technology reduce costs of production, then we would expect the supply curve to shift
down/right. Nature can also play a role. For example, an earthquake that destroys some of
the market's capacity would result in a shift left/up in the supply curve. This happened in the
semiconductor market after the 1999 earthquake in Taiwan. The number of suppliers also
affects supply, with more suppliers tending to mean greater quantity supplied at any given
price. Similarly, an investment in a new plant would result in a shift right/down in supply.

Lecture Notes 53

Once again, it is important to distinguish changes in price from changes in other factors. The
former is a movement along the supply curve, whereas the latter results in a shift of the whole
supply curve.

Market Equilibrium

The price of a product is the result (in market settings) of the interaction of buyers (demand)
and sellers (supply). We refer the point at which the supply and demand curves intersect as
equilibrium. See Exhibit 1. Its an equilibrium in the sense that none of the market
participants have an incentive to change their behavior: buyers are buying what they want at
that price (the point is on the demand curve) and sellers are selling what they want (its on the
supply curve). If the price were higher than the equilibrium price, fewer people would want to
buy than sell. The excess of sellers would tend to drive the price down. Conversely, if the
price were lower than the equilibrium price, fewer people would be willing to sell than buy.
The excess of buyers would cause the price to tend to rise to the equilibrium price.

We thus have an answer to one of the questions with which we started: a reasonable price is
one at which demand and supply are equal. In practice, we need to know what the supply and
demand curves look like, which requires both data and enough expertise and judgment to
construct the two curves. But the logic is just what we've described.

Creating Surplus

Perhaps its not immediately obvious, but trade creates surplus. If trade is voluntary, this has
to be true, or people wouldnt do it. But we can be more precise in the case illustrated by
Exhibit 1. For buyers, the demand curve represents their willingness to pay. The difference
between the demand curve and the market price (labeled A in the exhibit) is thus surplus to
buyers; we call it consumer surplus. Similarly, the difference between the price and the
supply curve represents surplus to sellers: the difference between the market price and the
price at which theyd be willing to sell (labeled B). We call it producer surplus. Total
surplus generated by trade is the sum of the areas A and B.

Comparative Statics

Our other questions concerned changes in market conditions. If the demand or supply curve
shifts due to changes in one of the other factors discussed above, then a new equilibrium
price will be established. The term comparative statics is used by economists to describe the
exercise of looking at what happens to equilibrium if an exogenous factor changes. We would
represent this by shifting the supply or demand curve and noting the change in the equilibrium
price and quantity.

If you work through an example, youll see that the impact on price depends on the slopes of
the supply and demand curves. The effect of a shift in the supply curve depends on the slope
of the demand curve. (This sounds a little strange, but its true because the demand curve
hasnt shifted, so the change in equilibrium is a movement along the demand curve.) If the
demand curve is steep, price will react more to a shift in supply than quantity. If the demand

Lecture Notes 54

curve is flat, the impact of a shift in supply will be predominantly on quantity. For shifts in
demand, the impact depends on the slope of the supply curve.

Thus a key ingredient to any market analysis is an assessment of the slopes of the supply and
demand curves: how sensitive the decisions of buyers and sellers are to changes in price.
Consider the California electricity market. Supply is virtually fixed, since the capacity of local
power plants cant be changed much without building new ones. Similarly, the lines to bring in
power from other states also have limited capacity. Hence the supply curve is very steep
(vertical?) and the impact of an increase in demand (the result of growth of the California
economy) is reflected almost entirely in the price. Another example is the September 20, 1999
earthquake in Taiwan, which damaged plants producing a substantial fraction of the world's 64
MB DRAM chips. This can be viewed as a left/up shift in the supply curve (a reduction in
supply). The price of the chips increased about 50%, from which we might infer that the
demand curve for chips is fairly steep. See Exhibit 2. A third example is copper, whose
quantity has increased substantially while price declined slightly over the last century. (This is a
general property of primary commodities: their prices have tended to go down, not up, over
time.) How is this possible? One explanation is that demand has increased as the world
economy has grown, but supply has shifted, too, as new technologies make copper extraction
more efficient. In other words, both the supply and demand curves have shifted to the right,
with supply shifting slightly faster.

Example

Vitamin C is a generic vitamin that is produced by many companies: brand names are not very
important, entry is easy. A good friend a world-renowned orthopedic surgeon from New
Jersey tells you that he is about to publish in The New England Journal of Medicine (a highly
respected and widely quoted medical journal) a study indicating that daily doses of 500 mg of
vitamin C tends to improve the muscle tone and increase the physical stamina of adults, with
no adverse side effects. Though a very good doctor, he is woefully ignorant about the basic
workings of markets and wants to know what is likely to happen, and why in the short run and
in the long run to the price of vitamin C, to the quantity sold, to the profits of the producers,
and to the number of firms that produce it. Summarize what you would tell him.

Answer. One would expect demand to increase as a result of the NEJM article. In the short-
run, supply is fixed. So, we will observe a move along the supply curve, with both price and
output going up. The extent of the price hike depends on the steepness of the supply curve:
the steeper the short-run supply curve is, the greater the price increase.

In the long-run, one would expect the supply function to expand, as new producers enter the
market and existing producers expand their capacity. Assuming that demand is kept at the
same level, this would correspond to a movement along the demand curve, with output going
up and price going down.

To summarize: We would expect price to go up in the short-run, then back down in the long-
run, possibly to almost the same level as the initial level. As to output, we would expect it to
go up, with a greater increase in the long-run than in the short run.

Lecture Notes 55

Exhibit 1
Supply and Demand

Comment: Price and quantity tend to adjust until buyers (demand) are reconciled with sellers
(supply) at equilibrium with price p* and quantity Q*. The triangle marked A represents
consumer surplus, the triangle B producer surplus.

Exhibit 2
Impact of Taiwan Earthquake on DRAM Market

Comment: Earthquake shifts supply left/up, resulting in rise in price and fall in quantity. Note
that the impact depends on the shape of the demand curve.
Price
Quantity
Demand
p*
Supply
Q*
Price
Quantity
D
p
S
Q
S
Q
p
Equilibrium
E
E
B
A

Lecture Notes 56

Utility

An existing or new business needs to have some idea how a customer might decide whether or
not to buy their product. The relevant concept is utility, or the units of happiness that a
customer gets from owning the product.
1

We think of consumers as wanting lots of stuff, but having to make tradeoffs in their budget
between different kinds of stuff, and tradeoffs between stuff and investments (such as an
MBA). There are different ways to represent these tradeoffsIn class we will concentrate on
the most basic way, money-metric utility. For other applications (especially Finance) you will
need slightly more sophisticated concepts of utility.

Well start with the more sophisticated notion of utility, because it will be easier to see when
you can use money-metric utility, and when its a bad idea.

A. Utility and Risk
Lets suppose that having more stuff actually does make us happy (because we act as though it
does, contrary to all psychological evidence), and lets say that we can measure that happiness
in units called utils. So, in my case, the first latte of the morning gives me 10 utils, and the
fifth latte of the morning might give me just 1 util, or maybe even negative utils.

Measuring utility in utils allows us to start thinking about risk. Lets say that your
grandmother gives you a birthday check for $300, and you know exactly what you want to
spend the money on; lets say youll get 400 utils from the stuff you will buy with the $300.
Suppose instead your grandmother had given you a check for twice as much, $600, you would
be happier, but probably not twice as happy (because the things you wanted the most urgently,
you would have bought with the $300, so the extra $300 gives you less). You might get 800
utils from the stuff you could buy from $600. If she gave you $900, you might get 1000 utils.
Your utility from different amounts of money might look like this:

1
This is obviously not a very subtle philosophical concept of happinessthe first economists (Jeremy
Bentham, John Stuart Mill) were reasonably deep philosophers, but its been downhill pretty much ever since.
One particular difficulty faced in thinking philosophically about utility is: how do I make interpersonal
comparisons? That is, when the government makes a policy (or an individual or company makes a decision)
that affects other people, how do you trade off making one person more happy with another person less
happy? (Amartya Sen is an exception to the recent trend toward shallowness in economists, and his work is
very interesting on this point.)
Firms and Markets
Lecture Notes

Lecture Notes 57

Now suppose that you had the misfortune to have an experimental economist for a
grandmother, and instead of $600, she decided to give you a gamble: you flip a coin, and she
gives you $300 if its heads, and $900 if its tails. In expected value that will give you $600. But
in terms of utils, it will give you less than 800 utils:

500u

0.5 = 0.5 (500u) + 0.5 (1000u) = 750u
Nature

1000u

This is a simple way of expressing risk: we get more utility from having $600 for sure (=800u)
than from having a risky gamble that pays $600 in expected value.

money
utils
$300 $600
800u
500u
$900
1000u
$300
$900
0.5

Lecture Notes 58

(The straight line on the graph below is the payoff to different gambles that are a mix of
getting $300 and $900; the gamble that gives $300 with probability 0.5 and $900 with
probability 0.5 is the middle of the straight line.)

B. Utility and Budgets
We use utility to think about how a consumer makes tradeoffs within a particular budget.
There are several complicating factors:
1. we buy lots of different goods
2. if you buy more of a good, you get less utility from each additional unit: for example,
suppose you are buying more and more pairs of exactly the same kind of shoes (and you
are not Imelda Marcos); its useful to have extras, but not as useful as the first pair.
This is called diminishing marginal utility (arent you glad you now know?).
Lets just stick to those two complicating factors, and think of our budget as a fixed amount of
money for a given period of time (like a month).

Lets start with a world where there are just two types of goods to buy, A and B, that come in
very small units (such as ounces of Beer). Lets say that the total utility I get from consuming a
certain amount of A and B is

Utility = (Amount of A)
0.4
(Amount of B)
0.4

So consuming 10 units of A and 20 units of B gives me 8.32 utils.

We might express different combinations of purchases in a diagram with quantities of product
A on the horizontal axis and quantities of product B on the vertical axis. In 3-D it would look
something like this, but 3-D is hard to visualize:
money
utils
$300 $600
800u
500u
$900
1000u
750u

Lecture Notes 59

Indifference curves. We want to find a 2-D way to represent our tastes for different combinations.
We do this graphically with indifference curves: lines that represent all the combinations of
the two products that we regard as equally good. For example, given the utility function
described above, we think of 2 units of product A and 1 of product B as just as good as 1 unit
of A and 2 units of B (because both give us 1.32 utils). A line connecting all such points about
which we are indifferent (i.e., that we like equally well) is called an indifference curve. The way
to think of an indifference curve is like a level curve on a topography map (there is a line on a
map showing all the points that are at 100m of altitude; here there is a line showing all the
points that give us the same number of utils.

Generally, indifference curves are downward-sloping, since we need more of a product to
compensate us for less of the other (the more is better principle). We also assume they are
curved away from the origin, since as we consider combinations with more and more of one
product, we need increasing amounts of it to keep us equally satisfied.

If we put lots of indifference curves on our graph, we can get a complete description of our
tastes. Since more is better, indifference curves that are up and to the right (farther from the
origin) represent higher levels of satisfaction.

q A
q B
IMore utility
1
1
2
2

Lecture Notes 60

Budgets. The other ingredient in our analysis is what the consumer can afford: the budget
set. With two products, the budget might be expressed in the equation:

Income q p q p
B B A A
s +

If we solve for q
A
we see this is the area below the following downward-sloping line:

B
A
B
A
A
q
p
p
p
Income
q =

This is a straight line whose position and slope depend on income and prices. If you increase
income, for example, the budget line shifts out. If you increase p
B
, the line twists clockwise
(around the vertical intercept, Income/p
A
). And if you increase p
A
, the line twists
counterclockwise (around the fixed horizontal intercept,
pB
Income
).

Demand. Putting together tastes (represented by indifference curves) and possibilities
(represented by the budget line) we can find out what our hypothetical consumer should do.
With indifference curves and budget line on the same graph, the consumers best choice is the
highest indifference curve the budget line touches. The point where they touch gives us the
quantities demanded of products A and B. Implicitly, these demands depend on tastes (these
are built into the indifference curves). They also depend on income (since a change in income
shifts the budget line and therefore leads to a change in demand) and prices (for the same
reason).

q
A
q
B
A
p
Income
B
p
Income
Increase
in income
Increase
in p
B

Lecture Notes 61

We can summarize and abstract from the underlying indifference curves and budget sets by
writing down a demand function, q
i
(p
i
), denoting the quantity demanded, q
i
, for a given price of
the good, p
i
, (and supposing that income, the prices of all other goods, and any other relevant
factors are not changing). Equivalently, we can write an inverse demand function, p
i
(q
i
),
which denotes what the price must be if the quantity demanded is q
i
.

To summarize, the analysis suggests that the quantity of a product demanded depends on:

- The tastes of the individual, expressed by indifference curves.

- The price of the product. Generally, the lower the price the higher the demand.
Depending on the curvature of the indifference curves, a change in price might have a
small or large impact on the number of units demanded.

- The price of other products. Decisions are not made in isolation. If we spend less on one
product, that necessarily leaves more to spend on others.

- Income. At higher levels of income, we can buy more of everything (and generally do).

Finance and the Risk-Return tradeoff
Why do you care about these utility curves? Good question. Mostly you care if you are going to
pursue finance, because one particular type of utility curve figures very prominently in Finance:
the risk-return tradeoff. Suppose you are buying $100 of financial assets, and (as weve
discussed) you are not too keen on risk. Then when you are buying assets, you are going to
trade off return and risk: you will only be interested in a more risky asset if it has a higher
return (i.e. if the expected payoff is $115 rather than $110). Your level curves will look like this:

q
A
q
B
Increasing utility

Lecture Notes 62

Lets say that Treasury bills are considered almost zero-risk, and that they pay $103; then I will
feel be indifferent between buying T-bills and buying an asset with a higher expected return
and higher risk, that lies on this line. This tells me how much I would have to offer this
customer, before he would buy my asset rather than buy T-bills.

C. Money-metric utility
So that was all very complicated. To keep it simple in class, we will still to money metric
utility: the happiness I get from buying a good is represented in dollars, not utils, and the
unhappiness that I feel from parting with my money is represented in dollars, as the price. So
if I see a pair of shoes that gives me $200 of utility, and it costs $150, then if I buy it I walk
away with $200 $150 = $50 of utility.

Whats limited about money-metric utility?
- Money metric utility doesnt let us express risk, the way we did in part A: obviously, if I
get $300 from my grandmother, it gives me $300 of happy, and $600 gives me twice as
much happiness. We need utils to express risk.
- Money metric utility lets us talk about budgets in a very limited way: I am sad about the
$150 I had to pay for my shoes. Why? Because thats $150 of other goods that I wont
be able to buy. Other goods are all represented in terms of the sadness of giving up
dollars in payment. This is OK if we are talking about small items in your budget (for
some people, shoes are not a small expenditure category!), but if we want to talk about
big expenditures, we want to look at the tradeoffs in more detail.

Expected return on $100
variance = risk
$103

Lecture Notes 63

Demand

Any business must have some idea how many people might buy its product at different prices.
The relevant concept is the demand curve. Stated differently, the demand curve tells us how
much people are willing to pay for a given number of units of a product. Below, we describe
properties of demand: how sensitive it is to the products price, to prices of other products, to
income, and so on. These concepts are basic tools of business analysis.

The Demand Curve
Imagine that you had 1000 customers, all with different willingness-to-pay. The first customer
is willing to pay $999 for your product, the second is willing to pay $998, the third $997, and so
on (until the 999
th
customer is willing to pay $1 and the thousandth customer only wants the
good if it is free).

You are charging one price for all, that is, you cannot charge different customers different
prices (later well explore strategies for doing just that). You can use the above information to
figure out how many customers will buy, at different prices: For instance, if you set the price at
$999, there will only be one customer. If you set the price at $990, there will be ten customers.
The way to represent this information is a demand curve. It tells you exactly how many
people will buy, at any price.

In the story we told above, each person in our market wanted only one unit, and for a different
price, but that is not always the case. Suppose we were selling pizzas: the demand curve would
Price
$674
336 units
Quantity
More formally, we can write down the
formula for this demand curve:

Q = 1000 P

(Write in the price, and it tells you how many
units people want.)
Firms and Markets
Lecture Notes

Lecture Notes 64

be a lot lower, since the maximum price would probably be around $20. Some people want
only one pizza, and some people will want several. For now, the only important thing to know
is how the total quantity demanded changes as we change the price: If our pizzas get cheaper,
some customers will order more (theyll be more willing to order an extra one and be sure not
to run out of pizza), and some customers will switch from other types of pizza.

NOTE: When we come to advanced pricing, we will care about more than just total demand
at any given price. We will want to know more details about customers: if different types of
customers behave differently; if most customers buy one unit (the car market, for instance), or
if most customers would buy different amounts based on the price (cell phone minutes, for
instance).

Price Elasticity of Demand
We presume (based on lots of evidence) that demand for a product any product! increases
if we lower price. This doesnt have to be the case, but it invariably is. The question is how
sensitive demand is to price. We measure sensitivity by the price elasticity of demand:
q
p
p
q
q
p
dp
dq
p
dp
q
dq
A
A
~ = = c .
In words, the price elasticity is the ratio between the percentage change in quantity and the
percentage change in price for a small change in price. Note that elasticity is not the same
thing as slope, although slope is an input to it; see above, where the slope is
dp
dq
. One of the
Price
$10.50
290 units
Quantity
This is the demand curve for pizzas from
our shop, for a given week. The demand
curve might be something like

Q = 500 20P

Lecture Notes 65

advantages of elasticity over slope is that it does not depend on units of measurement and thus
is more easily compared across products.

[Technical point: A useful variant of this formula uses logarithms. You might recall that

p d
q d
p
dp
q
dq
log
log
= .

(This is basic calculus; the equation is valid for any base logarithm, including natural log,
sometimes written ln.) If we do our usual approximation, replacing ds with s, we have

p
q
p d
q d
p
dp
q
dq
log
log
log
log
A
A
~ = = c .

The approximation is exact if the elasticity is constant along the demand curve. In this case,
the demand function is

log q = a + b log p

and the elasticity is = b. Given two sets of price and demand combinations for a given
product, (p
1
,q
1
) and (p
2
,q
2
), we can recover the elasticity from

1
2
1
2
log
log
log
log
log
log
p
p
q
q
p
q
p d
q d
=
A
A
~ = c
.

since log x log y =
y
x
log . For more on logarithms, see the appendix.]

Note that elasticities are generally negative, since demand declines with price. The question is
how negative. We say that products in which > 1 are elastic, meaning the quantity
demanded is sensitive to price. The higher || is, the more sensitive to price. Conversely, if
|| is small, we say that demand is inelastic, meaning that demand is relatively insensitive to
price. Note that the elasticity is defined at a point: it generally differs from one point to
another along a demand curve.

This property of demand depends on tastes (about which there is no argument!), but we can
nevertheless get some insight through examples. Demand for personal computers is elastic,
with the result that the demand for PCs has expanded dramatically as the price has fallen.
Demands for gas and energy are relatively inelastic, at least in the short run, since there is little

Lecture Notes 66

people can do easily to alter the amount they use.

Consider the demand for gasoline as an example to highlight an important point about
elasticities over time. If gasoline increased from $1.50 per gallon to $3 per gallon overnight,
most people who commuted by car would still fill their tanks in the morning and drive to
work. In the short run, they would not be willing or able to find a substitute means to avoiding
the higher gasoline prices. Thus, as we said above, gasoline would be inelastic in the short
term.

However, if a news report revealed that the government was planning to levy a 100% tax on
gasoline for the next 3 years, consumers behavior would undoubtedly change. In the next day
or week they would probably still drive their cars, but in the longer term their demand for gas
could change for many reasons. They might buy more fuel-efficient cars, carpool, or take the
bus or train to work, or work from home. As a result, the quantity of gasoline demanded at
the new price would decrease. This illustrates a general feature of elasticities: demand tends to
be more elastic in the long run than in the short run.

Another useful point is the difference between the elasticity of demand for a product category
(personal digital assistants, say) and specific products in that category (Handsprings Visor
Deluxe). Generally the demand for a specific product is more elastic than the demand for the
category as a whole. Why? Because when the price of a specific product rises, people are
willing to buy fewer units. Some of this reduction leads to purchases of other products in the
same category (a Palm IV), and part to a reduction in the category as a whole. Only the latter
shows up in the elasticity of the category as a whole, so its typically less elastic.

Elasticity and Revenue

From a firms perspective, the elasticity of demand is a critical piece of information, since it
determines the change in revenue that results from a given change in price. Recall that revenue
is simply price times quantity demanded. It is easy to see that by increasing a products price
by 15%, each unit sold will yield more money. But if overall demand for the product drops as
a result of the price increase, the positive effects of higher per unit prices will be offset by a
decline in the number of units demanded. Which effect is larger?

Formally, the percentage change in revenue induced by a (small) change in price is

p
dp
p
dp
q
p
dp
dq
p
dp
q
dq
p
dp
pq
pq d
) 1 (
) (
c + = + = + =

That is, the percent change in revenue following a price change is (1 + elasticity) x (% change
in price). Since < 0, the direct effect of the price change (the 1) and the indirect effect of
demand (the ) are opposite. If demand is elastic ( < 1), the demand effect is larger and an
increase in price reduces revenue. This is a basic point, but one that some have missed: that to
increase revenue in markets with elastic demand, you need to lower price, not raise it.

Lecture Notes 67

Cross-Price Elasticity

We have seen that demand for a product depends not only on its own price, but on prices of
other goods. When there are lots of other products its easy to lose sight of this, but its always
there. Some examples are obvious. The demand for ski boots depends on the demand for
skis: if skis get more expensive, we might expect people to buy fewer boots, too. And as we
saw above, the demand for commuter rail tickets may be influenced by the price of gasoline.

We summarize the sensitivity of demand to the price of another product with the cross price
elasticity:
Cross-price elasticity =
2
2
1
1
p
dp
q
dq
.
That is, the ratio of the percent change in demand for product 1 to the percent change in the
price for product 2.

The essential distinction here is between substitutes and complements. If the cross-price
elasticity is positive, we say that the products are substitutes. Hence Coke and Pepsi are
substitutes: If Coke gets more expensive, wed expect some people (but not all) to switch to
Pepsi. Similarly, gas and commuter rail tickets are substitutes, since an increase in the price of
gas would lead some people to switch from car to train travel. Conversely, if the cross-price
elasticity is negative, we say the products are complements. The language of business is
filled with competitive metaphors for which substitutes seem appropriate (Coke v. Pepsi). But
there are lots of examples of complements, in which a price reduction for one product
increases demand for others. Skis and boots are one example. Others include: Windows OS
and Intel microprocessors, beer and pretzels, gas and cars.

Income Elasticity

Changes in income also affect demand. Higher income generally means greater demand for all
products, but some products benefit more than others. We define the income elasticity of a
product by Income elasticity =
y
dy
q
dq
.
i.e., percent change in demand induced by one percent change in income (denoted by y).

Economists have names for products with different income elasticities. Inferior goods have
negative income elasticities. Although inferior goods arent all that common, its fun to try to
think of examples. Spam comes to mind, on the assumption that anyone with enough money
would buy something else. (You can see the source of the term.) Normal goods have
positive income elasticities. Within normal goods, those with elasticities between zero and one
are referred to as necessities, and those with elasticities greater than one as luxuries. Can you
explain why?

Lecture Notes 68

Measuring Demand

In an ideal world, firms would know the demands for their products. In practice, its not easy.
One reason is that its hard to get reliable market data: how much was bought by whom and at
what price? Another is that its inherently difficult to tease out the effect of price from the
effect of other variables, especially when the latter might be changing at the same time as price
(or, even worse, when they are not known to us).

Another approach one that is increasingly common is to do experiments in markets. Thus
catalog companies sometimes send out catalogs to different customers in which some of the
prices are different. These experiments run the risk of alienating customers (what if you find
out you got the high price?), but you can see the value to the firm of knowing the demand for
its products.

Numerical Examples

Example 1. Demand for a product is estimated to be

Price Quantity Revenue
10 6.31 63.10
11 5.63 61.90
12 5.07 60.84
13 4.61 59.87
14 4.21 58.99
15 3.88 58.18

What is the elasticity of demand at price 10? We approximate it by the change formula,

~ (q/p)(p/q) = [(5.63-6.31)/(11-10)](10/6.31) = 1.08.

This is approximate, since were using discrete changes. (In fact, the numbers were generated
by the demand curve, log q = log 100 1.2 log p, which has an elasticity of 1.2.) We could
get a better approximation by using a finer price grid.

Example 2. Village microbrew raised its price from $10 to $12 a case (wholesale). Shipment
quantities at the two prices were

Price Quantity
10 10,500
12 8,100

What is the elasticity of demand? We can approximate it by the change formula,

Lecture Notes 69

~ 77 . 1
12
100 , 8
12 10
100 , 8 500 , 10
=
=
A
A
p
q
p
q
.

This is approximate, since were using discrete changes. If we assume that the elasticity of
demand is constant then could get an exact solution by using the log formula:
= 423 . 1
) 12 log( ) 10 log(
) 8100 log( ) 10500 log(
log
log
=
=
A
A
p
q
.

Did revenue rise or fall? Since < -1, the increase in prices led to an overall fall in revenue. (If
you want to make sure then calculate the revenues before and after the price change).

If the elasticity is constant, what is the demand at $9? If the elasticity is constant then the log
formula calculates the elasticity exactly and in addition we know that:
423 . 1
9
10
log
500 , 10
log
9
=
q

where q
9
is the demand when the price is $9 per case, so (after a little bit of algebra).
199 , 12
9
10
log 423 . 1 ) 500 , 10 log( exp
9
=
|
.
|
\
|
+ = q

Example 3. Suppose the demand for product 1 is given by the demand curve,

q
1
= a b
1
p
1
+ b
2
p
2
,

which youll note also depends on the price of product 2. The price elasticity is

= (dq
1
/dp
1
) (p
1
/q
1
) = b
1
(p
1
/q
1
).

Note the demand is elastic for high values of (p
1
/q
1
) and inelastic for low values. The cross-
elasticity is
(dq
1
/dp
2
) (p
2
/q
1
) = b
2
(p
2
/q
1
).

Suppose a=500, b
1
=10, b
2
=5, and p
1
=p
2
=50. Then q
1
=250, the elasticity is 2, and the cross-
elasticity is 1. The positive sign of the cross-elasticity means the products are substitutes.

Consumer surplus is the area under the demand curve but above the price. Since demand is
linear, this is a triangle. With the same numbers, the consumer surplus is
250 x (7550)/2=31,250.

Lecture Notes 70

Example 4. The table below gives the own and cross-price elasticities for selected automobile
models.

Model 323 Cavalier Accord Taurus Century BMW
Mazda 323 -6.4 0.6 0.2 0.1 0.0 0.0
Cavalier 0.0 -6.4 0.2 0.1 0.1 0.0
Accord 0.0 0.1 -4.8 0.1 0.0 0.0
Taurus 0.0 0.1 0.2 -4.2 0.0 0.0
Century 0.0 0.1 0.2 0.1 -6.8 0.0
BMW 735i 0.0 0.0 0.0 0.0 0.0 -3.5
(Source: BLP, 1990 data.)

Why are the own elasticities so high? Answer: These are models for which many substitute
models are available. Thus, even if the demand for cars is not very elastic, the demand for a
particular model is.

Are the Accord and Taurus complements or substitutes? Answer: The cross-price elasticity of
the Accord with respect to the price of the Taurus is given by 0.1, a positive value. The two
models are therefore substitutes. In fact, no two models in this sample are complements.

What are the Tauruss closest competitors? Answer: Looking at the Taurus row, we see that
the cross-price elasticity is highest for the Accord. In other words, a 1% change in the price of
the Accord would have a greater impact than a 1% change in the price of any other model
(other than the Taurus).

If GM lowers the price of its Chevy Cavalier, does it cannibalize its Buick Century? Answer:
The cross price elasticity of the Century with respect to the price of the Cavalier is negligible.
Therefore, a decrease in the price of the Cavalier will reduce the demand for the Century by
only a tiny amount.

Why is the direct elasticity for the Mazda not lower than the elasticity for more expensive
models (as the rule of thumb would suggest)? Answer: As suggested by the qualitative analysis
of demand elasticity, luxuries tend to have higher elasticity than non-luxuries. However,
another rule of thumb to keep in mind is that the elasticity for a particular product is always
higher than the elasticity for the group of products it belongs to. As it happens, there are many
more compact car models than there are luxury car models. Therefore, even though the
elasticity for luxury cars is higher than the elasticity for compact cars, the elasticity for a
particular luxury model may not be much greater than the elasticity for a particular compact.

Suppose Honda sold 300,000 Accords in 2001. In 2002, the price of the Accord decreased by
2%, whereas the price of the Taurus decreased by 3%. What is the likely change in Accord
sales? Answer: The percent change in demand is approximately given by (-2%)*(-4.8) + (-
3%)*(0.1) = 9.3%. We would expect an increase in Accord sales of approximately 9.3%, or
.093*300,000 = 27,900 units.

Response to change in price of:

Lecture Notes 71

Appendix: Logarithms

Why logarithms? They just turn out to be useful in lots of places: elasticities, compound interest,
growth rates, and so on.

What are logarithms? The natural logarithm of a number x comes from the power of a number e,
which is approximately equal to 2.718. If x = e
y
, then y is the logarithm of x. We write
y = log x.

There are logarithms based on powers of other numbers (any positive number, not just e).
Well stick with e. In Excel, the natural logarithm of x is written ln(x).

Suppose, instead, you know the logarithm y of x. How do you find x? From the definition, x
= e
y
. In Excel, this is written exp(y).

How do they work? The most useful properties of logarithms are

log (xy) = log x + log y
y
x
log = log x log y
log (x
y
) = y log x.

In addition, the derivative of the natural log of x is the inverse of x:

x dx
x d 1 log
= .

Can you give us an example? We showed how to compute elasticities earlier. Another application
is compound growth. GDP per capita in Korea was $1000 in 1960, $6000 in 1990. (These are
rough estimates, measured in 1990 US$ i.e., corrected for inflation). What was the average
annual growth rate? Were looking for a number g satisfying
6000 = (1+g)
30
1000

How do we find g? One way is to use logarithms. Note that

log
000 , 1
000 , 6
= 30 log (1+g)

Since log 6 = 1.792, log (1+g) =
30
792 . 1
= 0.0597, and 1+g = exp(0.0597) = 1.062. The growth
rate was 6.2% a year, which is extraordinarily high. Similar calculations show up in present
value calculations.

Lecture Notes 72

Economic Profit and Costs

Suppose you need to decide whether to start up a plant or shut it down, whether to eliminate a
product line, whether to produce at capacity or below. The answers to these questions depend
on costs. But which costs matter? Cash flow measures of cost exclude some costs that matter
(opportunity costs) and include others that do not (sunk costs). We refer to the costs that
matter as economic costs, where economic in this context means the ones that matter for
the purposes of decision making.

More formally, lets say you are making a decision, and you only have two choicescall them
A and B. Using a decision tree is a way to think about the tradeoffs clearly:

A Revenues(A) Expenditure(A)

You

B Revenues(B) Expenditure(B)

In making the choice between A and B, you care only about the difference between those two
payoffs. This leads to our definition of Economic profit:

Economic profit of choosing A instead of B = top line minus bottom line
= [Total balance after choosing A] [Total balance after choosing B]
You can think of your total balance as how much money youd have in bank or in
investments at the end of the year. Notice that the first thing that drops out of the
comparison (we didnt even bother to write it down) is the amount you made on other projects
during the year, and any other money you have in the bank. Those amounts will cancel out
when we take the difference. That leads to two key points:

1. Sunk costs do not matter (i.e. sunk costs are not a part of economic costs).
Sunk costs are costs that you have already incurred, and can never retrieve. Lets suppose for
instance that you spend $500,000 on advertising a new product line earlier this (accounting
year); that amount is sunkyou cannot get the money back from the advertising firm, even
if you discontinue the product line.
2
Another category of sunk costs (sometimes called fixed
costs) are costs that you have not yet incurred, but that you cant avoid paying: your payroll
for the upcoming month, for example (if none of your employees are on flexible contracts).

2
If you spend $500,000 on an asset (a building, a piece of equipment), and you can only re-sell it for $80,000,
then $420,000 of that expense is sunk.
Firms and Markets
Lecture Notes

Lecture Notes 73

You are trying to decide whether to discontinue the product line or to keep it going. The
economic profit of that decision will not depend on the sunk costs:

keep Revenues(keep) Cost(keep) 500,000

You

discontinue 0 500,000

When we look at the difference between these two payoffs, the 500,000 will cancel out.

Say the British Government was trying to decide whether to continue rail service between two
cities or instead to change to bus service. The revenues gained by both choices are 120. The
costs of the rail company are 30 for interest on bonds used to finance the rails, 50 to lease
trains, and 50 for labor (or labour, since were talking about the UK). The costs of running the
bus service are 60 to lease the buses and 50 for labor. At first glance the government is losing
10 on the Rail Company and would make a profit on the bus company, so it would make sense
to switch to buses. The fact is that the British Government is not going to default on the rail
bonds, so the interest has to be paid regardless of which option is chosen. The interest cost is
a sunk cost, then, and the rail company is not only profitable in an economic sense, but more
profitable than the bus company. That is, economic profit = 120 (30 + 50 + 50 30) = 20
for the rail company, and economic profit = 120 (60 + 50) = 10 for the bus company. This
may not seem to make sense, but it does in an economic sense since you would have to pay the
cost of the interest anyway. If you could default on the railway bonds, it would be a different
story and it would be worthwhile to do so and change to buses.

Further examples:

- You have spent $15 on a ticket to a football game. Come Sunday, it is raining cats and
dogs and there is no way for you to sell the ticket. You should make your decision on how
much you value watching the game versus how much you hate the rain. The price of the
ticket should not come into the equation since it is a sunk cost. Regardless of the option
chosen you will have spent $15 on the ticket.

- A company has built a dam. There was a huge initial investment in the dam that comes out
to 5 per unit of energy when amortized over the life of the dam. The operating costs of
the dam add 5 to the unit energy cost so that the total cost of power from the dam is 10.
Once the dam is completed, a new source of energy is discovered so that the price of
energy falls to 7. Should the dam continue to operate since it will be losing 3 on every unit
of electricity generated? The answer is that the dam should continue operating as long as
the price is expected to average over 5 in the short to medium term (that is, it might make
sense for the dam to absorb short term losses if there is a cost for shutting it down and
reopening it). The cost of building the dam does not come into the equation since it
cannot be recovered if the dam stops producing electricity. Thus it makes economic sense
to continue to operate the dam as long as the operating costs are covered.

Lecture Notes 74

- Say you have purchased a laptop computer for $2,000. A week later a new, improved
model is introduced for $1000 that is 3x as fast as your laptop. The company offers to sell
you the new computer for $400 if you trade in your old laptop. This seems to be paying
$2,400 for a computer that costs $1000; should you do it? In order to decide whether to
take the trade in, you need to place a value on the extra speed of the new laptop. You then
need to ignore the $2,000 as a sunk cost. The only thing important to you is if the extra
speed is worth $400 or more to you, in which case you should trade the laptop in.

- Airbus is a jointly owned consortium that makes commercial aircraft. The four European
governments that own the consortium invested huge sums in subsidies to develop the
company as a competitor to U.S. aircraft manufacturers. Currently the company is
profitable, but it would have to be enormously more profitable in order to recoup the
amount spent on subsidies. If there had been loans made instead of subsidies, the
company would not be profitable. What should the governments of the countries do if
there had been loans to pay off instead of subsidies? Provided that they are not willing to
default, the company should continue to produce aircraft even if loan payments made the
company permanently unprofitable. This is true provided there is enough cash to stay
solvent (but it would make sense for the governments to inject more cash or to assume the
loans anyway). As long as Airbus is covering its operating costs, it should ignore the sunk
costs of the hypothetical loans and keep churning out aircraft.

- Haas (the business school at Berkeley) installed swing Macs in their classrooms for
academic purposes due to the fact that they could run both Macintosh and IBM software.
They proved unable to run this software well, however, and this became an impediment to
learning as well as an embarrassment to the staff. A proposal was brought up in a meeting
to replace these machines (which cost $5,000) with $1,000, functional PCs. A faculty
member said that this would be fiscally irresponsible since the school had spent so much
on the swing Macs in the first place. The economics faculty at this point chimed in that it
made no sense not to replace the Macs since they were a sunk cost.

2. Opportunity costs matter
According to this decision tree, choices A and B are mutually exclusive. (If they are not
exclusive, then the tree should have a third branch, entitled A+B.) That means that one
cost of choosing A is that you cannot choose Byou give up on the opportunity to do B. This
is called an opportunity cost.

Calculating economic profit using a decision tree means that you take opportunity cost into
account. Lets say that Stern has space in another building off of Washington Square which it
could use to run an extra section of Langone, or it could rent the building out for office spaace
Run extra section

Stern

Rent space out

Lecture Notes 75

Economic profit of running an extra section
= [Total balance after running extra section] [Total balance after renting space]

The profits from renting the space appear as a cost when we think of the economic profit of
running an extra section, because they are subtracted at the end. Why? Stern only wants to run
an extra section if that branch has a higher payoff, that is, if the economic profit of running it
is positive. Otherwise, Stern should rent the space out.

*** Note on opportunity costs *** Beware of double-counting!! If you are not using a decision
tree, but just using the intuition derived from the trees, you just go ahead and subtract them at
the end of your profit calculation. But if you are using a decision tree, dont subtract them
along one branch of the decision tree!! The decision tree is already calculating opportunity
costs for you

Examples of Opportunity Costs

An example of opportunity costs is going to the movies. Say you are considering going to the
cinema on a Friday night and a ticket costs $10. If your alternative to going to the cinema is to
sit at home and do nothing, then there are no opportunity costs (presuming doing nothing
holds no value to you). Thus the economic cost of going to the movies is $10. Let us say,
however, that you have a part time job and you have been offered a four-hour shift delivering
pizza for $10 an hour. Assuming that you do not mind the pizza delivery job (that is, not
delivering pizza does not have a value for you on that particular Friday), then the opportunity
costs of going to the movie is the money you are foregoing by not working (i.e. $40).
Consequently the economic costs of going to the movie in the second case would be $50.
Note that you may have the opportunity to baby-sit for those same four hours for $8, but we
use the $10 for the pizza delivery job since that is the best alternative use of your time.

Another example of opportunity costs is in the use of Carnegie Hall. Say you are managing the
facility and the agent of a concert pianist comes to you and suggests holding a recital. You
expect that the net revenue for the recital would be 100. The alternative to holding the concert
would be holding a rock concert, which would net 150. The impresario may try to convince
you that you would be gaining 100 from this recital, but actually you would be losing 50 (net
economic revenue of 50) since you need to add in the opportunity costs to get to the
economic costs. Say you had total revenues of 900 and expenses of 800 for the recital. In that
case you would calculate the net economic revenue as 900 (800 [actual expenditures] 150
[opportunity costs]). Alternatively you can make decisions by comparing the net benefits (i.e.
150 is better than 100).

A third example would be using a part of an NYU building to house short-term visitors such
as executive education students. Say that this use brought in 80 in room fees but required 30 in
cleaning and other expenses. The next best use of the space is to lease it out for office space at
60. You can look at this in the same two ways as before. In net benefit terms the 80 30
would lead to a benefit of 50, which is less than 60. The right decision would be to rent out
the space for offices. Similarly the net economic revenue of the student housing is 80

Lecture Notes 76

[revenues] (30 [cleaning, etc.] + 60 [opportunity costs]) = -10. This is less than zero, and
thus should be rejected as an option. Thus, either way you look at it, renting out offices is the
way to go. It would be different if there were other benefits to having housing for executive
education students that were not being considered.

All these examples illustrate a common feature of opportunity costs: its not easy to quantify
them. We have to put a value on an alternative use, and such values inherently involve some
judgment. Perhaps thats why financial statements tend to leave them out.

Economic Profits and Cash Flows

Economic profits are used to make decisions. They tell you whether choice A or B generates
higher profits and how much higher those profits will be. They do not replace accounting
measures, since it is quite possible to have the amortization of sunk costs bankrupt the firm
even though it makes economic sense to keep the firm producing. Similarly, often cash flows
are not useful for making economic decisions. The costs of choice A are defined only in
relation to the choice B that it is being compared to. You can look at the costs and benefits in
different ways, but they give the same answer.

Further differences between accounting and economic profits: Economics ignores the past if
you cannot do anything about it, but cash flow statements look backwards. Opportunity costs
are not reflected in cash flows, but they do affect decisions (if theyre made correctly).
Conversely sunk costs involve cash flows, but dont affect economic profits: They dont
matter. Economic profits are not the same as cash flows.

Examples

Example 1 (health club). If you own a health club that makes $3,800 in revenue, should you stay
open? The following is the cost structure of the health club: Actual expenses of $2,500 for
labor and $2,000 for lease payments. The lease is unbreakable and bankruptcy is not possible.
There is a possibility to sublet the space, however, for $1,200 a month. Thus the $2,000 lease
payment is a sunk cost, and the $1,200 is the opportunity costs for keeping the health club
open.

Answer. With this cost structure, one is losing $700 a month on an accounting basis. On an
economic basis, however, the decision tree is:

3,800 - 2,500 - 2,000

You

1,200 - 2,000

- 2,000

stay
open
close & lease
close & do
nothing

Lecture Notes 77

(Do nothing is an option people often forget! In this case its obviously worse, so we ignore
it.) Economic profits of staying open relative to the next best option (leasing) are $3,800 -
$2,500 - $1,200 (opportunity cost) = $100 a month. In this case it makes sense to keep the
health club open.

What if bankruptcy were an option? In that case, the obligation to pay the lease goes away
when you declare bankruptcy, so you may not have to pay it.

3,800 - 2,500 - 2,000

You

1,200 - 2,000

0

Now, economic profits of staying open relative to the next best option (declaring bankruptcy)
are $3,800 - $2,500 - $2,000 = - $700 a month. In this case it makes sense to close and declare
bankruptcy (if there are no other businesses that are affected by bankruptcythat would
require a more complicated tree!).

Example 2 (Redsyke quarry, real example from UK). There was an expansion of the M6 motorway
that required the extension of a quarry, including an 8-hectare plot used for grazing. The
quarry offered the farmer approximately 3x the market value of the land. The farmers lawyer
suggested that the farmer ask for compensation for foregone earnings based on the value from
the cows and sheep that would be displaced. What would you give as testimony if you were
called as an expert witness (your testimony could vary depending on which side you were
supporting)?

Answer. The question is what the opportunity cost is to the farmer. If the farmer could rent
adjacent land, then 3x market value seems high. Otherwise the profit on cows and sheep is a
legitimate opportunity cost. Note: the farmer lost.

Why all this effort to belabor what may seem obvious? The history of business is filled with
huge mistakes from an incomplete understanding of the nature of costs. Opportunity costs
involve no payment of money and are therefore easy to overlook. Large sunk costs often
come with political supporters, whose reputation hinges on the success or failure of the
project. But from a purely economic point of view, theres nothing you can do about them:
theyre gone.

stay
open
close &
lease
close &
declare
bankruptcy

Lecture Notes 78

Perfect Competition

When talking about pricing, we consider the monopoly benchmark. What happens when,
instead of monopoly, we are at the opposite extremea competitive environment? In
particular, how are prices determined in such markets? How can we assess the impact of a
given exogenous event on market prices?

We have thought about the demand and supply before but having a better understanding now
of the underlying individual consumer decisions that feed into demand and the underlying
individual firm decisions that contribute to supply, we can understand that classic diagram, the
circumstances under which it is appropriate, and its implications.

Further, the competition benchmark will allow us to establish an important result, namely that
perfectly competitive markets are efficient (in a sense to be specified). This provides a
rationale for anti-trust policy and promotion of competition, a topic that will appear frequently
in the rest of the course.

Perfect Competition

The perfectly competitive industry, an industry with no barriers to competition, is a useful
benchmark. It is also a reasonable approximation to many important industries, including
many sectors of agriculture, some sectors of labor markets and reasonably heavily traded
financial securities. Consider the competitive forces in this situation:

- Atomistic firms. Many firms, all small relative to the market and unable to affect the
market price through their actions. Moreover, the minimum efficient scale is small
relative to the size of the market.
- Homogeneous product. Competitors produce exactly the same product (and therefore
compete head to head on price, the only remaining variable).
- Perfect information about price and quality. Everyone knows its the same.
- Free entry and access to technology. Imitation is possible: others can enter the
business if its profitable with (eventually) the same costs as incumbents.

Under these conditions, which we term perfect competition, wed expect intense
competition.

How reasonable is this? More so in some industries than in others. There are industries where
differences in profit rates across companies can last for decades. Although the source of these
differences remains somewhat of a mystery, they appear to be related to industry
characteristics, scale economies, market share, and research and development intensity.
Related work suggests that one of the more persistent sources of high performance is
organizational capital, the processes that underlies the organization of the firm. Thus
Firms and Markets
Lecture Notes

Lecture Notes 79

Toyotas production system has proven very hard to imitate. Similarly, existing airlines have
found it difficult to reproduce the low-cost, high-value delivery system used by Southwest
(although new entrants seem to find it easier).

This perfect competition benchmark is useful not only in itself, but also as a comparison to
consider violations of these conditions. With one or few firms, we have a monopoly or an
oligopoly (which leads to complex strategic interactions that well study later in the course).
With differentiated products, we moderate the impact of competition (a recurring theme) and
open up the issue of product positioning. With imperfect information, we create the
possibility of branding (a brand often being a signal of quality). With entry and access to
technology, we raise the issues of scale economies, patent protection and intellectual property,
entry preemption, and network effects (which we discuss later). Stated differently: if a firm has
a sustainable competitive advantage, it must lie in a violation of one of these conditions.

Firm supply and market supply in competitive markets

Lets look more closely at a market or industry under conditions of perfect competition. First,
each firm faces effectively a flat (infinitely elastic) demand curve: it can sell all it wants at the
market price and can sell nothing at any higher price. Why? Because each firm is small relative
to the whole market and has no impact on market price. People sometime refer to the market
as atomistic, since each firm is small like an atom. In these circumstances a firms total
revenue is TR(q) = pq, hence the firms marginal revenue is p. The usual profit-maximizing
condition that marginal cost equals marginal revenue simplifies to

MC(q) = p.

Thus under these conditions, each firms supply decision is governed by its marginal cost
curve. It supplies the quantity at which price equals marginal cost.

Each firms supply curve gives the quantity that the firm would want to supply at a given price.
Specifically, firm i would want to supply q
i
such that MC(q
i
) = p. If we put all firms together,
we sum their supply at each price to come up with the industry supply curve. Specifically, if we
add up each firms value of q
i
, then we obtain the industry supply curve S(p) = q
1
++q
n
.

Consider the specific example of an industry with three firms, each with constant marginal
cost. Firm 1 has a capacity of 200 at MC = 5. Firm 2 has a capacity of 100 at MC = 8. Firm 3
has a capacity of 100 at MC = 10. If the three firms are competitive (this is a question well
want to come back to), what is the industry supply curve? How much does each firm produce?

The hard part is finding the industry supply curve is getting reliable cost data. Once we have
cost data (or at least estimates), we basically graph the demand and supply curves and find the
point where they cross. The supply curve is flat at p = 5 for 200 units, then jumps up to p = 8
for 100 units, and up to p = 10 for another 100 units.

Lecture Notes 80

Market Equilibrium

Whether considering long run or short run outcomes, the price of a product is the result (in
market settings) of the interaction between buyers (demand) and sellers (supply). We refer to
the point at which the supply and demand curves intersect as equilibrium. as we have
discussed in the earlier note on supply and demand.

Short-run and long-run equilibrium

We define the short-run as the period when the number of firms is fixed, whereas, in the long
run, we also consider the possibility of entry and exit.

Consider an industry with a given number of firms (short run) and an aggregate supply curve
as described above. The intersection of supply and demand determines the equilibrium price.
Depending on the number of firms in the market, this may result in a short-run equilibrium in
which price is above average cost (meaning firms are making above-normal profits), below
(firms are losing money), or equal. But firms are doing their best, given their prior choice to
enter the market.

The next stage of competition concerns the decisions of these firms, plus potential entrants, to
stay, enter, or leave the industry. Suppose price is greater than average cost for all firms in the
industry. Then we would expect incumbent firms to expand their capacity and new firms to
enter. As a result, the market supply curve shifts to the right, driving down the price. This
may take some time, but the ultimate effect is to drive firm profits down to zero (by which we
mean normal levels). Conversely, if price is below average cost, then some firms will leave the
industry, the supply curve will shift to the left, and firms will eventually attain a normal level of
profit.

Specifically, lets think about how this might work in a competitive industry in which all firms have
the same cost structure. The market price might allow firms to make money or not, depending on
the relation between p = MC and AC. If market price p is below average cost AC, then the
firm is not making enough to cover its fixed costs. When it gets a chance, it might choose to
leave the industry. This would tend to reduce supply and increase the market price.
Conversely, if at the market price p is above average cost AC, then the firm is making money
and would likely choose to stay. If other potential entrants face similar costs, they might
choose to enter, too. This would tend to increase supply and reduce the market price.

Eventually, wed expect this process of entry and exit to drive profits and losses out of
industries and leave firms with p = MC = AC. At this point, note that the industry is operating
at minimum cost.

It is important and interesting to highlight that if firms have different cost structures then in a
long run competitive equilibrium (with many price-taking who are small relative to the overall
market), the marginal firm will be one for which p = MC = AC, however some firms lucky
enough to have cost advantages will be earning rents. To earn such rents these factors need to
be impossible or hard to replicate. For example, for many minerals mining is a commodity
industry but some mines are lucky enough to find seams located near the surface while others

Lecture Notes 81

incur greater costs of extraction. The marginal mine will be earning no profit, but some mines
may be earning profits, which are termed rents on their superior location. The challenge for
managers is to create or develop products, practices or skills on which they can earn such rents.

Let us go back to the example above. Suppose market demand is Q = 540 20 p. At p = 10,
demand is Q = 540 200 = 340. Then Firm 1 produces q = 200, Firm 2 q = 100, and Firm 3
produces q = 40. Since the supply curve is such that price equals 10 for any output level
between 300 and 400, we confirm that equilibrium is given by p = 10 and Q = 340.

Is this equilibrium a long-run equilibrium? If the technology available to potential entrants is
like firm 3s, then the answer is yes: an additional entrant would not be able to make positive
profits. Notice that, in this equilibrium, while firm 3 earns zero profit (it is the marginal firm),
both firms 1 and 2 earn positive profits (rents from a superior cost function).

Other interesting questions we may ask with this framework: If demand falls to D(p) = 400
20p, what happens to the profits of Firm 1? (They fall, since the price falls.) What happens to
the profits of Firms 1 and 2 if Firm 3 reduces its MC to 8? (They both fall.)

The Fundamental Theorem and The Invisible Hand

Why do economists (and sometimes even policy makers) wax lyrical about markets? One
reason (and there are other perhaps more compelling ones) is that competitive markets are
efficient, meaning that in a competitive market the equilibrium level of output and prices are
such that they lead to the largest level of total surplus. To economists this is a sufficiently
important and striking result that it has won the (somewhat grandiose) designation The
Fundamental Theorem.

Adam Smiths famous invisible hand refers to the role that prices play in achieving a good
allocation of resources in the economy. Though buyers and sellers may have disparate and
conflicting preferences and capabilities, the price and market mechanism ensures that
consumers who value goods most receive them, and those firms for whom it is cheapest to
produce the goods produce them. Moreover, in the equilibrium of a perfectly competitive
market there are no remaining consumers whose willingness to buy one additional unit is
greater than what it would cost any firm in the economy to produce it. That is, all trades such
that willingness to pay is higher than marginal cost take place.

This no doubt is a good thing; however, a couple of observations are worth highlighting. First,
the Fundamental Theorem is a statement about efficiency rather than equity, that is, if concerns
the size of total surplus, not its distribution. In particular, note that in calculating total surplus
we value consumer and producer surplus equally. We presumably attach value to profits
because they are eventually returned to shareholders of the firm. However, in as much as
shareholders tend to be wealthier as a group than consumers, some would argue that firms
profits ought not to be weighted as much as consumer benefits. Institutional arrangements or
regulations that, for example, raised consumer surplus by the equivalent of $20 million and
decreased firms profits by $25 million may be judged to improve welfare even as a cost of $5
million in terms of total surplus.

Lecture Notes 82

Another important caveat is that the Fundamental Theorem applies to competitive markets,
which, as outlined at the beginning of this note, correspond to some fairly strong assumptions.
Where producers or consumers are large enough to affect market prices, products are
differentiated, there is less than perfect information about price and quality or entry into the
industry is restricted, then the Fundamental Theorem does not necessarily hold. Further, an
important resultthe Theory of the Second Bestsuggests that if there are some frictions
in a market (if some of the assumptions for the Fundamental Theorem fail), then reducing
others may well reduce welfare. Relaxing the assumptions of an idealized competitive market
and thinking through their implications will keep us busy through the rest of this course.
Another important assumption that arises in the context of the Fundamental Theorem is an
implicit assumption that a consumers enjoyment of a good does not depend on other peoples
consumptiontechnically speaking, the assumption that there are no externalities. We shall
return to consider relaxing this assumption in our discussion of networks.

Summary

- In a competitive marketa useful benchmark and approximationfirms are atomistic
and act as price takers, competitors produce a homogeneous product about which
there is perfect information, and there is free entry into the industry and access to
technology.
- For a competitive firm (one that cannot directly influence price), the (short-run) supply
curve is the MC curve. Short-run here means that firms have already committed to pay
whatever fixed costs are required to enter the industry.
- For a competitive industry, the short-run supply curve is the sum of the supply curves
of the individual producers.
- In the short run, the price is the intersection of demand and the short-run industry
supply. In the long run, firms enter profitable industries and exit unprofitable ones.
The interesting result is that price is governed by the minimum average cost.
- In a competitive industry (and under the strong requisite assumptions) the
Fundamental Theorem applies: Efficiency obtains in the sense that the total of
producer and consumer surplus is maximized.

Lecture Notes 83

Pricing

Most firms have some control over the price they set. Although they may have competitors,
they can charge a higher or lower price, and generate less or more demand as a result. The
question is how high to set the price. A high price generates more revenue per unit, but fewer
units are sold. A low price generates less revenue per unit, but more units are sold. Which is
best? It depends on the elasticity of demand and marginal cost.

Pricing: intuition

To see this most clearly, consider a monopoly: an industry with a single producer. You can set
your price however you like, but you sell only what the market demands at each price. You can
either price low and sell a lot, or price high and sell a little. The key insight is that for a
monopolist to sell an additional unit, it must lower the price on that unit but also on all other
units it sells. Thus the overall benefits are less than the price it gets for the extra unit.

Consider an example. Suppose you sell 80 CDs for $15.10, and to sell 81 you must reduce the
price to $15. Is the marginal revenue $15? No! Your revenue rises from 1208 (=15.10x80) to
1215 (=15x81), an increase of only $7. This is less than the $15 price, because we had to
reduce the price on the first 80 CDs by 10 cents (leading to 0.10x80=$8, the difference). Thus
the marginal revenue of an extra unit is less than the sales price.

Now how much should we produce? We produce until the extra benefit of an additional unit
is balanced by the extra cost: when marginal revenue equals marginal cost, or simply MR =
MC. As we will see, for a competitive, price taking firm, the demand curve is flat, and thus
marginal revenue is equal to price. It follows that profit maximization implies p = MC. Since
MR is less than price, and marginal cost is level or increasing, this results in less output under
monopoly.

Another way to see this is to recall that competitive firms are price takers: they are too small
to affect the market price through their own actions. A competitive firm cant raise its price
above the industry price, because it would then lose all its customers to competitors. Generally,
a monopoly will set a higher price, and sell fewer units, than competitive firms in the same
industry. The same analysis applies to any firm that has some control over price. In markets
with differentiated products, for example, each firm might be considered a monopolist for its
own version of the product (although the existence of close substitutes will affect its demand
curve).

Pricing: calculus

We can see the same thing mathematically. If the inverse demand curve (price as a function of
quantity) is p = D(q), then revenue (expressed as a function of output, q) is
Firms and Markets
Lecture Notes

Lecture Notes 84

R(q) = pq = D(q)q.

Marginal revenue (the extra revenue from an additional unit) is

MR(q) =
dq
q dR ) (
= D(q) + D(q)q = p + D(q)q < p,

where D(q) denotes derivative of D with respect to q. The last step follows from our standard
assumption that demand declines with output (hence D<0).

The optimal output level can be found by maximizing profit,

Profit(q) = R(q) C(q),

which we do by setting the derivative equal to zero and then verifying that we have a
maximum. The derivative equated to zero leads to

dq
q d ) ( Profit
= MR(q) MC(q) = 0,

or simply MR(q) = MC(q). In words, the profit maximizing output level is such that marginal
revenue equals marginal cost.

Technical Aside: To check that we have a maximum, we look at the second derivative:
2
2
) ( Profit
dq
q d
= MR(q) MC(q).

If this is negative, MR = MC gives us a max. Its sufficient that MC increase with q (MC > 0)
and MR decrease with q (MR < 0). For most of the problems well consider, well assume MR
< 0 and MC 0, which is enough to cover us without ruling out constant MC.

Elasticity rules

The math is useful for some people because it is clear and concise. It is also useful for
demonstrating an interesting relationship between seller margin and elasticity. Define the
margin as m =
p
MC p
. For a monopolist, the margin is generally positive. But how large is
it? The answer is m =
c
1
, where is the price elasticity of demand (and is negative). It
implies |
.
|
\
|
+
c
1
1 p = MC.

Lecture Notes 85

We will refer to both as versions of the elasticity rule. Note that the margin depends on how
sensitive demand is to price. If it is very sensitive (|| large), then the margin is small. Even
for a monopolist it is hard to squeeze out a large margin, because customers will simply decline
to buy the product. But if the elasticity is small, the margin can be quite large. (There is a limit
of one, for reasons related to the fine points above).

Where did this rule come from? Recall that

MR = p + D(q)q = MC.

Then ,
p
q
q D
p
MC p
) ( ' =
.

Except for the minus sign, the expression on the right hand side is just the inverse of the
elasticity of demand.

Examples. Local telephone service is a must for most people, so demand is inelastic (||
small). An unfettered monopolist would presumably choose a high markup and price. The
classic role of telecomm regulation, therefore, is to keep the local operator from exploiting its
position. Similarly, demand for prescription drugs is inelastic, both because theyre necessities
(how often have you said: Sorry, that drugs too expensive, Id rather be sick?) and because
in most cases the customer isnt paying anyway.

Conversely, a competitive firm might be regarded as one facing an infinitely elastic demand
curve (horizontal), which produces a markup of zero. Thus the key issue for a monopoly is
how elastic its demand is. If it is very elastic, the cost of monopoly is low.

Numerical examples

Example 1. Consider a monopoly facing linear demand and constant marginal cost. Demand is
q = a bp,

with a=12 and b=2. Costs are C(q)=cq, with c=1, so MC=c=1. What is the
optimal price for the monopolist?

Answer. The inverse demand curve (we need to solve for p) is
b
q a
p

= .
Then profits are

cq q
b
q
b
a
cq q
b
q a
cq pq q =
= =
2
1
) ( Profit .

To maximize, we differentiate with respect to q and set the result equal to zero, implying
MR=MC:

Lecture Notes 86

c q
b b
a
=
2
.

Optimal output is therefore
2
cb a
q

= , substituting in a=12, b=2 and c=1 yields that the
optimal output it 5, price is 5 . 3
2
=
+
=
=
b
bc a
b
q a
p , and profit is 12.5.
As for the elasticity rule: At the optimal price, the elasticity is

5
7
5
5 . 3
2 = = = =
q
p
b
q
p
dp
dq
c .

Thus the markup is
c
1
7
5
5 . 3
5 . 2
= = =
=
p
MC p
m , as promised.

If c increases by 1, then substituting in a=12, b=2 and c=2 yields that the optimal output is 4,
the optimal price is 4 and profit is 8.

Example 2. Consider a similar problem with a log-linear demand curve. This kind of demand
curve is uglier mathematically, but is convenient because it has constant elasticity of demand.
In practice, its generally a much better approximation than the linear demand curves we often
use as examples. Lets say that the inverse demand is

log p = o + | log q,

or

p = oq
|
.

(Why inverse? Because we solved for p, rather than q.) The elasticity of demand is c=1/|.
If marginal cost is a constant c, then profit is

Profit(q) = pq cq = oq
|+1
cq.

To find the optimal quantity:

0 ) 1 (
) ( Profit
= + = c q
dq
q d
|
o | .

If you recall that oq
|
= p, this turns into

c p = + )
1
1 (
c
,

Lecture Notes 87

which is equivalent to the elasticity rule.

Example 3. A firm sells one million units at a price of $100 each. The firm's marginal cost is
constant at $40, and its average cost (at the output level of one million units) is $90. The firm
estimates that its elasticity of demand is constant at 2.

Should the firm raise price, lower price, or leave price unchanged? Optimal pricing implies m
=
c
1
, where m=
p
MC p
is the margin. In this problem, we have
100
40 100
=
=
p
MC p
m
or 0.6, which is greater than 5 . 0
2
1 1
= =
c
. This tells us that the price/cost margin is too
high, so a lower price (for practice you can check that it would be $80) would be optimal.
Note that the margin depends on MC, not AC.

Lecture Notes 88

Advanced Pricing

Consider a situation faced by a stylized biotech firm:

Ken, its brilliant science, but I just dont see how we can make any money from it.
CEO Allison Schreter of biotech startup BioStern is discussing strategy with CFO
Charles Miller and Research Director Ken Goldman. It will cost at least $20m to
complete the research and get regulatory approval. If we charge $10/dose to meet the
US market, we can't expect to recoup more than $15m. But at that price no one
outside the US can afford it. If we charge a lower price, we can attract the overseas
markets, but our overall profit picture is even worse.

Ken is crestfallen. Two years ago, he left a tenured research position at NYU to pursue
an idea. Whatever its commercial merits, he felt his team was doing pathbreaking
science, and he hoped that commercial success would make further breakthroughs
possible.

Charles, meanwhile, was doodling on his Palm V, apparently oblivious to the
discussion. Suddenly he stood up. I have an idea, he said. What if we charge a
high price in the US, where people can afford it, and a lower price elsewhere? If we do
it right, we might be able to make enough money to cover our development costs and
keep the firm going for another five to ten years.

Allison wondered. If Charles was right, there was hope. But she needed to base her
decision on more than hope.

We won't be quitting our jobs any time soon to write screenplays. But you get the picture:
there may be situations in which you can increase profits by charging different prices in
different markets, or more in general using a pricing strategy that may be more sophisticated
than just posting one price.
We refer to any scheme that induce different customers to pay different prices as price
discrimination. In practice, price discrimination is connected to a number of related decisions,
including product versioning and bundling.

Price Discrimination

Our analysis to date has been based on uniform pricing: a firm charges the same price for a
homogeneous good to all buyers. But it is not hard to think of examples of non-uniform
pricing or price discrimination:
- Student, child, and senior discounts. Kids, for example, pay less at New York movie
theaters, although they use the same size seat as anyone else.
Firms and Markets
Lecture Notes

Lecture Notes 89

- Airfares. Business customers pay far more for seats, and in return get more room (large
seats), better service, and greater flexibility (the tickets can be purchased late and changed
with no additional cost).
- Site licenses for software. A classic example of a volume discount.
No doubt you can think of other examples yourself.

Roughly speaking, firms can generally increase revenue and profit by extracting some of the
consumer surplus. The extreme case is to charge each customer exactly his/her willingness to
pay, thereby extracting all of the consumer surplus.

The hard part is doing it. There are two basic questions a firm must address in considering
whether price discrimination makes sense. The first is whether it can distinguish between
markets. If you cannot tell the customers who are willing to pay more from those who are not,
you are in trouble. The second is whether you can keep the low-price customers from reselling
to the high-price customers.

Take the biotech example. The answer to the first question is that we distinguish between
customers by where they live. Presumably you can do this reasonably well. The second is
more difficult. What if (say) the Brazilians buy at a low price, but some unethical
intermediaries manage to buy in bulk and resell in the US? In finance, we call this arbitrage
and think of it as a good thing. Here it may be either good or bad from a social perspective
(it depends), but it clearly puts a dent into the producer's US profits, and might even make
the drug unprofitable.

Or take airfares. Here the airlines let the customers distinguish themselves. When a family
goes on vacation, it may be sensitive to price and willing to book in advance. But a
management consultant or investment banker often travels on short notice and is less sensitive
to price, the cost of the plane ticket being small relative to the value of the trip. So airlines
have learned that differentiating tickets on how long in advance they are booked is a good way
to segment the market.

Like other things in this course, it's as much art as science. What follows is a catalog of
approaches.

Two-Part Tariffs

One way sellers can extract a consumers surplus is by setting a two-part tariff. Suppose that
each consumer has a known downward-sloping demand curve. By setting the price
corresponding to MR = MC, the seller would be leaving money on the table (a positive surplus
that is captured by the consumer and a potential surplus that is neither captured by the seller
nor the buyer). Suppose however that the seller sets a two-part tariff: a fixed fee and a variable
fee. Suppose moreover that (a) the variable fee is equal to marginal cost; (b) the fixed fee is
equal to consumer surplus given that price is equal to marginal cost. With this two-part tariff,
it is easy to see that the seller would be extracting all of the buyer's surplus.

Lecture Notes 90

Discrimination Among Groups

Firms usually cannot ask every consumer her willingness to pay, but they can segment their
consumers into various groups with differing elasticities and demand curves. For example,
some mail order companies send out retail catalogues to various U.S. markets. These
catalogues are identical with one important exception: prices. A New York City consumer
would receive a catalogue with prices of, say, 10% higher than that for a consumer in Des
Moines. In addition to geographical segmentation, firms can offer special rates to certain
segments (students, senior citizens, etc.). In this way, they can capture more consumer surplus
by better targeting various consumers willingness to pay.

Geographic differentiation is particularly common across countries, where international
borders and local regulations help to reduce arbitrage. Thus similar cars and electronic
appliances (TVs and stereos) still sell for significantly different prices in the countries of the
European single market.

Self-Selection Schemes

Firms often do not have the required information to know your willingness to pay, and thus
are unable to capture your surplus. Yet firms in many industries are adept at inducing
consumers to reveal that information about themselves.

For example, airlines measure differing consumers elasticities in several ways. If you need a
ticket for a business trip, you are likely to be willing to pay a lot for it: you are less price elastic;
you need to take the trip and you need to take it now. Other consumers who are very price
sensitive may book their flights far in advance so they can shop around many rates. In any
event, the amount of time you wait to book a flight reveals important information about your
elasticity. Likewise, firms can offer a product line with various attributes across products to
discriminate between consumers who will choose their selection based on their particular
demand and elasticity. PalmOne, for example, offers a full line of handheld devices. The
PalmOne Tungsten handheld costs $450 whereas cheaper versions of the handheld with fewer
features and less aesthetic quality cost under $100. By giving consumers this choice, they will
reveal their willingness to pay by their choice of product. If PalmOne only had offered one
version, say priced at $250 with a medium amount of features, they would have lost sales
margin from some high-end consumers and sales from some low-end consumers.

When setting different prices for different versions of a product (versioning), the seller must
be careful not to set prices so that the high-end consumers prefer to buy the low-end product.
That would defeat the whole purpose of versioning.

Coupons are another well-known strategy. Some customers will take the time to lug 50
coupons to the checkout counter while other, more inelastic customers will rush to the check-
out without caring about paying the full price.

Lecture Notes 91

Another strategy is discrimination over time. The idea is that the high valuation customers are
less patient, so you can start with a high price and reduce it once the impatient have bought it.
Many information goods work this way (computer software, books).
Bundling and non-linear pricing (e.g., quantity discounts) are also examples of customers
revealing themselves.

Bundling

Bundling is the practice of selling two or more goods together. A supermarket sells apples in
cases rather than individually, a car manufacturer sells cars with several optional add-ons such
as A/C and stereo, a theater selling subscriptions for an entire season of shows, are all
examples of bundling.

Bundling can be pure or mixed. A seller does pure bundling when the only way his
products are sold is as a bundle. In other words, the customers do not have the choice to buy
the goods individually. Examples of pure bundling are most of the double CDs on the market.
Mixed bundling is the practice of selling the goods separately but also to give the customers
to buy the bundle at a discount. For instance, theater tickets could be bought separately, but
buyers usually have the choice to buy a subscription at a price that is lower than the sum of the
individual prices of the tickets.

Typically, bundling is a useful practice if the seller faces a demand that is heterogeneous in
tastes what is called negative correlation of demand i.e., the customers that like one good the
best are the ones that dislike the other good the most.

Example: Double CDs. Consider the situation in which a rock band just recorded 20 new
songs. These 20 songs can fill two CDs. The composition of the fist CD is slightly more metal-
oriented, while the second CD contains a easy-listening ballad.

The typical buyers, who are the fans of the band, can be divided evenly into two groups of
about 50,000 individuals each, the metal-oriented (type A) and the romantic (type B), and their
willingness to pay are the following:

CD1 CD2

Type A 10 9
Type B 9 10

Let us assume MC=0. If the CDs are sold individually, the optimal prices are P
1
=P
2
=9, and the
profit is going to be $1.8M0. However, if the CDs are sold together, the optimal price for the
bundle is $19, and the profit is going to be $1.9M.

Notice that the increase in profit is due to the fact that the demand displays negative
correlation.

Lecture Notes 92

Suppose now that CD2 contains a song that becomes very popular with a segment of people
that are not the usual band fans. They are 1M people. Then, for instance, suppose that the
demand is given by

CD1 CD2

Type A 10 9
Type B 9 10
Type C 0 10

Type C customers does not value CD1 at all and as a result they will not buy the bundle at the
$19 price. On the other hand, by decreasing the price of the bundle we loose surplus with the
other customers. In these situations, in which it is optimal to prune some purchase out, a
mixed bundling strategy can help. Suppose now that we sell the CDs at $10 each and the
bundle for $18.99. At these prices, types A and B will keep on buying the bundle (as they get 1c
of surplus rather than 0 if they just buy their preferred disk), and type C will buy only CD2.
The total profit is almost $19M+$10M=$29M.

Numerical Examples

Example 1. You own a pizzeria that produces at a marginal cost of $6 per pizza and operates as
a local monopoly in a small college town. At lunchtime only students come into the restaurant,
whereas in the evening while students are studying (or doing whatever it is that students do in
the evening) faculty come in. You calculate that students have an elasticity of demand for
pizzas of 4 and that faculty has an elasticity of demand of 2. What are optimal prices?
Suppose both faculty and students come throughout the day. What challenges do you face to
maintain the same revenue as before?

Answer: It will be profitable to charge one price p
L
for the lunch menu and a different price p
D

for dinner. To determine exactly what these prices should be recall the elasticity rule for a
monopolist, which implies that you should charge

6 )
2
1
1 ( =
+
D
p
and
6 )
4
1
1 ( =
+
L
p .

Solving these equations, we get p
L
=$8 and p
D
=$12.

If both faculty and students come throughout the day, this scheme would not work: faculty
would pay the lower price at lunch and you would lose the students custom at dinner. What
alternatives are possible?

Lecture Notes 93

Example 2. Biotech startup BioGar has developed Xamoff, an over-the-counter medicine that
reduces exam-related anxiety. A patent currently protects Xamoff from competition. BioGar
is now thinking of entering the European market but wonders if it should charge the same
price in the two markets. They estimate that the demand curves have the form

q
i
= a
i
b
i
p
i
.

In the US (market i=1), the parameters are a
1
=12 and b
1
=2. In the EU (market i=2), the
parameters are a
2
=4 and b
2
=1. The marginal (and average) cost per unit is c=1. All of these
units are millions. How much could BioGar gain by charging different prices in the two
markets?

Answer. Consider first the problem of setting one uniform price. Total demand at price
p
1
=p
2
=p is

Q = q
1
+ q
2
= (a
1
+a
2
) (b
1
+b
2
) p = A B p.

(The idea is to save ourselves some writing by defining A=a
1
+a
2
and B=b
1
+b
2
.) To find profit
as a function of output Q, we solve for price p=(A-Q)/B and substitute:

Profit(Q) = pQ cQ = cQ Q
B
Q
B
A
cQ Q
B
Q A
=
2
1
.

To maximize, we differentiate with respect to Q and equate to zero, which yields

Q = 5 . 6
2
=
cB A
.
Price is
p= 17 . 3 =
B
Q A
.

Finally, quantities are q
1
=5.67 and q
2
=0.83, and Profit=14.08.

Now find the best prices in the two markets separately. The presumption is that we can avoid
parallel imports from Europe (which we guess is the cheap location) back to the U.S.
Mathematically, this is two separate monopoly problems, but well do them simultaneously.
Profit is in this case depends on both quantities, but otherwise we follow similar logic:

Profit(q
1
,q
2
) = ) (
2 1 2
2
2 2
1
1
1 1
q q c q
b
q a
q
b
q a
+
.

This time we differentiate with respect to q
1
and q
2
(one at a time) and set each derivative equal
to zero. The result is

Lecture Notes 94

5
2
1 1
1
=
=
cb a
q and 5 . 1
2
2 2
2
=
=
cb a
q .

The prices are now p
1
=3.5 and p
2
=2.5, and Profit=14.75. We conclude that, by setting
different prices in Europe and in the U.S., profits increase from 14.08 to 14.75, an increase of
about 4.76%.

For your own enlightenment: Verify that the elasticity rule applies to each market. Show, too,
that total consumer surplus falls when we charge different prices.

Example 3. We had the baby Mac, then the iMac; its now time for the baby iMac. As
head of marketing of Apple Computer, you decided you could do better than the current
situation. Last year, the company sold 1 million iMacs for $1,500 each. This is the most you
can get from the market segment that currently buys the iMac. According to a marketing
study, there is a second market segment of 2 million people willing to pay up to $500 for a
stripped-down version of the iMac. Your market researchers also tell you that (i) the first
segment would be willing to pay up to $800 for the stripped down version, (ii) the second
segment would not be willing to pay more than $600 even for the full-fledged version of the
iMac. Finally, your production people tell you that it costs $300 to produce an iMac, be it the
standard version or the stripped-down version.

What is your optimal pricing policy?

Answer. A first possible strategy (benchmark) is to only sell the full version and charge $1,500.
This would lead to selling 1million units, for a total profit of (1500-300) x 1m = $1.2b. A
second possible strategy would be to hit each segment by charging $500 for the stripped-down
version and $1500 for the full version. But would this work? No: high-end consumers get
zero value from buying the full version (its priced at exactly their value), but 800-500=$300
from the stripped-down version. Thus they would buy the stripped-down version. An
alternative strategy is to charge $1,200 for the full version (think of it as slightly less than
$1,200) and $500 for the stripped-down version. This will lead high-end users to pay $1,200
and low-end users to pay $500. Total profit is now (500-300)x2m + (1200-300)x1m = $1.3b,
an improvement over the current solution.

Example 4. Monthly individual demand for hours at the NPNG (no pain no gain) gym is Q=10
P, where P is price per hour. Marginal cost is zero. Find the optimal price and quantity
under standard pricing. What is the profit per customer? Suppose the gym charges monthly
passes. What is the optimal price of a monthly pass? How much do customers pay per hour?
What is the profit per customer?

Answer. The standard monopoly pricing problem implies an optimal price of P=5 and output
Q=5, for a profit per customer of 25 (note that this is variable profit and does not take into
account any fixed cost that might be incurred).

Lecture Notes 95

Suppose now that the gym can charge a monthly fee. Following the logic of two-part tariffs, it
should charge an hourly fee equal to the marginal cost, P = 0, in which case demand will be 10;
and a fixed fee equal to the consumer surplus at this price, that is,

50
2
10 * 10
= = F .

In this case, profit per customer is 50, a clear improvement over simple pricing.
Public Policy

Price discrimination is not per se illegal. But it raises issues if used in a way that restricts
competition. It also raises political issues (think of AIDS drugs). Perhaps for that reason,
many price discriminations schemes are disguised, either through bundling or versioning or
more complex means. See how long it takes you to figure out the Mets ticket pricing scheme!

Lecture Notes 96

Market Power

Previously, we have discussed competitive markets and how they generate efficient outcomes.
We highlighted that the Fundamental Theorem rests on some key assumptions. In this
chapter, we relax the assumption that firms have no market power. What is market power?
How (and why) do governments tend to set rules limiting anti-competitive behavior? These
are some of the central themes we will address.

Industry Structure

Industries vary widely in range of products (broad or narrow), geographic scope (cement is
harder to transport long distances than CDs), cost structure (how does the minimum efficient
scale compare to the size of the market?), number of firms (presumably related to costs), and
intensity of competition (vaguely related to the number of firms).

Exhibit 1 describes several mostly-US industries in broad terms. In addition to a name, each
industry comes with an SIC code (standard industrial classification), which can be used to
identify it. Most are 4-digit codes, which define industries with a modest amount of detail.
Even so, many of the industries listed cover a wide range of products. Telecommunications,
for example, covers local service, long distance, wireless, and a variety of specialist providers.
If we wanted to understand the industry in any depth, wed want to look at finer categories.

These industries differ in a number of respects. One is how concentrated they are: how
dominant the largest firms are in terms of their market shares. The 4-firm share (sometimes
called the 4-firm concentration ratio) is the market share of the largest 4 firms (these market
shares are computed from revenue). They range from 1.8% for legal services (there are lots of
law firms, and even the largest isnt very big relative to the market overall) to 84.8% for aircraft
manufacture (where, to be honest, a national number is less meaningful than a global one).

Another measure of concentration is the Herfindahl-Hirschman Index (HHI), which is
widely used by regulators to guide decisions on mergers. The HHI is computed by

HHI = (s
1
)
2
+ (s
2
)
2
+ (s
3
)
2
+ + (s
n
)
2
,

the sum of the squared market shares (expressed as percentages) of the firms in the industry.
If this sounds somewhat esoteric, note that a monopoly would have HHI=10,000 [=100
2
], an
industry with many small firms would have HHI close to zero, and an industry with 4 equal-
size firms would have HHI=2500 [=25
2
+25
2
+25
2
+25
2
] (approximately what we see for motor
vehicles).

Firms and Markets
Lecture Notes

Lecture Notes 97

An alternative route to measuring market powerone that we have considered in the notes on
pricingis the margin (the premium above marginal costs that the firm is charging). However,
margins may be difficult to determine without accurate cost data. Moreover, if products are
bundled or firms use other advanced pricing schemes it may be difficult to determine the
relevant price margins. Thus, in practice margins are used less frequently than market shares.

Concentration measures are relatively crude. Depending on the industry, one might want to
see finer product categories and/or regional differences (for airlines, for example,
concentration is considerably greater at individual airports). While concentration measures
dont tell us exactly how intense the competition is, theyre a good start. Generally, wed
expect to see intense competition if there are many firms, no competition if there is one firm,
and something in between if there are a few firms. We have looked in some detail at the two
polar cases (competition and monopoly) and plan to turn to the more complex issue of how a
small number of firms (an oligopoly) might interact.

Monopoly

A firm that is large enough to influence price will generally restrict output and raise price.
Thats good for the firm, in the sense that the monopoly position is profitable. But is it good
for society as a whole?

The economic basis for competition policy goes back to Adam Smiths invisible hand. We
expect the forces of competition to allocate resources where they are most valuable, thus
increasing the welfare of consumers. A monopoly, however, restricts output, raises price, and
attracts fewer resources to the industry than one would get under competitive conditions. In a
sense, not enough resources are allocated to the industry. Theoretically, we might note that the
sum of consumer surplus and producer surplus (profit) is lower under monopoly. The
difference is termed the deadweight loss of monopoly.

All this could change if there are economies of scale. In this case, costs would be higher with
competition. The question, then, is whether the cost advantage of monopoly exceeds its
markup.

Government Policy

Firms and industries range from government-owned and operated (Post Office) to regulated
(telecommunications, local utilities, pharmaceuticals, financial services), to unregulated but
subject to laws on fraud, anti-trust, and so on. As economist Andreu Mas-Colell once said:
The invisible hand of the free market apparently needs a lot of help. Thats true even in the
US, the bastion of free markets.

Competition policy. Most countries have laws enforcing competition: outlawing price agreements
among competitors, restricting mergers that substantially increase market concentration, and
limiting the behavior of monopolies and near-monopolies.

Lecture Notes 98

Most countries outlaw price fixing (when two or more firms get together to agree on price and
carve up the market). In effect, price fixers agree to act collectively as a monopoly (the term is
cartel). The DeBeers diamond cartel is a good example.

Many countries also limit mergers that they feel would reduce competition in one or more
markets. In the US, the Department of Justice has guidelines based on the HHI. As a rule of
thumb, mergers that lead to an HHI of 1000 or less ordinarily lead to no further scrutiny; those
that raise the HHI by 100 or more and produce a post-merger HHI between 1000 and 1800
raise significant questions; and those that raise the HHI by 50 or more and produce a post-
merger HHI above 1800 raise significant concern. All of this is contingent on the definition of
the market (geographic, range of products).

Monopolies themselves are not illegal, although monopolies generally have an obligation not to
exploit their monopoly advantage to enter other markets. That was an issue in the Microsoft
case: whether Microsoft had used its monopoly position in operating systems to extend its
reach to the browser market.

International issues have become increasingly important, the ill-fated GE/Honeywell deal
being a recent example. In this case, a merger of two U.S. firms was blocked by the EUs
Competition Commission on the grounds that it would reduce competition in some European
markets (specifically, aircraft engines).

Regulation of monopolies. Typically, industries in which the minimum efficient scale leaves room
for only a single producer are regulated. The basic theory of regulation is to force a price equal
to what wed see under competition. Experience tells us, though, that poorly implemented
regulation can be as bad as unfettered monopoly. Governments continue to experiment with
new approaches.

Essential facilities. A critical issue in monopoly regulation is what to regulate. For years,
governments regulated telecommunications in their entirety. But more recently, governments
have loosened restrictions on long distance service, which is increasingly competitive. Local
service, however, is generally regulated by states. The argument is that the local connection is a
natural monopoly (its expensive to build more than one) and should be regulated, but long
distance is not. Similar arguments have been applied to electricity.

Lecture Notes 99

Exhibit 1. Properties of Selected Industries

Industry SIC Code 4-Firm Share HHI Region Source (Date)
Electric power 2211 14.3 US Commerce (97)
Pharmaceuticals 3254 32.2 446 US Commerce (97)
Computers 3341 37.0 465 US Commerce (97)
Semiconductors 334413 52.5 1080 US Commerce (97)
Motor vehicles 3361 82.4 2506 US Commerce (97)
Aircraft 3364 84.8 1637 US Commerce (97)
Air transportation 4811 25.5 US Commerce (97)
Motion pictures 5121 32.5 US Commerce (97)
Sound recording 5122 53.1 US Commerce (97)
Telecommunications 5133 42.3 US Commerce (97)
Commercial banking 52211 17.3 US Commerce (97)
Securities brokerage 52312 37.7 US Commerce (97)
Wholesale banking 716 World Roy Smith (98)
Legal services 5411 1.8 US Commerce (97)
General hospitals 6221 73.5 US Commerce (97)
Performing arts cos. 7111 8.5 US Commerce (97)

Sources. US Department of Commerce web site, Roy Smith.

Lecture Notes 100

Auctions

Many business transactions are now handled with auctions, rather than through conventional
markets in which sellers post prices and buyers decide whether and how many to buy.
Common examples include fine art, electro-magnetic spectrum rights, US treasury securities,
many equity markets, some IPOs, privatizations of state-owned companies, wine, eBay, and
on-campus interviews. You get the idea. These are all sales, but there are many auctions by
buyers, too: government contracts to buy goods and services, for example.

Why auctions? Our usual market story is that sellers get together with lots of buyers, and after
enough transactions they figure out a reasonable price: an equilibrium. But in most of the
examples we just mentioned, the assets are unique; we dont have the luxury of experience to
establish a reasonable price. Once a Vermeer painting or 3G spectrum rights are sold, theyre
gone. If you price such a product/asset low, you get less revenue than you could. If you price
it high, you get no takers (and no revenue). The solution is to run an auction, and let buyers
reveal what they think its worth.

But what kind of auction? A conventional English auction? A Dutch auction in which the
price falls until someone bids? A sealed-bid auction? Over the last few decades, there have
been enormous advances in our understanding of auctions and comparable advances in their
range of use. Well try to give you some sense of both.

The theory of auctions has taken great strides in recent years, and builds on the theory of
mechanism design (the topic for which Hurwicz, Maskin and Myerson won the 2007 Nobel prize
in economics). The broad field of mechanism design (as the name might suggest!) looks at
designing a system of rules and regulations in a way that achieves good outcomes and takes
into account that those involved will change their behavior as the rules change. In the context
of auctions, the auctioneer can choose the type of auction that is run and will wish to consider
the impact on the revenues raised and (for example if the auctioneer is the government) the
efficiency of the outcome (does the person who gets the good value it most, is the contractor
who gets the job the one who can conduct it most efficiently etc.)

Auction design has hit the headlines, in its use to sell public assets. Historically, and still in
some countries and some sectors public assets were sold through beauty contests in which
firms submitted long-detailed proposals and a committee of officials would choose which they
liked best. However, in part following the theoretical work, auctions are being increasingly used
and with some success. Sales of radio spectrum (necessary to run a mobile phone operation)
throughout Europe and the US have raised hundreds of billions of dollars, and seemed to raise
much more revenue than the previous system of beauty contests. The telecom auction in
Britain alone raised $35 billion!

Firms and Markets
Lecture Notes

Lecture Notes 101

Common Auction Mechanisms

The most common auction types are:

- The (ascending) English auction, in which buyers make increasingly higher bids until
only one bidder remains. The highest bidder buys the object at the highest bid. This is
how auctions for antiques in auction houses (Sothebys or Christies) are conducted

- The (descending) Dutch auction, which works the other way around: starting with a
very high bid, the price is reduced until someone is willing to bid. The first and only bidder
buys the object at the bid price. Flowers are sold in this way in the Netherlands and this
can be a particularly fast way to run an aucion.

- Sealed bid auctions, in which potential buyers submit sealed, or secret, bids for an
object. The highest bid buys the object for the bid price. If you think about this, you
might realize that the sealed bid and Dutch auctions are equivalent, in the sense that
bidders do not have any information about other bids and the highest bid wins.

o A variant is the Vickrey auction, in which buyers submit sealed bids. The
highest bidder wins, but must pay only the second highest price.

There are lots of variations on these basic designs, but this is a good start. The obvious
question is which of these auctions works best. For example, if youre the seller, which auction
generates the most revenue? The answer is the usual: it depends. Well be fairly concrete
what it depends on before were done.

How Bidders Value the Object
Probably the most important ingredient in an auction is the values placed on the object by
potential buyers. Well assume that the seller does not know the values placed on the object by
potential buyers; if she did, she could simply charge a price equal to (or just below) the highest
valuation. Well also assume that buyers do not know each others valuations. They do know
their own valuation: we say that it is private information, since the knowledge is their own.

However, to progress we need to make some assumptions about what the auction designer and
bidders believe about what the distribution of values that other bidders might be. In trying to
figure out what price to bid on in a Dutch auction, for example, a bidder will clearly bid
differently if he expects rival bidders to have say a 90% chance of having a valuation of $10
and a 10% chance of a valuation of $12 than if he expects rival bidders to have an equal
likelihood of any valuation between $5 and $8.

Key for the design and performance of auctions therefore would appear to be what is known
about bidders valuations. The way theorists think about this is to assume that we know the
distribution from which their values are drawn. (This is like much of statistics: we use a pdf to
summarize the extent of our uncertainty.)

Lecture Notes 102

As well as taking great care in what bidders know about each others valuations. Theorists (and
practitioners!) must be careful to distinguish between private values and interdependent values.
Bidding with Private Values

If I buy a picture solely to put on my wall to enjoy, then the benefit that I get from that does
not depend on others valuations. We call this a private values case and here, bidding
strategies are relatively simple.

For example, suppose that Luis wants to sell his apartment and the only potential buyers and
John and Heski.

To see how they should bid under different auction formats, we need to be more specific
about the value of the object being auctioned. Well start off by assuming that for each of
Heski and John their valuation for the apartment is equally likely to take any value between 0
and $1 million.

Consider first the English auction: Heski and John should each keep bidding until the going
price reaches their valuation, after which they should keep quiet. The winner will therefore pay
the losers valuation. Since the losers valuation is equally likely to be anything less than the
winners valuation, on average the winner ends up paying half his valuation.

It is now easy to see that the English auction raises as much as the Vickery auction. Suppose
Heski and John submit their true valuations as their bids in the Vickery auction then Luis sells
his house to the highest bidder at a price equal to the second highest bid: with the same
average payoff as the English auction.

Note that we must of course ask whether they would want to submit their valuations as bids. It
is easy to see that they do. No matter what John does, Heski wold like to bid his valuation. If
John bids lower than Heskis valuation, then by bidding his valuation Heski can guarantee
winning without affecting the price. If John bids higher than Heskis valuation then Heski
doesnt want to win! and so might as well bid his valuation.

Bidding strategies for the first price auction, or the Dutch auction are a little more subtle. In
such auctions, you wont want to bid your true valuation, since if you win then you pay your
bid and so you have pay your full valuation and dont gain anything by winning, and so a
bidder shades their bid down below their valuation; however the more you shade your bid, the
lower the probability of winning. Trading off the probability of winning against the gain from
winning turns out to follow a similar analysis to the margin-volume trade-off for a price-setting
monopolist, though we leave the analysis for another day

Bidding with Interdependent Values

When agents have private values they learn nothing in the course of the auction. However
there are cases where the agents have interdependent values that is where if Heski has a
higher valuation then John is also likely to have a higher valuation. For example, suppose that
instead of buying a house to live in, Heski and John were both looking to buy as an investment

Lecture Notes 103

(rather than to live there). They might have different knowledge of the market (leading to
different estimates).
If John naively bids his full valuation in the Vickery auction, for example, then on winning the
auction he will usually be disappointed. Why? Because his winning the auction suggests that
Heski had a lower valuation, which suggests that Heski had relatively bad news about the state
of the market and that Johns estimate was an over-estimate.

This phenomenon is called the winners curse and has been widely discussed in the context
of oil tracts. For all oil drilling companies, the key determinant in valuing a tract is the amount
of oil to be found there. Different bidders estimates are likely to rely on differing geological
surveys some are likely to be more optimistic and some more pessimistic. If you win then it
is likely that you had a relatively optimistic survey and so bidders should shade down their bids
to take into account this effect.

Since an English auction allows bidders to gain some information about other bidders
valuations in the course of the auction, it allows bidders to realize when other bidders are
optimistic and so will typically raise greater revenues than a Dutch or sealed bid auction.

Practical Considerations

Beyond the considerations addressed above there are a number of further practical
considerations to be considered:

- In general we might think about the costs of making a bid. Do bidders have to devote
resources to prepare business plans etc ahead of bidding? Are bidders forced to
disclose valuable commercial information?

- In general, a key determinant of revenue for an auctioneer is the number of bidders (an
auction with only one bidder, regardless of the format is unlikely to raise much
revenue---this sounds obvious but there are examples of telecom auctions when there
were no more bidders than objects being sold). Encouraging bidder participation is
therefore critical.

- How long does the auction take to run? This is a particular issue for perishable goods
and affects the design of fish markets in Tokyo and flower markets in Amsterdam?

- In the Vickrey auction you need to trust the auctioneer, who has an incentive to change
your bid, or make claims about others bids that can be hard to verify.

- An English auction can make it easier for bidders to collude in their bids (it is easier for
those colluding to verify that the others are sticking to their agreements).

- Is the winners curse a problem?

Botton line: This gives you the idea that posting prices is only one way to sell.

Lecture Notes 104

Further Reading

This is a topic about which experts have real knowledge: they simply know more about how
auctions work than the rest of us. If youre interested, a number of interesting sources are
easily availabe:

- Paul Klemperer has a free and excellent online book Auctions: Theory and Practice at
http://www.paulklemperer.org/index.htm. reading the whole book might be for those
only very interested, but the introductory chapter is an excellent introduction.
- John McMillan, Selling Spectrum Rights, Journal of Economic Perspectives, Summer 1994,
145-62.
Paul Milgrom, Auctions and bidding, Journal of Economic Perspectives, Summer 1989,
pages 3-22.

Lecture Notes 105

Game Theory: Introduction to Strategy

With perfect competition and monopoly, the interactions between firms are minimal to
nonexistent. In a perfectly competitive market, the impact of competitors is summarized
entirely by the market price. In a monopoly, there are no competitors to worry about. In
markets with a small number of firms, however, the interactions can be substantially more
complex. If I cut prices, will my competitors cut them, too? If I enter a new market, will
others follow? If I introduce a new flavor of ice cream, will my competitors imitate or look for
other niches in the ice cream market?

This note is an introduction to the formal analysis of interactions among a small number of
players: what we call game theory. Well do this in settings that are very. For example, there
will typically be only two players. Why? Because its helpful to reduce distractions, even if
that makes the situations less realistic. Of course, the challenge is to recognize strategic
situations in practice, where they often come in heavy disguise.

A Strategic Situation

Suppose you own a software business. You only have one competitor. You can decide to
launch a new software package this year or not to launch one. You dont know whether your
rival has decided to launch her own new software package. However, you are pretty certain
what will happen in the following four scenarios. If I launch and my rival launches, we both
lose money. If I launch and she doesnt, only I make money. If she launches and I dont, only
she makes money. And if neither of us launch, we both lose money but less than if we both
launch.

You wish you knew what your rival was going to do, but you dont. So how do you proceed?
You might say to yourself: If she thinks that I am going to launch, she wont. But then you
think: If she thinks I think she will launch and thus I dont, she will launch and I will not
make money. But that seems too simple. So after pondering a bit more you think: If she
thinks that I think that she thinks that I think she will launch, she might think I will actually
launch and so she wont. And so on, and so on. Pretty soon youre totally confused and have
no idea what to do.

How do we break out of this logical conundrum? How do we reason strategically when
confronting uncertainty and risky decisions in a competitive setting? How do we predict what
our competitors will do given what we will do when our decisions are based on what our
competitors do (and vice versa)? Well see.

Firms and Markets
Lecture Notes

Lecture Notes 106

As we work through some formal games and learn how to use game matrices and trees, well
come across a few general ideas that underlie strategic behavior. Among them:
- Understand your competitor.
- Timing matters.
- Information matters, too.
- Some games are better to play than others.
- Some games have win-win solutions and some dont.

These platitudes will become more than that in a few weeks.

Two Ways to Represent Games

Formal games have a number of ingredients: a list of players, rules that govern how they play,
a set of possible strategies (moves) for each player in every possible situation, and the
payoffs (benefits) to each player for every possible combination of strategies. Well focus here
on relatively simple games, in which two players make one move each from a short lists of
choices. A key part of the rules is the order of moves. In a so-called normal-form game,
the players choose their moves at the same time without knowledge of the others move. We
represent such games in a matrix (table) (sometimes we will refer to these games as matrix
games). In a so-called extensive-form game, players choose and make their moves
sequentially, one after the other. (Think of chess.) We represent these games with game trees
(sometimes we refer to them as tree games).

Lets talk about normal-form games first. An example might be:

L R
20 10
T 15 5
5 15
B 10 20

In this game, Player 1 and Player 2 choose moves at the same time. Player 1 chooses between
top (T) and bottom (B). Player 2 chooses between left (L) and right (R). The numbers in the
matrix tell us the benefits to the two players of each combination of strategies (moves), with
Player 1s benefits in the lower left corner of the relevant box and Player 2s in the upper right
corner. For example, if Player 1 chooses T and Player 2 chooses R, the benefit to Player 1 is 5
(in whatever units seem appropriate to Player 1) and the benefit to Player 2 is 10.

Player 2
Player 1

Lecture Notes 107

Extensive-form games have similar elements. A sample game tree might be:

Here, the boxes tell you who moves when and the branches illustrate the possible choices at
each point. Thus Player 1 moves first, choosing (again) between T and B. Once Player 1 has
moved, Player 2 chooses either R or L. In this case, the two 2 boxes show you that Player
2s choice might depend on the choice made by Player 1. The payoffs at the end tell you the
benefits to the two players for every combination of strategies, with payoffs to Player 1 listed
first.

A point to keep in mind: Even though these two games look similar (same players, same
choices of strategies, same payoffs), theyre different games. Why? Because the order of
moves is different. Well see shortly that this can make a difference to the outcome.

Outcomes to Games

The next step is to consider reasonable outcomes of games. What would we expect Player 1 to
do? Player 2? What is the outcome of the game and the payoffs received by each player? As
an illustration, let us consider the following normal-form game:

L C R

T
5 6 7
9 8 1

M
3 5 6
1 2 0

B
7 6 8
2 3 4

1

2

2

(15,20)
(5,10)
(10,5)
(20,15)
T
H
B
L
H
R
L
R
L
S
H
Player 1
Player 2

Lecture Notes 108

Dominant and dominated strategies. In some games, there are moves that are good in all situations,
in the sense that their payoff is greater for every strategy chosen by the other player. Such
strategies are dominant. Conclusion: play them. In the above game, Player 2s R is a
dominant strategy: its better than both L and C for any choice of Player 1.

In some other games, you may find that one strategy has lower payoffs than another strategy
for every move by the other player. We say that such strategies are dominated (by the other
strategy). In the game above, Player 1s M is dominated by T (and by B, too!): Regardless of
Player 2s choice, T is better than M (and so is B). Clearly youd never play a dominated
strategy. The converse need not be true: just because you wouldnt play a strategy doesnt
mean its dominated. We conclude: do not play dominated strategies. We can cross them out
and proceed without them.

One important note, taking the above game as an example: If Player 2 expects Player 1 to
choose B, you might think that Player 2 is better off by choosing L: even though Player 2s
payoff is lower than by choosing R, Player 1s payoff is lower in the (B,L) combination than in
(B,R). However, each players goal is to maximize his payoff, not to minimize the other
players payoff or maximize the payoff difference. True, in some real-world applications the
rivals payoff matters. For example, if the rival firms profit is sufficiently low then the rival
firm may decide to exit, leaving me as a (more profitable) monopolist. If that were the case,
then we should include this additional benefit as part of my payoff. In other words, the
payoffs in the matrix already reflect all of the above considerations.

Best responses. In many games, the best strategy choice depends on the other players strategy.
We call the best choice for each move by the other player the best response. If theres a
dominant strategy, its the best response for every move by the other player. In this case,
Player 2s best response is to play R always. We mark those payoffs in bold. How about
Player 1? If Player 2 chooses L, the best response is T. If Player 2 chooses C, the best
response is also T. And if Player 2 chooses R, the best response is B.

Nash equilibrium. We consider a reasonable outcome in a game to be a combination of
strategies in which each player chooses its best response. Such a combination is called a Nash
equilibrium. In the preceding example, (B,R) is a Nash equilibrium. At a Nash equilibrium,
neither player has an incentive to change its strategy unilaterally.

Summary:
- Dont play dominated strategies.
- Do play dominant strategies.
- In general: Each player chooses best response to every choice of the other player.
- Equilibrium: Each player makes the best response to the others possible move. An
outcome in which each player chooses his best response is called a Nash equilibrium.

Lecture Notes 109

Historical note. Early pioneers in game theory include the brilliant mathematician John von
Neumann (Theory of Games and Economic Behavior, 1944, with co-author Oscar Morgenstern) and
John Nash. Nash is a particularly intriguing case. After doing path-breaking work in the early
1950s, schizophrenia kept him out of circulation until he recovered in the 1980s. He received
the Nobel Prize in economics in 1994. Theres a terrific book about him, Sylvia Nasars A
Beautiful Mind (theres also a fairly inaccurate movie of the same name).

Examples

The most famous example of a game is the prisoners dilemma. A business version is this:
Firm A and Firm B both have to set output levels. They cannot discuss their choice of output
levels, because thats illegal. Each would like total supply to be low (to generate a high price)
but sell a lot itself. You can see the conflict. An example might look like:

High Low

High
10 5
10 20
Low
20 15
5 15

You can see the incentive to get together to divide the market: both firms do well if they agree
to limit supply (Low, Low). But you can also see the incentive to produce more: once (say)
Firm B chooses Low, the best thing A can do is produce High.

Lets examine this formally as a game. The problem is that High and High are dominant
strategies. The only Nash equilibrium is therefore for both firms to produce a lot (High,High),
which drives down price and profit. This shows why cartels are so attractive to firms: they
could both make more if they were allowed to collude. It also shows why in the absence of a
binding agreement cartels tend to fall apart.

The Prisoners Dilemma is striking in showing how a bad outcome (for the firms) might be an
equilibrium. But not all games work this way. You might try slapping down some numbers
and see where they lead. Some games have win-win possibilities. Consider below the game we
looked at earlier, with some local color added.

Whats going on here? Richie and Dave are friends and would like to do something together.
Richie prefers the Opera and Dave prefers the Knicks, but its more important to each of them
to do something together than to get their first choice. Lets assume that they cant do the
obvious thing and simply agree on what to do: Each must make his choice without knowing
what the other will do. Whats the Nash equilibrium? In this case, there are two Nash
equilibria, (O,O) and (K,K). The difficult part is how to choose among them, and game theory
has little to say about that.

Firm B
Firm A

Lecture Notes 110

Knicks Opera
20 10
Knicks 15 5
5 15
Opera 10 20

Many students find this confusing, so lets take a closer look. A common question is: If were
moving simultaneously, how do we know which Nash equilibrium to play? And if we dont
know which one, couldnt we end up at a non-equilibrium square? (Yes!) A defense of the
Nash equilibrium outcome might be that neither player would want to change its strategy
unilaterally. Or we might decide that the concept of Nash equilibrium is simply not enough in
this case to tell us what the outcome of this game should be.

Whatever its complexities, the situation turns up in the business world routinely. One example
is standards for new products. Two companies might have a preferred technical standard, but
in some cases theyre each better off if the other chooses the same standard. That is, each firm
would prefer to use the competitors standard than choose different standards and confuse the
market. Things like industry associations are sometimes used as ways to accomplish this.
(Technical agreements between competitors are not per se illegal.)

Timing Matters

The same game can be used to illustrate the impact of timing on the outcome. Suppose Richie
gets to move first: he chooses, and calls Dave to tell him where to meet. Dave then decides
whether to meet him or choose the other event. The game tree looks like this:

How do you find the outcome for a game like this? Answer: you start at the end. When its
Daves turn to move, Richie has already chosen. If Richie chose K, then the best thing Dave
can do is choose K (20 is greater than 10). If Richie chose O, then Dave chooses O, too (15 is
Dave
Richie
R

D

D

(15,20)
(5,10)
(10,5)
(20,15)
K
T
H
O
B
K

O
K
O
S
H

Lecture Notes 111

greater than 5). Richie now takes Daves choices into account in choosing his own strategy.
He knows that hell get 15 if he chooses K and 20 if he chooses O, so he chooses O. The
equilibrium: (O,O).

You can see how important the timing of moves is here. If both move at the same time, its
not clear which Nash equilibrium occurs. If Richie moves first, they go to the Opera. If Dave
were to move first, theyd go to the Knicks. You might say that Dave has a first-mover
advantage in this game. The point, however, is not that going first is good (sometimes it is,
sometimes it isnt), but that timing has an impact on the outcome.

Additional examples

Example 1. The tables below represent a series of two-player games which illustrate the rivalry
between Time magazine and Newsweek. Each magazine's strategy consists of choosing a
cover story: Impeachment or Financial crisis are the two choices. (Times strategy is in
row, Newsweeks in column. In each cell, the first value is Times payoff, the second
Newsweeks payoff.)

Impeachment Financial crisis
Impeachment 35,35 70,30
Financial crisis 30,70 15,15
(i) Time and Newsweek are evenly matched

(ii) Time is more popular than Newsweek

(iii) Some readers will buy both magazines

The first version of the game corresponds to the case when the game is symmetric (Time and
Newsweek are equally well positioned). As the payoffs suggest, Impeachment' is a better
story but payoffs are lower when both magazines choose the same story. The second version
of the game corresponds to the assumption that Time is a more popular magazine (Time's
payoff is greater then Newsweek's when both magazines cover the same story). The third
version of the game is the case when the magazines are sufficiently different that some readers
will buy both magazines even if they cover the same story.

For each of the three versions of the game,
(a) Determine whether the game can be solved by dominant strategies;
(b) Determine all Nash equilibria;

Lecture Notes 112

(c) Indicate clearly which assumptions regarding rationality are required in order to reach
the solutions in (a) and (b).

Answer. In (i), Impeachment is a dominant strategy for both players. It follows that
(Impeachment, Impeachment) is the unique Nash equilibrium. All we need to assume to reach
this conclusion is that players are rational and know their own payoffs.

In (ii), Impeachment is a dominant strategy for Time, but not for Newsweek. Given that Time
chooses Impeachment, Financial Crisis is the optimal choice for Newsweek. It follows that
(Impeachment, Financial Crisis) is the unique Nash equilibrium. This solution assumes that
Time is rational and knows its payoffs; and Newsweek is rational, knows the payoffs for both
players, and believes Time is a rational player.

In (iii), there are no dominant strategies. There are two Nash equilibria (in pure strategies):
(Impeachment, Financial Crisis) and (Financial Crisis, Impeachment). In this context, the
concept of Nash equilibrium presupposes that players know the payoffs of both players;
moreover, it is common knowledge (I expect that you expect that I expect...) that the particular
equilibrium will be played.

Example 2. Consider the situation where Ericsson and Nokia compete in the market for 4G
handsets. Each firm contemplates setting one of two possible price levels: $100 or $90.
Production cost is the same for both firms: $40 per handset. Market demand as a function of
the prices set by each firm is given by the following matrix (notice that this is not the payoff
matrix; each firms payoff is profit, not output).

100 90
800 1100
100 700 400
700 900
90 800 600

Suppose firms choose prices simultaneously. Describe the game and solve it.

Answer. Based on the values provided, we can compute profits for each price combination.
For example, if Ericsson sets a price of 100 and Nokia a price of 100 too, then Ericsson gets a
demand of 700k and total profit $42m = (100 40) 700k. Doing the same for all price
combinations we obtain the following profit matrix (in $m per quarter):

100 90
48 55
100 42 24
42 45
90 40 30

Ericsson
Nokia
Ericsson
Nokia

Lecture Notes 113

Nokia has a dominant strategy: to price at $90. Knowing that Ericsson should price at $90 too,
though it is not a dominant strategy. The Nash equilibrium of this game is therefore (90,90),
leading to profits of (30,45) for Ericsson and Nokia, respectively. Notice that both Ericsson
and Nokia are worse off than they would be by pricing at $100. This game has a similar
outcome as the prisoners dilemma.

Example 2 (continued). Suppose that Nokia has a limited capacity of 800k units per quarter.
How would the analysis change?

Answer. The constraint binds in the case when Nokia sets a price of $90. In this case, Nokias
sales are now 800k, not the values previously reported. Everything else remains the same. The
new payoff matrix is:

100 90
48 40
100 42 24
42 40
90 40 30

It is now a dominant strategy for Nokia to price at $100. Knowing that Ericsson should price
at $100 too. The new equilibrium is therefore given by (100,100), leading to profits (42,48) for
Ericsson and Nokia, respectively. This example shows that one firms capacity constraint can
soften price competition and benefit all firms in the market.

Example 2 (continued further). Suppose you work for Ericsson. Your CIO is unsure whether
Nokia is capacity constrained or not. How much would you value this piece of information?

Answer. This information would have no value (as far as this game is concerned). In fact,
regardless of whether Nokia is or is not capacity constrained, Ericssons dominant strategy is to
price at $90. Information only has value if it would lead me to change my course of action and
improve my prospects.

Other Issues

In the coming weeks, well look at additional strategic issues. Among them:

- Sources of leverage. When firms interact repeatedly, either over time or across markets,
this gives them additional leverage over each other. Sometimes that can be used to
moderate the impact of Prisoner-Dilemma-like situations. Example: American Airlines
and United compete on the NY-Chicago route. If American reduces fares on this route to
gain market share, United might not only meet the lower fare, but also retaliate by reducing
its fare on the Chicago-Dallas route, which is important to American. This makes fare-
cutting less attractive to both players.

Ericsson
Nokia

Lecture Notes 114

- Information. When one player knows something the other doesnt (or knows it earlier),
that can affect the game, too. Example: Youre a new entrant to the PC market and
buyers do not yet know the quality of your products. How do you convince them that
your products are good?

- Reputation. This is essentially a combination of the first two. There are times when a
firms actions have an impact not only on the current situation, but also on customers and
competitors interpretations of your future actions. An example is branding: A company
with a strong brand and products will be less willing (presumably) to sell shoddy products
under the same brand. (Can you see why?) Or: Intel suing AMD sends the message that
Intel will aggressively defend its property (and maybe more).

- Choose the right game. Some games lead to bad outcomes and you should avoid them if
you can. Example: The Browser Wars drove the price of web browsers to zero. Could
Netscape and Microsoft have played a different game with a better outcome?

Lecture Notes 115

Price Competition

Weve acquired some tools to analyze games in which players are taking a short-run
perspective, that is, they are trying to do as well as possible in the game today, without thinking
about the long-run implications. The relevant concept is Nash equilibrium: players settle into
playing a best response to what the other players are doing.

Well start by looking at differentiated products, which are markets in which consumers
perceive a difference between different brands: a button-down white mens shirt from Ralph
Lauren is not the same as a button-down white mens shirt from The Gap, though each may be
preferred by different customers. Then well look at goods that are identical from the point of
view of the customer, such as gas or sugar or flour; these are sometimes called commodity
businesses. Well see how serious the short-term price competition perspective can be, in that
environment.

A. Differentiated Price Competition

Lets think of Time and Newsweek, and imagine that every week they are choosing their price
for the week, not knowing what the other magazine has chosen: thats a simultaneous game.
Here prices are the strategic variable, so we will think of all payoffs in terms of price, and
actions in terms of prices (i.e. not in terms of quantities!).

Time and Newsweek are looking for the stable point, or Nash equilibrium, of this game. That
involves figuring out the best responses for each player to what the other player does. For
instance, Newsweek will have to ask the following question: Suppose I had ESP, and I knew
that Time was going to charge $3.00 this coming week, how much would I want to charge?
Then it will have to ask Suppose I had ESP and I knew that Time was going to charge $3.10
this coming week, how much would I want to charge? and so on, for all possible prices that
Time might want to charge. What Newsweek will come up with is whats called a Best
response curve (or reaction function), that indicates what price Newsweek would want to
charge, based on Times price:.

Firms and Markets
Lecture Notes

Lecture Notes 116

Why is it upward sloping? Its probably easiest to think in terms of a price cut: Suppose both
newspapers are priced at $4, and Time decides to cut price to $3.50. Newsweek will probably
want to cut its price, too, to stay competitive, but maybe not all the way to $3.50. Newsweek
might price at $3.88, for instance. Why doesnt Newsweek match Times price cut penny for
penny? Imagine a market of readers: some are very loyal to Newsweek, some prefer
Newsweek, some slightly prefer Newsweek, some dont care, and some prefer Time. Quite a
few of the loyal customers will still buy Newsweek even if Time is a little cheaper. Newsweek
wants to compete with Time for the less loyal customers, but it doesnt want to lose so much
revenue on loyal customers. So it does not cut price quite as much as Time. In the extreme, if
Time were to cut its price to zero, Newsweeks best response would not be to cut its price to
zero: after all, it has a positive marginal cost. Probably it would keep the price a little above its
MC, and still have a few loyal customers.

Meanwhile, Time performs the same calculation: Suppose I knew that Newsweek would
charge ___, what would my best response be? This gives a best response curve for Time.

(In this case weve drawn them as identical, but if Time and Newsweek has different costs or
readers of different loyalty, the curves would be different.)
Newsweeks
Price
$3.00
$4.00
Times
price
Newsweeks best response curve
$4.00
Times Price
$3.00
$4.00
Newsweeks
price
Times best response curve
$4.00

Lecture Notes 117

Now, in order to graph both curves on the same picture, we are going to flip the Times best
response graph around, by flipping the axes:

Now we can draw both curves on the same graph, and find the point that is a best response for
both players:

If Newsweek knew that Time would charge $4.00 this week, Newsweek would want to charge
$4.00; and if Time knew that Newsweek would charge $4.00 this week, Time would want to
charge $4.00. So that is the Nash equilibrium, or stable point. (Notice that both charging $5.00
is not an equilibrium: if Time were going to charge $5.00, Newsweek would want to charge
$4.25.) We expect that prices would settle down here, once they had competed for a few
weeks.

One thing to note, though, is that the prices are lower than if a monopolist owned both Time
AND Newsweek, and were setting the price for each: the monopolist might choose a price of
$3.00 $4.00
Times price
Times best response curve
$4.00
Newsweeks
price
Newsweeks
Price
$3.00
$4.00
Times
price
Newsweeks best response curve
$4.00
Times best
response
curve
$3.00

Lecture Notes 118

$5.50 per magazine, for example. This isnt something you can see from the graph, so we will
give a mathematical example.

Math example:
This is purely for nerds who find it easier to follow once youve seen the math! Ignore this example if it is not
helpful. You are not expected to know this math.

Suppose we have the demand for Time and Newsweek:
q
T
= 5 p
T
+ 0.5p
N

q
N
= 5 p
N
+ 0.5p
T

and we have A and Bs marginal cost: MC
A
= $1, and MC
B
= $1

How do we find Ts best response curve?

Question Time needs to ask: What price should T charge, to maximise her profits, if she knows that
Newsweek will charge a price p
N
? (In maths, we now treat p
N
as a number that we know, like $3.00 or $4.00,
even though we write it in algebra.)

Max Profits
T
= p
T
q
T
- MCq
T
= (p
T
- 1)q
T

This is differentiated Price competition put everything in terms of price

Profits
T
= (p
T
- 1)q
T
= (p
T
- 1) (5 p
T
+ 0.5p
N
)
= 5p
T
p
T
2
+ 0.5p
N
p
T
- 5 + p
T
0.5p
N

= 6p
T
p
T
2
+ 0.5p
N
p
T
- 5 0.5p
N

Remember, profits are maximised at the point where marginal profits are zero (or, equivalently, where MR =
MC)
maximise taking p
N
as a known constant number: find the derivative with respect to p
T
, and find the best
response p
T
* at which its zero (the top of the hill).

Marginal Profits = d(Profits
T
)/dp
T
= 6 2p
T
+ 0.5p
N

to find p
T
* that sets marginal profits equal to zero: 6 2p
T
* + 0.5p
N
= 0
p
T
* = 3 + 0.25 p
N

Thats Ts best response curve.

Doing the same maximisation for Newsweek gives Ns best response curve.
p
N
* = 3 + 0.25 p
T

As before, we solve by finding the Nash equilibrium: p
T
* and p
N
* are a Nash equilibrium if
- T wants to charge p
T
* if N is charging p
N
*
- N wants to charge p
N
* if T is charging p
T
*

When both T and N are at a best response, were at the Nash equilibrium solve the two together:
p
T
* = 3 + 0.25p
N
*
AND p
N
* = 3 + 0.25p
T
*

Lecture Notes 119

And then solve the two together:
p
T
* = 3 + 0.25 p
N
* = 3 + 0.25(3 + 0.25 p
T
*)
= 3.75 + 0.0625 p
T
*
0.9375 p
T
* = 3.75
p
T
* = 3/(0.75) = 4

p
N
* = 3 + 0.25 p
T
* = 3 + 0.25(4) = 4

We said that these are lower prices than a monopolist would charge. A monopolist would choose p
T
and p
N
to
maximize the sum of profits:

Monopolist: Max p
N
, p
T
{ p
T
q
T
- MCq
T
+ p
N
q
N
- MCq
N
}

= {(p
T
- 1) (5 p
T
+ 0.5p
N
) + (p
N
- 1) (5 p
N
+ 0.5p
T
)}

Usual approach: Take the derivative of profits with respect p
T
and set equal to zero:
6 2p
T
+ 0.5p
N
+ 0.5p
N
0.5 = 0
p
T
= 2.75 + 0.5p
N

Take the derivative of profits with respect p
N
and set equal to zero:
0.5p
T
p
N
+ 6 2p
N
+ 0.5p
T
= 0
p
N
= 2.75 + 0.5p
T

Solving these two equations together (as above) gives p
N
= p
T
= $5.50
So prices are higher under a monopoly that owns both magazines.

B. Price Competition in Identical Goods

What happens if a small number of firms compete on price, that is, they produce identical
products and try to attract business by pricing below the competition? In some situations, the
result can be that price is driven down to marginal cost and no one makes any money. This is
particularly painful in industries with high fixed costs: marginal cost is well below average cost
and firms pricing at marginal cost lose money. A good example is airlines, where price
competition in the 1990s resulted in pretty much everyone but Southwest losing money.

The point of this session is that the price-cutting game is a prisoners dilemma: its a bad
outcome, but the outcome is inherent in the game. The outcome is sometime referred to as
the Bertrand trap, named after the 19th-century French economist Joseph Bertrand. The
lesson to you should be: price competition is hazardous. Or as they say in the business world:
avoid commodity businesses.

The Price-Cutting Game

Consider an industry with two firms producing a single homogeneous product. Suppose each
has the same cost structure. Well show that each has an incentive to reduce its price below the

Lecture Notes 120

others as long as price is above marginal cost. The result: progressive price-cutting reduces
the price to marginal cost.

Lets look at this a little more formally. Suppose marginal cost is a constant c and demand for
the product is D(p). If one firm has a lower price than the other, it gets the whole market,
D(p). If they charge the same price, they each get half the market, D(p)/2. (This is arbitrary,
but since its not an essential part of the story we want to keep it simple.)

Now lets look at the incentive to cut price. If each firm is charging the same price, then they
split the market. If the price is above marginal cost, then both firms make a profit. But what if
one firm charges one penny less? Its profit per unit falls (just a little, since the price cut was
one penny), but because it gets the whole market its profits go up. (Profit should almost
double.) Therefore it has a clear incentive to reduce price. The picture changes only when
price falls to marginal cost. At this point, further reductions in price are unprofitable. We
illustrate this graphically in Exhibit 1, where for each firm we graph the optimal price for one
firm given the price charged by the other. We call the lines showing these optimal choices
reaction curves. Theyre analogous to best responses in matrix games. In fact, the only
difference is that price is a continuous variable, not discrete, so its easier to describe strategy
graphically rather than in a table or matrix.

The optimal strategy for each firm, then, is to reduce price below the other, unless price equals
marginal cost. (Also, if the other firm were pricing above monopoly price, then my optimal
price would be monopoly price. In fact, I would be a de facto monopolist, since my rivals
high price effectively places it out of the market.) Since each firm is trying to undercut the
other, we end up with price equal to marginal cost. In game theory terms, this is a Nash
equilibrium, the intersection of the reaction curves of the two firms (Exhibit 1). In economic
terms, we have reproduced the competitive solution (price equal to marginal cost) with only
two firms. Its a remarkable result and an important business lesson: you can have intense
competition even with a small number of firms.

Escaping the Trap

Well spend some time exploring ways to escape the Bertrand trap of competitive price-
cutting, but its worth a quick look now. Heres a partial list:

- Buy your competitor. This has clear anti-trust problems, but you can see the incentive for
firms to do so. Alternatively, one firm might decide to exit (since its not making any
money) leaving the other with a monopoly.

- Collude on price. Also illegal. But what if you and your competitor simply come to
understand that price competition hurts you both? Are there legal ways of interacting that
avoid the hazards of price competition? More on this soon.

- Develop a cost advantage. Suppose you (as Firm 1) have a cost advantage over your
competitor (Firm 2). Then the game plays out like this: you reduce your price to just

Lecture Notes 121

below your competitors cost. At that point Firm 2 drops out and you get the whole
market. Your profit is determined by the difference in your costs. See Exhibit 2. This is
yet another example of the value of low cost.

- Differentiate your product. Generally, products are not homogeneous: buyers perceive
differences between them that moderate the tendency for the low price to take the whole
market. In such cases, the pressures of price competition are less severe, but they are
generally there.

Examples

Airlines. Some observers think that airlines increased capacity too much in the 1990s, buying
new planes and expanding into new markets. With fewer planes, they would have less ability to
expand output and drive prices down.

Telecommunications. In hindsight, too much capacity was added in the late 1990s, much of it
fueled by cheap venture capital. The result (this is a high-fixed-cost, low-marginal-cost
business) was a sharp drop in prices. WorldCom and Global Crossing are just two of the
casualties.

Movie theaters. In the 1990s, capacity expanded dramatically with no obvious market reason.
The result is that prices fell and margins paid to producers rose.

Manufacturing. A few years ago, a senior exec from a New Jersey-based company spoke at the
annual meeting of the Asian Business Society. At the time, many firms were discussing the
possibility of buying Asian companies in or near bankruptcy. In his view, excess capacity in his
industry had led to overly aggressive price-cutting. He mentioned trying to buy three Asian
manufacturing facilities, with the idea of using one and closing the other two to reduce capacity
in the industry.

Numerical examples of competition in IDENTICAL goods

Example 1. You are the sole domestic producer of the generic antidepressant Sensitrum. Your
marginal cost is $2 per dose. Demand is given by Q = 400 50 p (Q in millions of doses, p in
$). There is a second producer in India whose marginal cost is INR 145 (including
transportation cost to the US). Firms set prices simultaneously.
(a) What is your equilibrium profit at the current exchange rate of INR 48 / US$?
(b) An advertising and retailing campaign costing $80m is expected to increase demand by
40%. Should your firm go ahead with it?
(c) One macroeconomics expert tells you that it is likely that the rupee will appreciate in the
near future. How would this influence your decision?

Answer.
(a) First we must determine the Indian firms marginal cost in US$. This is given by 145/48 =
$3.021. In order to determine the price equilibrium, notice that we are in the situation of
homogenous product with different marginal costs. As we saw before, the equilibrium

Lecture Notes 122

consists in the lower cost firm, the domestic firm, to set a price just under foreign firms
marginal cost and get the whole domestic market. The domestic firms profit is (400 50 x
3.021) (3.021 2) = 248.95 x 1.021 = $254m.
(b) The advertising campaign would increase profits by 40% x 254 = $101.6. This is more
than cost ($80m), so do it.
(c) The possibility of the rupee appreciating only makes things better: it raises your
competitors cost. Let e be the spot exchange rate expressed as rupees per dollar. An
appreciation of the rupee means e goes down. This implies that the Indian firms marginal
cost in US$, 145/e, goes up. Since your competitors costs are higher, you can charge
more, your profits are higher, and the benefits of the advertising campaign are higher.

Example 2. You are currently the sole seller of ByeByeCold, a revolutionary drug that almost
instantly eliminates cold symptoms. Although the production cost is only $.10 per dose, you
sell ByeByeCold for $1.39 per dose, for a total profit of $900m a year. You are currently
considering licensing ByeByeCold to a second producer. One of your managers suggested,
since the firm would be sharing the market with a competitor, it would be appropriate to
charge a flat fee that covers half the current profits plus a generous margin; the value of $700m
was suggested. An alternative proposal would be to set a royalty fee of $.50 per dose. What is
your opinion?

Answer. If you license ByeByeCold for a flat fee, you will be competing with another firm
selling the same product and with a similar marginal cost. Except for the possibility of
collusion, this would imply approximately zero profits for both firms. It follows that half the
current profits would not be sufficient to compensate for the profit loss from licensing. In
fact, there exists no licensing contract that would be profitable for both parties.
Why do firms ever license, then? One possibility is that the second firm is more efficient in
production so that there are gains from bringing it on board. Also, there may be reasons why
production by a second firm increases the size of the market. If these were true, then it is
possible that a profitable licensing arrangement can be made.

Lecture Notes 123

Exhibit 1
Reaction Curves in Price Game

Comment: Each firm undercuts the others price, until price hits its marginal cost.

Exhibit 2
Reaction Curves in Price Game When Firm 1 Has Lower Cost

Comment: This time, the low cost firm reduces price to just below the others cost, which is
enough to drive it out of the market.
p
1

p
2

45
MC
2

Firm 2s
reaction
curve
MC
1

p
1
*
= MC
2
-
p
1

p
2

45
MC
2

Firm 2s
reaction
curve
MC
1
p
1
* = c
Firm 1s
reaction
curve
p
2
* = c
Firm 1s
reaction
curve

Lecture Notes 124

Competition and Cooperation

The language of business is frequently based on the language of competition, or even war. In
fact, business has elements that are competitive, and other elements that are cooperative;
elements that are zero-sum, and others that are positive-sum; elements focused on creating
value, others focused on distributing value. Without the cooperative elements, the competitive
ones wouldnt be worth the effort.

This note is devoted to restoring the balance. We will look at situations in which firms are as
much allies as adversaries and explore ways in which they might develop their common
interests. Examples include:

- Microsoft and Windows. Since both depend on the health of the PC market, they have a clear
common interest (which is not to say that their interests are the same).

- Management and labor. Other things constant, a firm makes more money if its labor costs are
lower. But strikes and disagreement between management and labor can shrink the pie and
potentially leave both worse off. Surely one of the reasons Southwest is more efficient
than US Airways or United is that they do not have a history of labor-management strife.

- International trade. The logic of trade, which has been understood by specialists (if no one
else) for almost two centuries, is that both countries benefit. Voluntary trade creates value
that benefits both sides.

- Stern students. There is a narrow sense in which one students success comes at the expense
of anothers: not everyone can get the highest grade. But there is a more important sense
in which students have a common interest: you learn from each other, share insights and
contacts, and the success of your fellow students increases the reputation value of a Stern
degree.

Such combinations of cooperation and competition were dubbed co-opetition by Novells
founder, Ray Noorda. The message is not that you should be nice in business (although
courtesy remains a virtue, even in the 21st-century business world), but that you should
understand (i) when you and others have similar interests and (ii) how to take advantage of
them.

Firms and Markets
Lecture Notes

Lecture Notes 125

Common Interests

Some decisions concern the creation of value, others how to divide it up: making the pie and
fighting over its pieces. Lets start by stressing the creating value part of business
relationships.

Consider the relation between a firm and its suppliers. On the one hand (the economist in us
speaking!), a firm does better if it gets a low price on its inputs. On the other, the firm and
supplier both benefit if the final product does well. Take two airlines, American and Delta. In
one sense, they are in direct (and often quite cutthroat) competition for market share, landing
slots, and gates. Yet these firms also complement each other. How? They both buy planes
from Boeing. These planes are expensive for Boeing to design and produce. If only Delta
purchased planes from Boeing, it might not be enough to allow Boeing to invest in improved
design features. Moreover, aircraft manufacture benefits from a steep learning curve: the
more Boeing produces, the cheaper they can sell them. So if American buys more, there is a
clear indirect benefit to Delta.

There is a similar logic for a firm and its employees. Suppose you are the representative of a
labor union, and I am an executive of a corporation. We are engaged in negotiations over the
next contract. We realize we do not have identical interests. I think that you want an
unreasonable wage hike and excessive vacation time; you think I am an indifferent manager
who wants to deny you any added benefits or share in the firms success. I think any gain you
receive will be at my expense. You think that any denial of your demands results in a win for
me. This is a distributive, zero-sum mindset where we each think there is no way to win
without the other side losing.

But picture a somewhat different scenario. Suppose I (the executive) believe that profits will
be way up in the next five years thanks to efficiency gains in the firms productive capital, and
thus its labor. As a result, I am willing to give quite a bit on wages to reflect these productivity
gains. Nevertheless, I am adamant in opposing any additional vacation benefits, which raise
cost without increasing productivity. Similarly, suppose you (the labor rep) are flexible over
vacation time, but adamant about wage increases. Suddenly, there is a clear win-win possibility.
You and the manager can cooperate on those issues on which you agree.

A more subtle example brings us back to the Bertrand trap of price-cutting: if we are direct
competitors, can we avoid the destructive temptation to undercut each others prices? Clearly
we both gain if we understand that cooperation on price benefits each of us. In this case,
theres a legal issue, too, since getting together to fix the price violates antitrust law in the US
and many other countries. But if we understand our common interest, perhaps we could
approach the same outcome without explicit agreement.

Lecture Notes 126

More Examples

- Sony and Electronic Arts. Sony needs good games to sell its Playstation machines, and EA
needs machines to sell its games. Nintendo recognized the common interest and has
produced both from the start.

- Intel and AMD. Right now, Intel is the standard for PC chips. But it wasnt that long ago
that Intel was trying to carve out a niche for itself. Its major customer in the early 1980s
was IBM and its fledgling PC, and IBM was concerned about sourcing from a single
supplier. To convince IBM to buy their product, they licensed the technology to IBM,
AMD, and several other manufacturers. In this case, Intel realized it had a common
interest with AMD and others to persuade IBM that the supply of chips was competitive,
so it supported them through licensing. Once Intel had a lock on the market, of course,
the common interest evaporated.

- GE and Westinghouse in the 1950s and 60s. A classic case, first, of price-fixing, and later
of clever cooperation that was ultimately judged to be illegal (but still clever!).

- National Basketball Association. There are at least two layers of common interest here.
First, the league needs credible competition to succeed. Its simply not optimal (most
believe) for a small number of big-market teams to win all the time. Among other things,
the primary source of revenue is television fees, which are paid to the league, not to
individual franchises. Second, owners and players have a common interest in the success
of the league. How do we align all these interests? Its a complex situation, but the salary
cap is an attempt to both even the playing field across teams and guarantee that owners
and players share in the overall success of the league (they are guaranteed a minimum
percentage of revenues). Anyone who follows the sport knows this is grossly simplified,
but the fact remains: its critical in a sport to balance the interests of teams, the league, and
the players.

Cooperation as a Nash Equilibrium

Lets continue with the price-cutting game we studied earlier. How might firms learn to avoid
behavior that is mutually destructive?

One answerthe one we will spend our time onis that when firms interact repeatedly over
time, they might be able to avoid the short-run temptation to cut prices or otherwise act in
their narrow self-interest. If two firms compete not only today, but tomorrow and the next
day and so on, this gives each of them leverage over the other. If one cuts prices too low
today, the other can instigate a price war to punish it tomorrow.

Take OPEC as an example to see how this might work. Oil-exporting countries meet
periodically to restrict output, setting output quotas for each country. Say in country X the
quota is 100,000 barrels per day. What if the country actually produced 120,000 to generate
additional revenue? After all, no judicial body enforces the cartel agreement. If others stick to
the agreement, this will keep the price high and increase the benefits of producing more. In

Lecture Notes 127

the language of game theory, country X has an incentive to cheat on its agreement with
OPEC to limit supply. If others cheat (after all, they have the same incentives), output
increases and the price falls, foiling the cartels efforts.

So what stabilizes a cartel? In fact, they tend not to be that stable. But when they work, there
are other sources of leverage at stake. One is that OPEC members play this game over and
over. If a country cheats in one round, then the cartel can retaliate against in the future by
cutting its quota or flooding the market with oil to drive prices down (Saudi Arabia is in a good
position to do this), i.e., it could launch a price war. There is a further difficulty with OPEC,
which is that members may not know exactly what the others are producing. We will return to
this later.

The general point is this: when firms interact repeatedly, the possibility of retaliation or
punishment for non-cooperation can support cooperation as a Nash equilibrium. One
example is the price-cutting: it may be less attractive to undercut your rival if such behavior
triggers a price war.

Related issues:
- Distressed firms. If a firm is worried that it may not survive to tomorrow, future
punishment no longer works as a deterrent. Thus we frequently see distressed firms
starting price wars. This has been a particular problem with airlines, many of which have
been in and out of bankruptcy throughout the 1990s.

- Multi-market contact. Some firms compete in more than one market. This gives them
each more leverage over the other. If they behave uncooperatively in one market, you can
punish them in another. A good example is airlines. A study showed that when airlines
compete on multiple routes, they are less likely to compete aggressively on price. The
media industry is filled with examples of multi-market contact, which presumably gives
them leverage over each other.

- Information. Generally it is harder to punish non-cooperative behavior if you do not
observe it. This is one reason antitrust authorities worry about transparency: it allows
firms to cooperate more effectively. It has been a major source of debate on B2B e-
commerce, since it puts a great deal of information into one or more competitors hands,
which might be used to enforce cooperation.

Legal Issues

Some kinds of cooperation are illegal (price-fixing, for example), others are legal (for example,
most industry standards groups), and others lie somewhere in between.

Lecture Notes 128

Numerical Examples

Example 1. Repeated pricing game.
Suppose that demand in an industry is given by Q = 100 p and there are two firms
producing, each with MC = 20.
(a) Determine monopoly price and profits.
(b) Determine the NPV of future monopoly profits, assuming interest rate r.
(c) Determine threshold value r below which cooperation (cartel pricing) is an equilibrium.

Answer
(a) We begin by writing the inverse demand function for the industry p = 100 Q. Then
the problem for a monopolist with MC=20 is to maximize profits. Taking the derivative to
maximize profits (or equivalently setting MR=MC) yields Q = 40, p = 60 and total profits of
1600. (Check that you can obtain these values.)
(b) The NPV of future monopoly profits is
r r r r
1600
...
) 1 (
1600
) 1 (
1600
1
1600
3 2
= +
+
+
+
+
+

(c) Therefore by sticking to cartel pricing each firm would earn
2
1600
in the current
period and future profits of
r
1600
2
1
in future profits.
If instead a firm chooses to break the cartel agreement, then by undercutting the other firm
slightly it would obtain the entire monopoly profits of 1600 (or extremely close to this value)
but would trigger a future price war in which profits would fall down to zero (recall notes on
the hazards of price competition). Thus the cartel is sustainable so long as the value of sticking
to the cartel agreement
r 2
1600
2
1600
+ is greater than the value that can be obtained by
deviating from the agreement, which is 1600 (+0 in all future periods). Thus the condition that
allows cooperation to be sustained as an equilibrium is:
1600
2
1600
2
1600
> +
r
,
which is equivalent to r < 1.

NOTE: This might seem like a non-binding constraint (r < 100%). However, we must recall
we (implicitly) made the assumptions that there are only two firms, that they are identical, that
there is no product differentiation (punishment profits = 0), that prices are perfectly
observable, and so forth.

Example 2. Secret price cuts (challenging example).
Consider the same set up as in Example 1 and the particular case when r = .3. Suppose,
however, that each firm cannot observe the rivals price (ready-mixed concrete might be an
example). All that a firm can observe is its own price and its own demand. This creates a

Lecture Notes 129

difficult inference problem, since zero demand can result from two causes: (a) rival
underpricing and (b) low market demand. Low market demand occurs with probability 20%,
independent from period to period.

Suppose firms revert to marginal cost pricing whenever they receive zero demand. Show these
strategies are a Nash equilibrium. Is this the best equilibrium? If not, what is a better
alternative?

Answer.
As before, monopoly profits in a given period are 1600 (when demand is not low). Assuming
that firms remain at monopoly price while demand is high, the NPV of future monopoly
profits is now given by

( ) ( ) 2560
5 .
1280
3 . 1
8 .
1
1
3 . 1
1600
8 . ...
) 3 . 1 (
1600
8 .
) 3 . 1 (
1600
8 .
3 . 1
1600
8 .
3
3
2
2
= =
+
|
.
|
\
|
+
= +
+
+
+
+
+

If the firm undercuts the rivals price today, it gets an expected profit of (.8) 1600 = 1280
today, and zero thereafter. The no-deviation constraint is thus
1920
2
2560
2
1600 ) 8 (.
1280 = + < ,
which is clearly satisfied.

Or to put it another way. We check for Nash equilibrium by seeing that for a combination of
strategies, no firm want to deviate. Well, here, what does the firm get by sticking to the
strategy?

With a 20% probability demand is low in which case it earns nothing today (since there is no
demand) and given that firms revert to marginal cost pricing following a low demand
realization, the firm will get nothing in the future. With 80% probability demand is normal and
the firm earns half the monopoly profits 1600/2=800. Further in the next period, it has the
chance to earn money, again by sticking to its strategy then with 20% probability there is low
demand and it earns nothing this period or for evermore, but with 80% probability earns half
the monopoly profits and continues on potentially earn more profits in the future.
The NPV of sticking to the strategy can thus be written down as:

0.2*0+0.8*(800+1/1.3*(0.2*0+0.8*(800+1/1.3*(0.2*0.+0.8*(800+1/1.3*(0.2*0.+0.8*(

which we can re-write more simply as 0.8*800+0.8/1.3*800+0.8^2/(1.3)^2*800+.=1920

Instead if the firm deviates and undercuts then it gets 0 with 20% probability and with 80%
probability the full monopoly profits today and nothing for evermore, that is it earns an
average of 1280.

Lecture Notes 130

Since 1920>1280, the firm has no reason to deviate. It is indeed playing a best response (and
since the game is symmetric) so is the other firm. If both firms are playing strategies that are
best responses to each other, then this is a Nash equilibrium.

There is one important difference with respect to the previous example. The equilibrium
strategies call for reversion to marginal cost pricing whenever a firm observes zero demand. In
the previous example, this never actually takes place, since both firms anticipate a price cut will
lead to an indefinite price war. By contrast, in the present example zero demand may simply
result from low market demand. Since the firms have no way to tell what causes low demand,
they will revert to a price war when in fact there is no culprit to punish.

You might think firms would be better off by not starting a price war, since they know that, in
equilibrium, zero demand is caused by market conditions, not by defection. However, if a firm
expects its rival will not start a price war then it has an incentive to defect, so that cant be an
equilibrium.

Although the above strategies form an equilibrium, there may be better equilibria. At some
point, market demand will be low and firms will plunge into an indefinite price war. This
seems a little too much, especially in light of the fact that the no-deviation constraint has a lot
of slack (1280 < 1920). So, instead of an indefinite price war, suppose that, whenever a firm
observes zero demand, a finite price war starts. Given the slack in the no-deviation constraint,
it must be possible to have finite price wars and still avoid the temptation of price cutting.

Example 3: Tit-for-tat in Iran-Iraq game.
Consider the following game in which Iraq and Iran must determine how much to produce.

2 4
42 44
2 46 26
22 24
4 52 32

Suppose this game is infinitely repeated and that the interest rate is 50%. Suppose moreover
that players start with Q
1
= 2 at t = 1, then Q
t
= rivals Q
t-1
(tit-for-tat) thereafter. Show that
these strategies form a Nash equilibrium. (Hint: consider a one-time deviation from
equilibrium.)

Answer.
If Iran and Iraq play the equilibrium strategies, then Iran will get

138
5 .
46
46 ...
) 5 . 1 (
46
) 5 . 1 (
46
5 . 1
46
46
3 2
= + = +
+
+
+
+
+
+
Iraqs output
Irans Output

Lecture Notes 131

This follows, since its strategy starts by choosing 2 and so does Iraq, then since Iraq in the
second period does what Iran did in the first, and Iraq does what Iran did in the first, both
again choose 2, and so for the third period etc.

Suppose Iran deviates in the first period by producing high output and sticks to tit-for-tat
thereafter. Then at t = 2, Iraq will choose high (tit-for-tat), whereas Iran will be choosing low
output. At t = 3, the situation is reversed: Iraq chooses low output and Iraq high output.

Given this sequence of play, Iran anticipates an NPV of

8 . 124 2 . 31 6 . 93
25 . 1
25 . 2
5 . 1
26
25 . 1
25 . 2
52
) 5 . 1 (
1
1
1
5 . 1
26
) 5 . 1 (
1
1
1
52
...
) 5 . 1 (
26
5 . 1
26
...
) 5 . 1 (
52
52 ...
) 5 . 1 (
52
5 . 1
26
52
2 2
3 2 2
= + =
= |
.
|
\
|
|
.
|
\
|
+ |
.
|
\
|
=
=
|
|
|
|
.
|
\
|
+
|
.
|
\
|
+
+
|
|
|
|
.
|
\
|
+
=
=
|
|
.
|
\
|
+
+
+
+
+
|
|
.
|
\
|
+
+
+ = +
+
+
+
+

which is less than 138.

For Iraq, the equilibrium NPV is

126
5 .
42
42 ...
) 5 . 1 (
42
) 5 . 1 (
42
5 . 1
42
42
3 2
= + = +
+
+
+
+
+
+

If Iraq deviates in the first period and then proceeds to play tit-for-tat, it expects an NPV of

6 . 105 4 . 26 2 . 79
25 . 1
25 . 2
5 . 1
22
25 . 1
25 . 2
44 ...
) 5 . 1 (
44
5 . 1
22
44
2
= + =
|
.
|
\
|
|
.
|
\
|
+
|
.
|
\
|
= +
+
+
+
+

which is less than 126.

We might try other deviation strategies, but it will still be the case that the best a player can do
against tit-for-tat is to play tit-for-tat as well. In other words, tit-for-tat by both players
constitutes a Nash equilibrium.

Is tit-for-tat the only equilibrium? Suppose that Iran chooses low output during the first 10
periods, high output thereafter. Is tit-for-tat Iraqs best response to Irans strategy? Would
Irans strategy and Iraqs best response for a Nash equilibrium?

Lecture Notes 132

Example 4. Finitely repeated game.
Let us continue with the Iran-Iraq game, assuming now that it is repeated for 10 periods. What
is the (natural) Nash equilibrium of this game?

Answer.
The best way to approach finitely repeated games is to think of them as a big game tree. As we
saw before, the way to solve these games is to start from the end and move backwards. So, let
us consider the very last period. Since there is no future to worry about, players treat this as a
one-shot game. Since this is a prisoners dilemma, we know what the equilibrium is: both
players choose high output. Consider now the period before last. Since players anticipate that,
no matter what, they will choose high output in the very last period, they treat the second-to-
last period also as a one-shot game. It follows that both choose high output. And so forth.
We conclude that the only equilibrium is to choose high output in every period.

(Technical note for aficionados: generally speaking, extensive-form games, including finitely
repeated games, may have more equilibria then the ones we find by solving the game
backwards. The additional equilibria, while satisfying the requirements for a Nash equilibrium,
are however not reasonable in the sense that they implicitly assume players promise to play
future actions that, when the time comes do take them, are not in their best interest. We will
return to this in the chapter on commitment.)

Lecture Notes 133

Commitment, Entry and Exit

One of the key areas of strategic decision-making relates to entry and exit in competitive
markets. A good example is the competition between Boeing and Airbus in widebody aircraft.
When they each decide whether or not to enter a given market segment, they realize that theyd
each be better off if its rival decided not to enter. So who enters? As in other situations of
strategic behavior, firms facing entry and exit decisions must attempt to see this game
through the eyes of their rivals and reason backwards to decide on an optimal strategy. In
some cases there is a clear first mover advantage, in others not. Either way, timing is critical.

Commitment

One aspect of timing is commitment: the ability to commit to a decision before your rival has
acted. For example, suppose Boeing could persuade Airbus that it was entering a given market
segment. If Airbus believed this, it might choose not to enter. We will see that in many cases,
a firm that makes a credible, irrevocable commitment to entering can increase its expected
profits by convincing rivals to stay out. We would say that commitment is valuable.

Commitment consists of reducing your options, since the idea is to lock into a decision now.
Why would that be valuable? Arent options good? To see why, consider an example (this is
becoming a clich, and may not even be true, but it makes a good story). When in the 16th
century Spanish conquistador Hernn Corts arrived in the New World (what is now Latin
America), he faced long odds. With only a small army backing him, he intended to conquer
and colonize a large population of unknown, potentially hostile indigenous people. Upon his
arrival in Mexico, he made the seemingly insensible move of burning his entire flotilla of ships
(except for one), ensuring that the local people observed this act. Why burn the ships? After
all, he was essentially destroying his escape strategy. He was also destroying assets that could
prove valuable from a military point of view. Was Corts insane or what?

Corts probably was crazy to some extent. (If you are ever in Mexico City, go see Diego
Riveras murals in the Mexican government headquarters where he unflatteringly portrays
Corts as a hideous-looking, shriveled-up pasty-white demon.) Nevertheless, Cortss ship-
burning was not crazy; it was most probably a calculated, rational move. He was
demonstrating his unwavering commitment to colonizing the New World. By burning his
ships in plain view, he was saying in a matter of words we are here to stay; and we will either
conquer you or die trying to do it. This commitment was credible because he was destroying
his ability to undertake any other course of action besides staying and fighting.

Note that commitments must be credible in order to work. If Hernn Costs decided to signal
his commitment by sending the Mexican natives a message stating we will fight, this would
not have much effect. After all, we know that talk is cheap. Similarly, any firm can make a
Firms and Markets
Lecture Notes

Lecture Notes 134

public statement of its intentions; but only by acting on these commitments in an observable
and credible way that is hard or expensive to retract can it ensure its commitment has an
enduring effect.

Preemption

Why does DR build stores just two blocks away from each other in Manhattan? Arent the
stores stealing each others market share? Firms can create barriers to entry by increasing their
capacity or product line in the hopes that they will capture demand that otherwise could be
appropriated by a potential entrant. We can say more formally that a preemption strategy by a
firm is one that makes entry by a new firm unprofitable. This is different from an unrestricted
monopoly strategy because, as the Duane Reade example suggests, it may intentionally sacrifice
profitability in exchange for controlling the market.

Predation

Another related dimension is predation: If you cant keep em out, drive em out. Predatory
pricing has a long and infamous history in the United States, ranging from Standard Oil to
railroad companies to AT&Ts longstanding clench of the telephony market. Predation occurs
when a firm prices its products below cost to force others out of the market or preventing
their entry. You may wonder why a firm would ever wish to price below cost and lose money.
In a static world, this would make no sense, but in a dynamic setting, predatory pricing can be
an effective deterrent.

There remains, however, a range of opinions about whether it should be a serious concern for
competition policy. At one extreme there is the so-called Chicago school. According to the
Chicago school, predatory strategies cannot be profitable; if a firm prices low, then this must
be considered a competitive, not anti-competitive, strategy. There are however several
explanations of why below cost might be a successful predatory strategy:

- Asymmetric information. If firms know more about their costs than do the public and rival
firms, they are able to exploit this fact by pricing low to convince rivals they are very
efficient and not worth competing against.

- Reputation-building. It can be rational to establish a reputation for irrationality. If firms
signal that they are willing to act irrationally from an economic standpoint to dominate a
market (after all, pricing below cost is not a rational economic decision, at least not in the
short run) they may create a reputation that scares firms off from competing with it.

- Learning curves and network effects. Firms can reduce cost as output rises through learning
effects and other means. Anticipating these effects, firms may price below cost in the
present to stave off rival entry. Once they attain their cost reductions, later they can return
to profitabilityespecially if they have driven rivals out of the market.

Lecture Notes 135

- Deep pockets. If a firm has more cash than its rivals, pricing below cost leads to a war of
attrition where the winner is the firm with deeper pockets. (Although we list this in the
fourth place, it is likely the most common reason why predatory pricing works.)

The above discussion notwithstanding, we should add that predation is not always so clear-
cut. (Look at the testimony of the Microsoft trial if you need convincing!) Moreover, it is
important to note that pricing below cost is not the only way for a firm to behave in a
predatory fashion. Firms can also enlist predatory strategies such as capturing demand through
product design (eg, MS Windows) and through striking exclusive contracts with other
members of the value chain that put them at an advantage with respect to the competition.

Why is predation illegal and bad for consumers if it leads to lower prices? Recall our earlier
discussions of monopoly firms. Once predatory firms drive other firms out of the market,
they can increase prices and act without regard to competitive dynamics that ensure
competition and choice for consumers.

Examples

Example 1. The word is out: a revolutionary discovery by a faculty member at NYU's Courant
Institute for Mathematical Sciences will finally allow the practical implementation of parallel
processing in personal computers. As holder of the patent, NYU has licensed the technology
to two firms, to assure PC manufacturers of two sources of supply. One firm, located in the
US, has already moved ahead with engineering plans for building a production facility. The
question is whether to go for a plant of size 1, 2, or 3. The second firm is located in Brazil. It
will also have to decide whether to build a plant of size 1, 2, or 3. However, it is a bit late with
respect to the US firm, and wont be able to move ahead before next year. When it does so, it
will know what plant size the US firm has opted for.

You have been hired as a consultant to the US firm. Your analysis shows that the profits from
various combinations of plant sizes are:

1 2 3
3.9 5.8 5.7
1 3.9 2.9 1.9

2.9 3.8 2.7
2 5.8 3.8 1.8
1.9 1.8 -0.3
3 5.7 2.7 -0.3

Question: As a consultant for the US firm, what size factory do you recommend?

Answer: As indicated, the US firm chooses output this year, whereas the Brazilian firm only
chooses output next year. We thus consider an extensive-form game, with the US moving
Brazilian firm
US firm

Lecture Notes 136

first. Suppose US chooses 1; then Brazil chooses 2 (which has the highest payoff) and the US
payoff is 2.9. Similarly, if US chooses 2, Brazil chooses 2 and the US payoff is 3.8. And if US
chooses 3, Brazil chooses 1 and the US payoff is 5.7. The last (3) generates the highest payoff,
so the US should build a plant of size 3.

Question: When you meet with the Executive Committee of the US firm, one executive
remarks: There's no need for haste. The Brazilians will not build a factory until next year, so
why dont we wait and build our factory next year as well? How would you reply?

Answer: The situation described above gives the US a large advantage: by choosing/moving
first, it can force Brazil to choose a low output level. If the US firm waits, then it will either be
choosing output at the same time as the Brazilian firm or after the Brazilian firm. If the two
firms choose capacities at the same time, the Nash equilibrium is (2,2) and the US firm has a
payoff of 3.8, substantially less than 5.7.

Lecture Notes 137

Imperfect Information

Many business situations are plagued by information difficulties, in which one party to a
transaction has better information than the other. (Think of the phrase: Trust me on this.)
We refer to these situations as having asymmetric information. Inevitably, the party with
superior information is tempted to exploit its advantage. And even if it doesnt, the other party
may think it will. And the uninformed part has some leverage, too: the possibility of walking
away from a deal may offer it some advantage. The result is one of the most interesting and
challenging aspects of business.

Consider these examples:
- An employer wants his sales team to work hard but cannot observe whether or not they are
working hard. What steps can he take to ensure that they work hard?

- Health insurance company faces pricing dilemma. A low price leads to small margins. But
a high price runs the risk of attracting only high-risk patients. What should it do? We refer
to its problem as adverse selection, since the high-risk patients are more likely to select
the plan. The problem is that the health of a patient may be known better by the patient
than the insurance provider.

- Large firm would like to spin off subsidiary. Potential buyers ask: Is this a strategic move
(the firm wants to concentrate on other businesses) or an attempt to unload what it knows
to be an unprofitable business?

- Price as signal of quality. New firm enters market for luxury luggage. It would like to
reduce price to gain market share, but worries that this might be interpreted by customers
as a sign of low quality.

Information can be a serious friction to doing business and to prevent markets from working
efficiently. We will first explore some of the ramifications of asymmetric information in
stylized settings where the issues are clear, even if unrealistic. We later return to reality and
discuss ways in which information problems might be ameliorated.

Games with Imperfect Information

Games with imperfect information are games where one player has better information than the
other (in particular following some hidden action).
We begin by considering an agency model, where a principal (employer, board of directors,
regulator etc), who would like to induce some kind of good behavior from an agent, cannot
directly observe the agents actions. In this case, known as moral hazard, (hidden action) the
agent has better information only with respect to the action that he takes and that the principal
Firms and Markets
Lecture Notes

Lecture Notes 138

does not observe. The key in games of moral hazard is that even though the principal might
not observe the agents actions, she might be able to observe outcomes that are related to
actions (e.g. output, an observable, is a function of effort, an unobservable). Then the
principal might want to set up a reward system based on observablesthough she should
beware about what kind of behavior this encourages.

A different kind of information problem arises if we suppose that before the game even starts
one of the players is relatively well informed (hidden information). We can find many examples
of such asymmetric information in day-to-day life. When people apply for health insurance
they often know more about their health status and history than the insurer. Manufacturers of
products often know more about the quality of their product than do consumers. Firms
presumably know more about their assets than potential buyers. In a legal setting, defendants
often know more than plaintiffs. And in Edgar Allen Poes Telltale Heart, the guilt-ridden
criminal knows about his crime while the investigators do not. (And, if he had kept his cool,
he would have probably remained a free, yet disturbed, man.)

We can consider two polar cases of asymmetric information. First, we can consider a game
where the relatively less-informed player moves first. In this case, adverse selection becomes
a risk for that first mover. Consider, for example, the case of litigation for medical
malpractice. Before the trial proceeds to the discovery stage, it seems reasonable to assume the
defendant has better information than the plaintiff. If the plaintiff makes the defendant an
offer for an out-of-court settlement, he should take into account that such offer will be
accepted if and only if liability is likely to be greater than the proposed settlementthe exact
situation in which the plaintiff would want the offer not to be accepted.

Second, we could consider a game in which the informed player moves first. In this case, the
first move may tell the other player something about the first players information, and so the
first player will take this into account. If a firm knows the quality of its product, could a buyer
infer quality from the sellers price? Typically the move by the first player contains information
that the second player can use to make a better decision. The first player knows this, of course,
and uses it to guide its choice.

The takeaway point on hidden information: the uninformed player must guess how the
informed player will act react. Who is likely to take up this kind of insurance? When the
informed player moves first, the move conveys information: if this is really such a good deal,
why is it being offered to me? The trick is to convey the right information.

Lecture Notes 139

Agency problems

There are numerous examples where an agent takes actions that cannot be observed by a
principal. The actions chosen by this agent will be determined by the incentives that he faces and
so the principal might try to structure transactions so that these incentives lead to the behavior
that she prefers.

The most obvious example and (in the light of the Enron debacle) perhaps most notorious is
pay for performance; however, similar issues are at play in numerous other examples. For
instance, if an insurance company sells a factory owner insurance against fire it will wonder
whether the factory owner (now that she has insurance) will do all that she can to prevent the
outbreak of a fire; or, if a commercial bank decides to lend money to an entrepreneur, it will
wonder whether the entrepreneur will be prudent or gamble the funds hoping for a big win
and spectacular IPO.

In many applications where explicit incentives are used, an important consideration is that the
incentive scheme might subject the agent to considerable risk. For example, performance pay
leads to higher variance in pay than fixed compensation. If, moreover, the agent is more risk-
averse than the principal (e.g., worker and corporation), then we have a trade-off: efficient risk
sharing suggests shielding the agent from risk but unless some risk is imposed on the agent
there is no motivation.

Another widely discussed and prominent issue is a concern about misaligned incentives, or, as
Steven Kerr nicely put it, the folly of rewarding A, while hoping for B. In his beautiful
paper, he mentions numerous examples, to which we add the perhaps apocryphal one from the
planning days in the Soviet Union: A particular nail factory was subject to an annual
production quota of 1 ton of nails. It fulfilled this requirement by spending a week
manufacturing a single 1 ton nail! In short, incentives work but expect to get exactly what you
pay for.

We now turn to consider problems of asymmetric information where one of the players is
relatively well-informed even before the game starts.

Selling Lemons

Suppose that the problem is one of hidden information and that the uninformed player moves
first; what considerations will guide her strategy? The classic example is the sale of a used car.
It made more sense in the 1970s than it does now that car quality is uniformly higher. But the
idea is that when you buy a used car from an individual, you probably know less about the
quality of the car than the seller. You move first in the sense that you dont observe any
informative moves by the seller before you make an offer. The risk is that you might buy a
lemon (meaning a low-quality car).

Consider a concrete example. The buyer makes an offer and the seller decides whether or not
to take it. Suppose that you, a potential buyer of a used car, know from reading Consumer
Reports that 50% turn out to be lemons. You decide that a lemon is worth zero and a non-
lemon is worth $1,500. You might then set your reservation price (the most you are willing to

Lecture Notes 140

pay) at $750 based on the 50% probability of buying a dud [750 = (.5x0) + (.5x1500)].
Anything more, and youd prefer to walk away. What about the seller? Let us say a seller
would be unwilling to part with a non-lemon for less than $1,000, the fair market value based
on the blue-book value of the car, etc. Now we have a problem. Given your lack of
information about whether you will get a lemon or not, you are willing to pay less for the
model than the reservation asking price of legitimate sellers. This means that anyone willing to
sell you a car for your price is probably selling you a lemon!

It only gets worse. Fearing that you will get a lemon at $750, you change your reservation price
downward because you think the likelihood of getting a worthless lemon is now 90%. Thus,
the new bid is $150 [=(.9x0) +(.1x1500)]. Now who on earth would sell you a car that, worth
$1,000 without defects, sells for only a fraction of this price? Only people selling lemons. As
this downward spiral continues, both buyers and legitimate sellers drop out of the market until
the market vanishes. This happens because buyers do not have the information they need to
make risk-free offers based on fair market values.

Its a classic example of bad products driving out the good: if its impossible to demonstrate
quality directly, no one would be willing to produce and sell high-quality products at the price
buyers (who dont know quality) will pay. Viewed from the perspective of game theory, we ask
ourselves the basic question: Why is the seller willing to part with the product? If we fear that
its because he knows that its of low quality, we decide not to buy and the market collapses.
We would say (formally) that the only Nash equilibrium is when no transactions take place.
This extreme outcome is the result of the parameters we started with, but more generally the
difference in information will lead some mutually beneficial transactions not to take place.
Asymmetric information is a barrier to beneficial economic transactions.

The takeaway point is that the absence of good verifiable information makes it difficult to
conclude transactions that are valuable to both sides. If only the owner of the non-lemon car
could convey that information to the buyer in a convincing way. In practice, there are many
institutions that play that role: warranties, legal systems, third-party ratings (Moodys and S&P,
Consumer Reports, Zagat Guides), the sellers reputation, free trials, and so on. Protecting the
integrity of these systems is often in the interest of sellers. For example, financial service firms
in the US have been the beneficiaries, by and large, of SEC regulation designed to protect
buyers.

Signaling Quality

The other application of asymmetric information we can consider is when the informed party
moves first. In this case, the informed player must understand that her action will convey
(signal) information to the less-informed party. For example, price is considered an
important signaling mechanism. If you choose to bring a bottle of wine to a friends house for
dinner and know very little about wine, you might buy a $20 bottle of wine instead of the $7.99
special. (Of course, beer is always good.) As a non-expert, you may not know how the two
taste, but clearly the more expensive one seems a safer bet. The vintner in this case is signaling
information that she possesses and you dont about her wines quality. Similar things can occur

Lecture Notes 141

with other products: stereos, consultants, surgeons, investment banking services. The key is
that the seller has demonstrably better information about the product than the buyer. Thats
one of the hazards about using a low price to gain market share: it might convey to buyers that
the product is low quality.

To understand this type of game better, suppose that there is one seller of stereos and many
buyers. The seller sets one price and buyers decide whether to buy. Only the seller knows the
true quality of the stereos many components. In some theoretical examples, we would find
that only a high-quality producer would charge a high price: If a stereo producer used high
quality parts, and those parts cost more, she should set a higher price. If she is using low
quality parts, which cost less, he should set a lower price.

How can we explain that, in equilibrium, price reflects quality? First, a high-quality product
would not be profitable at a lower price because its parts cost more; these costs must be
recovered via a higher price. But why would a low-quality producer not want to (falsely) signal
high quality setting a higher price? In many cases this will limit your market: fewer consumers
are willing to spend more for high-priced stereos to feel more secure about quality (or not
everyone has enough income to do so). Thus, by pricing your cheaper goods at a higher rate,
you would lose the larger market of people looking for more affordable equipment. Based on
what we know about revenue (price times quantity) we can see that we may very well lose more
in quantity by selling high than we gain from the higher price. In addition, as people found out
in the long run that your stereos were not worth their higher price, you would lose your market
to the higher-quality competition.

In short, there are games in which the equilibrium is for the high-quality product to sell at a
higher price than the low-quality product, which allows uninformed buyers to infer quality
from price. This works if the signal has a cost (if its free, the low-quality producer will follow
suit) and the cost of that signal is lower for high-quality producers. (In this case, higher
production costs imply lower costs of signaling high quality with high price.) In practice, other
factors play a role, too: the reputation/brand of the seller, warranties (which are presumably
more costly to the producer of the low-quality product), reviews by consumer product firms
and other buyers, and so on.

Historical Note

The 2001 Nobel Prize went to three economists who were instrumental in developing the
theory of economic behavior in situations of asymmetric information. George Akerlof is
responsible for the lemons model. Mike Spence developed the signaling model. In his
example, students used education to signal ability. We prefer to think of education having
value in its own right, but we suspect you can see his point. Joe Stiglitz applied the theory to
capital markets; in particular, to an entrepreneur trying to finance a new venture.

Lecture Notes 142

Examples

Example 1. Sale of business. Suppose a firm owns a business unit that it wants to sell. Potential
buyers know that the seller values the unit at either $100m, $110m, $120, $190m, each value
equally likely. The seller knows the precise value, but the buyer only knows the distribution.
The buyer expects to gain from synergies with its existing businesses, so that its value is equal
to sellers value plus $10m (there are gains from trade.) The buyer must make take-it-or-leave-it
offer at some price p. How much should the buyer offer?

Answer.
We can write down a table, which summarizes, for each offer that the buyer makes the
probability that the offer gets accepted, the expected value (to the buyer) conditional on having
the offer accepted and the overall expected profit from any given offer:

Offer Probability of Acceptance Expected value (to buyer) if
accepted
Expected profit
100 10% 110 1.0
110 20% 115 1.0
120 30% 120 0
130 40% 125 -2.0
140 50% 130 -5.0
150 60% 135 -9.0
160 70% 140 -14.0
170 80% 140 -20.0
180 90% 150 -27.0
190 100% 155 -35.0

The seller should offer either $100m or $110m.

Discussion. Suppose the buyer offers p = 100 (in $m). Then, in most cases offer is rejected.
Specifically, 90% of the times the offer is rejected. Offering more would imply a higher
probability of sale, but the expected value of the unit would increase by less than the price
paid. The intuition for this result is the force of adverse selection: the seller will only sell the
unit if its value is relatively low.

Example 2. You work for Pepsi. The company has just signed a major star endorsement. You
must decide how much to spend on your Summer ad campaign: $1m (Lo) or $2m (Hi).
Net profits (in $m) depend on how much you and Coke spend and on whether or not Coke
has signed a major star:

Lecture Notes 143

High Low
Hi
3 2
0 1
Lo
4 5
1 2

Cokes decision of whether to sign a major star has already been taken, but you do not know
what the decision was. Your CIO tells you that there is a 70% chance they did. You also
know that Coke will have a chance to react to your decision of how much to spend. Should
you go for a $1m or a $2m campaign?

Answer.
This is an extensive-form game with incomplete information. Let us solve it backwards and
include Nature as a player. If Coke did sign a star, then it will choose $2m if and only if Pepsi
chooses $2m. If Coke did not sign a star, then it will choose $1m regardless of what Pepsi
chooses. Now, moving backwards: If Pepsi chooses $1m, then Coke will choose $1m. Pepsis
expected payoff is 70% 2 + 30% 3 = $2.3m. If Pepsi chooses $2m, then Coke will choose
$2m with probability 70%, $1m with probability 30%. Pepsis expected payoff is 70% 0 +
30% 6 = $1.8m. Pepsi should therefore choose $1m.

(Additional question: What considerations are we leaving out of the analysis?)

High Low
Hi
0 1
4 6
Lo
2 3
2 3
P
Cokes advertising Cokes advertising
(a) Coke signed a major star
(b) Coke didnt sign a major star
no star (30%) star (70%)
2 1 2 1 2 1 2 1
1 2
no star (30%) star
70% 0 +
+ 30% 6 = 1.8
70% 2 +
+ 30% 3 = 2.3
2
5
1
4
3
3
2
2
1
2
0
3
6
1
4
0
Pepsis
Advert.
C C C C
N N

Lecture Notes 144

Additional examples of asymmetric information

Health insurance. Many policies exclude preexisting conditions. Why? Other policies offer
better rates to groups. Why?

Life insurance. Most policies require a physical. Why?

Downsizing. Firm must reduce payroll by 10%. But if it offers a standard package to all of its
employees, the best employees are the most likely to take it. Again, adverse selection. How
might it reduce its payroll without losing its most productive employees? For example, Merrill
Lynch announced such a plan in October of 2001. The plan allowed employees to apply for a
severance package, but the company could decide which applications to accept. Why do you
think they structured the plan this way?

Ethnic business ties. Some businesses involve important judgments about quality. The
diamond business, for example. We sometimes see that such businesses are dominated by
members of a single ethnic group. Why do you think this is?

American Expresss spinoff of Shearson. In 1993, American Express sold Shearson to
Primerica (now part of Citigroup). In the March 9 Wall Street Journal: Among the sticking
points in acquiring Shearsons brokerage operations would be the firm's litigation costs. More
than most brokerage firms, Shearson has been socked with big legal claims by investors who
say they were mistreated, though the firm has made strides in cleaning up its backlog of
investor cases. In 1992s fourth quarter alone, Shearson took reserves of $90 million before
taxes for additional legal provisions. When the deal was completed, Primerica bought most
of Shearsons assets but left the legal liabilities with American Express. Why do you think the
deal was structured this way? Was it fair to American Express?

Lecture Notes 145

Externalities

Do you normally order salad but go for the lobster when splitting the bill?

This might not seem like a key strategic issue or one of crucial economic importance, but the
same logic that drives diners to spend more when splitting the bill underlies numerous
applications from traffic to pollution abatements and large multi-divisional firms.
3

Externalities

Underlying the diners dilemma is the fact that, when splitting a bill, some of the costs of a
more expensive menu choice are borne by the other diners. The latter find themselves in the
unenviable position of suffering the consequences of someone elses actions without the ability
to do much about it.

Formally, when a persons utility or a firms profits depends on the actions of someone else, we
say that the latter exerts an externality on the former. Examples abound; perhaps more
common are negative externalities where the action of one party imposes costs on another
party (when I smoke, people around me suffer) but there are plenty of examples of positive
externalities when the action of one party benefits another party (when I plant flowers in my
yard, everyone who passes by benefits).

Externalities are prominent in organizations. In a large multi-division organization, for
example, division managers and heads are often rewarded only for the performance of their
own divisions, and so might not take into account the impact of their actions on the firm as a
whole. Even at a smaller scale, giving strong individual incentives to salesman can lead them to
go after each others sales and to ignore the negative externalities imposed on other salesmen
and the organization as whole.

Public Goods

An extreme case of an externality is what economists call a public good, where consumption by
one party in no way hinders consumption by another party. This sounds obscure, but consider
examples such as clear air, parks and national defense.

The public good has the possibility of exclusion if we can control the parties that can enjoy the
good. For example, we can restrict access to parks but it is hard to stop residents of some area
from enjoying clean air, if clean air is provided.

3
Regarding the restaurant problem, a recent study found that on average diners spend close to 40% more
when splitting the bill. See Gneezy, U., E. Haruvy, and H. Yafe. The inefficiency of splitting the bill: A
lesson in institution design The Economic Journal, April 4004, 114, 265-280.
Firms and Markets
Lecture Notes

Lecture Notes 146

Given that many people can enjoy a public good simultaneously (indeed by definition infinitely
many people can enjoy the good) providing a public good is therefore an activity with
considerable positive externalities associated with ita cure for a disease or a piece of music
which can bring pleasure to many. An important policy challenge is to ensure the provision of
such public goods.

Externalities and Market Failure

When externalities are present, the price of a good need not reflect its true social value, and so
firms may produce too much or too little. It follows that the equilibrium market outcome is
inefficient. Recall that an important assumption in our discussion of market efficiency (The
Fundamental Theorem) is that one persons consumption decision does not affect the utility of
any other agent and it is precisely this assumption that fails when externalities are present.

Overcoming Market Failures

Market failure in the presence of externalities provides a clear rationale for government
intervention or other market non-market mechanisms:

Social norms. Even in the case of the diners dilemma and splitting the bill, frequently we dont
order the Beluga caviar, Dom Perignon Champagne or whatever the most expensive item on
the menu happens to be. Why not? Often going out to restaurants is a repeated game: we
tend to go out with the same groups of friends or colleagues, and so a norm of behavior (with
a consequent threat of punishment) can arise.

Similarly, my front garden, if well tended, generates positive externalities for my neighbors as
do theirs for me. If each of us maximizes his own private utility, each of us might spend less
on our front gardens than is socially optimal, but it might be that we fall into some convention
or arrangement whereby if I let my garden go to seed then my neighbors will exert social
pressure on me to do something about it.

Explicit Regulation and Intervention. Everyone in New York knows (though they may have
different views about it) how Mayor Bloomberg (and increasingly many others) chose to deal
with the externality imposed by smokers on those around them: to ban smoking. Though
possibly (probably?) a measure that improves social welfare, notice that such explicit regulation
does not fully maximize social welfare. Suppose that some fabulously wealthy person came to
town and really, really wanted to light up in a small bar, to the point where he was willing to
pay $1m to each person in the bar. It is far from clear that we are maximizing social welfare by
stopping him from doing so. If there were some way, however, of cheaply instituting and
enforcing property rights over the air in the bar, then the situation might be improved. For the
air in the bar, there probably is not a market, but in other applications there might be, which
leads us to the next measure.

Property Rights and The Coase Theorem. A final method for dealing with externalities involves the
establishment of clear and unambiguous property rights. Consider the case of a steel plant
dumping waste in the river. Society might decide that (a) downstream parties have a property
right to clean water, or (b) that the plant has the right to dispose of its waste as it sees fit.

Lecture Notes 147

Property rights can be established in either way. But once those rights have been clearly
established (and so long as they are properly enforced), individuals can bargain over how they
exercise these rights. If downstream parties have the right to clean water, then the steel plant
can pay them in order to get permission to pollute. Alternatively, if the plant has the property
rights, then downstream parties can compensate the steel plant to restrict its pollution.

The Coase Theorem, named after its originator, Ronald Coase, deals with this situation. It
asserts that, so long as all parties have full information, bargaining is costless, and there are
well-defined property rights, then the outcome is efficient regardless of how property rights are
assigned. Essentially this is a market solution: clearly establish property rights and then let the
market bargaining determine the outcome.

Though in some contexts the underlying assumptions for the Coase Theorem might seem a
little too strong, in others contexts they are quite reasonable. In fact, the Coase approach has
been much discussed recently, for example, in the context of transferable emissions permits or
licenses to pollute.

Example. Common facility.
Different divisions within a firm frequently compete for a common resource. Suppose that
divisions 1 and 2 of a given firm share a common facility F. Let y
i
be the service level used by
division i (i=1,2). Division is gross benefit in terms of improved divisional earnings is given
by y
i
0.25 y
i
2
0.1 (y
1
+y
2
).
a) What are the equilibrium levels of y
i
if the various divisions act separately?
b) What are the optimal levels of y
i
from an overall firm point of view?
c) Explain the difference between the results in a) and b).
d) How can equilibrium and optimality be reconciled?

Answer.
a) Each division maximizes y
i
0.25 y
i
2
0.1 (y
1
+y
2
). Let us consider division 1. Since it
cannot control the other division, the optimal solution corresponds to maximizing y
1
0.25 y
1
2

0.1 y
1
= 0.9 y
1
0.25 y
1
2
. The solution is y
1
= 1.8. By symmetry, the same is true for the
other division, so y
2
= 1.8.

b) The firm maximizes total benefit, that is y
1
0.25 y
1
2
0.1 (y
1
+y
2
) + y
2
0.25 y
2
2
0.1 (y
1

+y
2
) = 0.8 (y
1
+ y
2
) 0.25 (y
1
2
+ y
2
2
). Maximizing with respect to y
1
we get y
1
= 1.6.

c) The equilibrium value is greater than the optimal value. The reason is that there is a negative
externality: when division 1 uses the facility more intensely, it does not take into account that
this costs division 2 lower divisional earnings. In other words, part of division 1s gain is not a
gain for a firm as a whole.

d) One possibility is to create a system of transfer pricing whereby each division pays for the
use of the common facility F. If the fee is set at .1 (the cost imposed on the other facility),
then the equilibrium solution will be optimal.

Lecture Notes 148

Networks and Standards

Almost all of us use PCs that run Microsoft Windows. Do we simply have the same taste, or is
something deeper going on? 20 years ago, Betamax was a technologically superior format to
VHS. Yet we now all use VHS (which has improved markedly in the meantime). 150 years
ago, hundreds of phone companies existed. But since they werent interconnected, the AT&T
Bell System eventually took over almost the entire market as a government supported and
regulated monopoly. The rationale behind all of these stories is what we term network
effects. The value of the product to customers depends not only on its quality but on the
number of other people using the same product. There is, in essence, a network of users, and
the size of the network plays a large part in its creating value. Thus we all use Windows-based
PCs because they support the same programs and allow us to share files easily. Similarly, VHS
won because once it got a large enough market share (another story), there was no reason for
video stores to stock tapes in a second format. AT&T won for a similar reason: once it got a
critical mass of customers, it attracted others who wanted to communicate with them.

Markets with network effects have a number of unusual features. First, theres a tendency for
the winner to claim all or most of the market, since buyers want not only a good product, but a
product compatible with similar products used by others. Second, such markets are frequently
subject to battles over standards, since a firm that is able to establish its own product as the
standard is often in a strong competitive situation. Third, the best product need not win. An
accident that gives a product a critical mass of users (an installed base) has a strong competitive
position, regardless of the quality of its product. Fourth, monopoly may be the natural
outcome of competition. How should antitrust authorities respond to such market power? In
short, it changes the nature of competition in fundamental ways.

Network Effects

Suppose I were to tell you that I had here, in my overcoat, a brand new Panasonic cordless
phone with all the bells and whistles on it, and that I would sell this telephone to you for $1.
Not a bad deal, right? Its even cheaper than a subway fare. Now suppose you knew the
following information. The year is 1789, and even though I am selling the phone to you in real
terms (i.e., the phone costs the 1789 equivalent of $1 dollar today), you would probably not
buy it. Why? Because no one else had telephones in 1789! What would the value of owning a
telephone be if you couldnt call anyone?

Now lets jump ahead to the early 20th century. Thanks to our friendly monopolist AT&T,
more and more people are connected to the telephone network across the US. Now you may
be able to call a very small percentage of your family members or business associates. All of a
sudden the phone that was worthless to you in 1789 begins to look better. Now suppose I tell
you that we have traveled back to the future and we are in 2001. You can call just about
Firms and Markets
Lecture Notes

Lecture Notes 149

anyone you know in the world on the phone not just a handful of people. This same piece of
electronic equipment, which hasnt changed one bit since I offered it to you 200 years ago, is
worth a lot more. Why? The answer is network effects.

Network effects arise when each consumers valuation of a certain product increases with the
addition of other consumers who use the product. We can extrapolate the telephone example
above to many other daily applications like email and fax machines.

We call such examples direct network effects because, as we show above, it is obvious how
these products utility increases as other people also demand them. Still another example of
direct network externalities is given by software. For example, consider Microsoft Word. On
one level, your valuation for this product relates to the features and functionality it provides
you (for example, the spell-checker if you are a terrible typist). But in addition, suppose you
insisted on using an old, DOS-based Word program that no one else uses anymore. It may be
great for you if you are used to it and love it, but what if you need to email someone a
document (and vice versa)? What if you need to print on a printing network that does not
recognize your software? What if you are working on a group project and splitting up a paper
assignment? Because millions of us use Microsoft Word together and are able to interconnect
through this shared application, the value of the program increases with each additional MS
Word user to the extent that we can all communicate easily and efficiently.

We also can observe indirect network effects arising from many products. One example is
given by ATM machines. My banking with Chase has nothing to do (directly) with your value
of using Chase directly, but the fact that we all are using Chase makes the firm more likely to
build ATMs all over the city which provides us all with increased utility.

We can probably guess at this point that the value of a network good depends on the number
of consumers who demand it. This point is not lost on firms, who realize that they must
achieve a so-called critical mass of users in order to reap the benefits of network effects. What
are these benefits? Firms like Microsoft and Chase create added inelasticity for their products,
keeping consumers from switching to other products and attracting new consumers thanks to
the bandwagon effect of demanding something that everyone around you also has.

Critical Mass

The logic of critical mass is as follows. Because the demand for network products increases
with the expectation that there will be additional users whose presence increases the networks
value (called an externality because you, as the additional consumer, do not capture this added
value that you create), the demand curve takes on an unusual humped shape instead of the
normal downward sloping line. For this reason, as price declines, a snowball effect may take
place as many, many additional consumers suddenly demand the product. At this point in the
demand curve we say that the installed base (the number of customers comprising the
network) has attained critical mass.

We can see why firms want to attain critical mass for their network products, but this is not
always easy. One confounding factor can be that of excess inertiathat is, the prevention of

Lecture Notes 150

critical mass from being attained by some stickiness to a rival product or an unwillingness to
adopt a new network product. A good example of this is DVD players. While firms are
gradually lowering the price of DVD players as their marginal cost declines, DVDs have been
slow to explode in the marketplace. Why? Because many consumers are unwilling to
substitute DVDs for their VCR players and the scores of tapes that would need to be replaced.
Even though everyone may find that they prefer DVD players to VRC players once they
switched, this leap is not easy for sellers to induce. Thus, sellers often take recourse against
excess inertia by introducing new network standards that are backwards compatible (that is,
they can be integrated with existing network goods) or by pricing at an extremely low level to
offset consumers switching costs.

Standards Wars

In the late 1970s, VHS and Betamax hit the scene around the same time with different,
incompatible, VCR formats. When VCRs started to be used for viewing movies (in the 1980s),
video stores typically offered both types of cassettes to consumers. While it is agreed that
Betamax was a better technology, we know today that VHS all but eclipsed the Betamax
standard. What happened from Betamaxs point of view? Some bad decision making in terms
of pricing and vendor relationships and some lucky factors for VHS led to more network
adoption of the inferior VHS product; eventually as the market was tipped in the favor of
VHS its installed base soared relative to that of Betamax. And once consumers were stuck
on the VHS standard, there was no looking back Betamax was left in the dust.

This example underscores the point that in standards wars between network products, usually
one standard eventually captures the entire market due to network effects. The other standard
generally is abandoned, or is relegated to a small niche of users that generate lower network
externalities (as in the case of the Mac OS). Related to this, we can see that the best standard is
not always the winning standard. Thus, firms must behave strategically to capture the
bandwagon of consumer behavior with respect to network products whether or not their
product is superior. Why? Because once they lose the ability to harness network effects,
consumer demand snowballs in favor of another standard for the reasons discussed above.

It is further important to note that not all outcomes in standards wars are controllable by firms,
even if they are acting strategically. While it is easy to cast hindsight onto many standards wars
over the years and point to why one product is victorious over another, the actual escalation of
network demand is determined by a conspiracy of smaller eventssome controllable by the
firm, some not. Sometimes economists use the expression path dependence to denote this
determinism by small historical events that irrevocably shape the future.

An additional strategic dimension of network products and standards wars is that of the
compatibility question. Why did IBM allow rival firms to clone its PCs? And why did Apple
refuse to integrate its computing platform with IBM-compatible PCs? Firms with standards
must weigh the potential costs and benefits of making their products compatible with
competing products. On the positive side, allowing your products to be used with a wider
number of applications, even though you are competing with those applications, creates
additional demand for your products that would otherwise not exist. IBM, from above,
figured that if it allowed for the number of IBM-compatible PCs to explode, its own

Lecture Notes 151

computersat the center of the standardwould be more demanded and other computing
standards (i.e., Apple) would suffer. But on the downside, compatibility also creates risks. If
you make your software compatible with that of a rival firm with superior software, you may
find your installed base migrating to their products over time.

There is no clear answer to whether compatibility or incompatibility is better for a firm; it
depends on the particular industry, the size of the firm, consumer behavior, and of course the
competitive landscape. We can generalize a bit by saying that if product market competition is
intense, then incompatibility is a superior strategy. Conversely, if competition for an eventual
standard is intensein other words, if there are two competing standards vying for critical
mass and only one will survivecompatibility may be the better option.

Example

Two firms, Firm 1 and Firm 2, have heavily invested in a new type of high-density versatile
disk. This new disk will be able to contain the information corresponding to about twelve two-
hour movies. The technology comes with a computer device that allows for reading and
writing the new disks at high speed. Each firm has come up with its own design, A and B.
Although each design performs approximately the same functions, the designs are not
compatible. The firms must now decide which design to market. The payoffs from such
decision can be summarized in the following table.

A B
u w
A u+v w
w u+v
B w u

Question: What values would you expect u, v and w to take? Why.
Answer: If network effects are very significant, it is likely that w is a small number, probably
close to zero. If the firms continue to support different, incompatible standards, then it is
likely that the market will be very confused and shy away from investing in high-density DVDs
(we always have the good old DVDs that we have now). We would also expect u and v to be
positive numbers. The idea is that firms gain from supporting the same design (u>0),
especially if that is the design they invested on initially (v>0).

Question: What are the Nash equilibria of this game?
Answer: Suppose that w<0, say w=0 and that u,v>0. Then there are two Nash equilibria, (A,A)
and (B,B).

Open question: What examples can you think of where a game like this was played? What
aspects of your example are not depicted in this game?

2012 NYU Stern School of Business.
Firm 2
Firm 1

CourseBrick Spring2012 Zhou

Caricato da

Informazioni sul documento

Descrizione originale:

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

CourseBrick Spring2012 Zhou

Caricato da

Copyright:

Formati disponibili

Firms and Markets

. So, if your two points are x

Potrebbero piacerti anche