3 Microeconomics Questions
Fokkusu
LectureNotes.pdf
Slides 15
(Chapter 22)
What do we mean by network effects?
Willingness to pay increases with the number of
people that acquire the good
Example: cell-phone, fax machine, Facebook…
People choices depend on the expected network size!
Often, there is a coordination problem!
3
How can we model demand with
network effects?
Setup for demand model:
Very large number of consumers. Thus, each
consumer doesn’t think he/she affects the overall
fraction that buys the good
Consumers have unit demand (e.g. they only want
one Facebook account)
Reservation values for the good are between 0 and
100, and uniformly distributed
Call these values vi
Price of good is p
4
How can we model demand with
network effects?
Setup for demand model:
Call the fraction of consumers buying the good f
Then individual demand is
Implication, even if a consumer has vi>>p, the
consumer won’t buy if nobody else is buying!
We are interested in aggregate demand f(p)
0 if ( )
1 if
i i
i
fv p q p
fv p
5
Let us start without network effects…
Then individual demand is
qi (p) = 0 if vi < p and qi (p) = 1 if vi ≥ p
If vi is uniformly distributed [0, 100] then
f(p) = (100 – p)/100
6
What happens with network effects?
The indifferent consumer satisfies vi = p/f
All consumers with vi > p/f buy the good!
His decision depends on i’s believes about what
everyone else is doing!
7
What happens with network effects?
We can think about aggregate demand in terms of
Nash equilibrium among consumers
Each consumer takes as given others’ choices while
deciding whether to buy the good or not
In equilibrium, each consumer selects the optimal
choice and correctly predicts the choices of others
8
How do we get aggregate demand?
Let all consumers believe f will get the good
Is f(p) = 0 a NE? Yes!
If f = 0, then vi f = 0 < p
This means that no consumer wants to get the good!
Is there another NE with f(p) > 0 ? Yes!
9
How do we get aggregate demand?
Is there another NE with f(p) > 0 ? Yes!
Let all consumers believe f > 0 will get the good
Recall indifferent consumer satisfies vi = p/f
Since vi is uniformly distributed [0, 100] then
f(p) = (100 – p/f)/100
The NE solve
100 f(p)2 – 100 f(p) + p = 0
Smallest of the two solutions f* is called critical mass
10
What are the profits of the firm?
For each p, profits depend on the selected NE!
It makes more profits in the largest NE
How can the firm affect equilibrium selection?
Slides 13
(Chapter 8)
2
Bundling
Firms sell goods as bundles
Selling two or more goods in a single package
Goods are close in nature, but slightly different
Goods are not necessarily complements
Complementary goods: buying one good increases your utility from buying the other good
3
Bundling example: Microsoft
Consider MS Office
Can buy Office package for $350. This contains Word, Excel, Outlook, and Powerpoint
Each component costs $140
So purchasing the bundle via individual purchases costs $560.
The bundle offers a $210 discount!
Why might Microsoft do this?
4
Bundling: an example
Two movies are available from the same producer
Star Wars and Superbad
Selling to two theatres: A and B. 1st degree price
discrimination is illegal
Willingness to pay is:
Theater A
Theater B
Willingness to
pay for
Star Wars
Willingness to
pay for
Superbad
$8,000
$7,000
$2,500
$3,000
5
Bundling: an example
Suppose component pricing (no bundling)
What price should producer set for Star Wars?
For Superbad?
Profits?
Theater A
Theater B
Willingness to
pay for
Star Wars
Willingness to
pay for
Superbad
$8,000
$7,000
$2,500
$3,000
6
Bundling: an example
Suppose component pricing (no bundling)
What price should producer set for Star Wars?
For Superbad?
Profits?
Theater A
Theater B
Willingness to
pay for
Star Wars
Willingness to
pay for
Superbad
$8,000
$7,000
$2,500
$3,000
$7000
7
Bundling: an example
Theater A
Theater B
Willingness to
pay for
Star Wars
Willingness to
pay for
Superbad
$8,000
$7,000
$2,500
$3,000
$7000
$2500
Suppose component pricing (no bundling)
What price should producer set for Star Wars?
For Superbad?
Profits?
8
Bundling: an example
Suppose component pricing (no bundling)
What price should producer set for Star Wars?
For Superbad?
Profits?
Theater A
Theater B
Willingness to
pay for
Star Wars
Willingness to
pay for
Superbad
$8,000
$7,000
$2,500
$3,000
$7000
$2500
$19000
9
Bundling: an example
Now suppose the producer tries pure bundling
This means the producer charges a single price at
which a theatre can buy both movies, or neither
No option to buy only one movie
What price to set for the bundle?
Theater A
Theater B
Willingness to
pay for
Star Wars
Willingness to
pay for
Superbad
$8,000
$7,000
$2,500
$3,000
10
Bundling: an example
Theater A
Theater B
Willingness to
pay for
Star Wars
Willingness to
pay for
Superbad
$8,000
$7,000
$2,500
$3,000
Total
Willingness
to pay
$10,500
$10,000
The producer can charge $10,000 and both theatres will buy the bundle
Profits are $20,000
This is greater than profits with component pricing! ($19,000)
11
Why did bundling work?
With component pricing, had to set low price for each component
Only way to have both consumers buy each product
With bundling, producer can exploit the aggregate willingness to pay
The theatre that dislikes Star Wars likes Superbad, and vice-versa
This is a form of price discrimination!
Will this always work?
12
Bundling: alternative example
Will stick with the same movie problem, but now
change one payoff.
Theatre A now will pay $7000 for Superbad, not just
$2500
Now what are component prices? Bundle price?
Theater A
Theater B
Willingness to
pay for
Star Wars
Willingness to
pay for
Superbad
$8,000
$7,000
$7,000
$3,000
13
Bundling: alternative example
Component prices are $7000 for Star Wars and
$7000 for Superbad. Profits with component pricing
are $21000 (sell only one Superbad)
Bundle price is $10000. Bundling profits are only
$20000
Theater A
Theater B
Willingness to
pay for
Star Wars
Willingness to
pay for
Superbad
$8,000
$7,000
$7,000
$3,000
14
Why didn’t bundling work in the
changed example? I only changed a number, and now producer is
better off with component pricing than pure bundling
What happened?
Producer was better off selling Superbad only to theater A
Theatre A also had a relatively high value for Star Wars
Bundling doesn’t allow the producer to capture A’s high valuation
15
When doesn’t pure bundling
increase profits? Original paper on this is by Jim Adams and Janet
Yellen, 1976
One punchline: component pricing may work better if buyers’ valuations are positively correlated
That is, if a buyer has a high valuation for one product, he/she will probably have a high valuation for the other product too
Can also use mixed bundling: offer both the bundle and separate components
Monopolist can extract even more value
16
Adams and Yellen paper shows
intuition through graphs Suppose monopolist has two goods: good 1 and
good 2
Total production cost is c(q1,q2) = c1q1 + c2q2 No economies of scale or scope
Each consumer i has a reservation price for each good: R1i and R2i Value of buying both is simply R1i + R2i Goods are not complements!
Valuations vary across consumers
Do graphs on board (also see section 8.1 of PRN)
17
Introducing tying
Tying (tie-in) sales occur when, if consumers
want to buy one product from you, you require
them to buy a particular complementary product
Notice the similarity with bundling
Main difference is that with tying, the two goods
are complements
Microsoft case example: consumers who wanted
to buy Windows were required to buy Internet
Explorer rather than Netscape
18
Introducing tying
In some ways, the difference is semantics…
Tying: “To buy A, you must also buy B”
Pure Bundling: “You can only buy A and B
together”
19
Two motives for tying
Motive one: Price discrimination
We’ve already talked a lot about this
Example: low price for inkjet printers but high price
for ink cartridges
Cartridges incompatible across brands
Tying for this reason is not illegal (firms do this all
the time, after all)
May be efficiency enhancing relative to single
product linear pricing
20
Two motives for tying
Motive two: Use your market power in one
product to gain market power in the other
Tying to enhance or maintain a monopoly is
illegal under the Sherman Act
This sort of tying was one of the issues in the
Microsoft case
And other cases as well (e.g., Kodak in the 1980s)
Slides 14
(Chapter 19)
2
Advertising: what do we want to know?
How do firms choose how much to spend on
advertising?
Do firms advertise too much or too little relative
to what is socially optimal?
Does advertising lead to market power?
How does advertising affect demand?
3
Much controversy on nature of
advertising: informational or persuasive?
Economists have typically tried to categorize
advertising as being of one of two types:
1. Informational advertising: conveys information about
a product’s quality and / or price
2. Persuasive advertising: Attempts to “change
preferences” to create product differentiation
(without actually providing info)
4
Much controversy on nature of
advertising: informational or persuasive?
What are the competitive effects of these two types of advertising? i.e. Do we expect each type of advertising to
raise or lower prices?
If informational advertising is making people aware of price differences, that means search costs are lower
Should see lower prices. That’s good.
Persuasive advertising artificially increases product differentiation.
Should see higher prices. That’s bad.
5
Historically, lots of concern in industrial
organization regarding persuasive ads
During 1960s and 1970s, lots of economists
looked at profits and price – cost margins
across many industries
Generally find that industries in which firms
spend lots of money on advertising are also
more profitable and have bigger margins
Conclusion: Does advertising increase market
power?
6
Explaining correlation between
advertising and price – cost margins
Two potential stories:
Could interpret regression results as advertising
increases market power and margins
Or could interpret regression as: in industries
where market power can be exerted, there are
also incentives to advertise
There exist theoretical models consistent with
both stories!
7
The Dorfman-Steiner advertising model
Second story basically says: If you have market
power, you have incentives to increase demand
Otherwise, no point in advertising!
Dorfman and Steiner (1954) have simple
model showing this intuition
Result: The amount of advertising you should
do (as a proportion of sales revenue) increases
with the elasticity of demand with respect to
advertising and decreases with the price
elasticity of demand
8
A different story: Sutton (1991)
Sutton wanted to explain why some large
markets have lots of small firms, while other
large markets have a few big firms
Example: large cities have lots of small
restaurants, but few newspapers
Sutton’s view: markets characterized by
endogenous sunk costs have high
concentration even when market size is large
9
What are endogenous sunk costs?
Endogenous sunk costs are initial investments
that increase the size of the market
Example: sunk cost of building the factory is
not endogenous: this doesn’t grow demand
Advertising is an endogenous sunk cost. The
more you advertise, the more you increase
demand
If you want to enter an industry where the
incumbents advertise a lot, you will need to sink
up-front $ into advertising
10
Advertising means that sunk costs
increase with market size
Sutton’s idea: in markets where demand is
responsive to investment…
There is an incentive to advertise
This incentive is bigger in larger markets
Thus firms advertise more in larger markets
This means that sunk entry costs increase with
market size
So barriers to entry increase with market size
So high concentration (and market power) even in
large markets
11
Which story is right?
Still an open question!
Maybe both are right?
12
Another question: is there too much
advertising?
It certainly seems as though there are a lot of ads
out there
What are the social benefits and costs of
advertising?
Do ads increase or decrease profits?
What about consumer welfare?
If informative ad: benefits come from lower prices,
consumers making better decisions
If persuasive ad: less clear
Ads potentially lead to higher prices
“Artificial” or “real” altering of preferences?
13
Advertising’s impact on profits: a
prisoners’ dilemma?
Maybe ads just lead firms to try to steal each
other’s business, without even increasing profits
Simple two firm example:
Coke and Pepsi set advertising expenditures AC and AP
Simplified profit functions are:
Coke: πC = (60 – AP)AC -
Pepsi: πP = (60 – AC)AP -
We work out Nash Equilibrium on the board
2
C A
2
P A
14
Advertising’s impact on profits: a
prisoners’ dilemma?
So both Coke and Pepsi spend $20 on advertising
Profits are $400
Could the firms have done better by cooperating?
Total industry profits are:
(60 – AP)AC + (60 – AC)AP - -
This is maximized when the firms spend $15 each.
Profits are $450 each.
2
C A
2
P A
15
Advertising’s impact on profits: a
prisoners’ dilemma?
If the firms cooperate (or merge), they internalize
the business stealing effects of advertising
No point stealing business from yourself
Like differentiated Bertrand vs collusion
Collusive prices are higher because you’re no
longer trying to steal business from other firms
But at least with price competition, consumers got
lower prices!
Here, they get more ads. Prices may increase!
16
Do consumers get anything out of
persuasive advertising?
Are some ads purely manipulative or do they
actually provide real value to consumers?
17
Do consumers get anything out of
persuasive advertising?
Maybe a brand is something that has value
When you buy a Nike shoe, you’re not just getting
a shoe, you’re getting value from the Nike brand
You get to “be seen” in cool shoes
Brand is a complement to the product
So maybe advertising really does make some
products better?
So in some cases, maybe there aren’t enough
ads?
All of this is an open issue
