Auction Theory - An Introduction

Felix Munoz-Garcia
School of Economic Sciences
Washington State University

February 20, 2015

Introduction
Auctions are a large part of the economic landscape:
Since Babylon in 500 BC, and Rome in 193 AC
Auction houses Shotheby’s and Christie’s founded in 1744 and
1766.

Munch’s “The Scream,” sold for US$119.9 million in 2012.

Introduction

Auctions are a large part of the economic landscape:

More recently:
eBay: $11 billion in revenue, 27,000 employees.

Entry of more …rms in this industry: QuiBids.com.

Introduction
Also used by governments to sell:
Treasury bonds,
Air waves (3G technology):
British economists called the sale of the British 3G
telecom licences "The Biggest Auction Ever" ($36 billion)
Several game theorists played an important role in
designing the auction.
Overview

Auctions as allocation mechanisms:

types of auctions, common ingredients, etc.
First-price auction.
Optimal bidding function.
How is it a¤ected by the introduction of more players?
How is it a¤ected by risk aversion?
Second-price auction.
E¢ ciency.
Common-value auctions.
The winner’s curse.
Auctions

N bidders, each bidder i with a valuation vi for the object.

One seller.
We can design many di¤erent rules for the auction:
1 First price auction: the winner is the bidder submitting the
highest bid, and he/she must pay the highest bid (which is
his/hers).
2 Second price auction: the winner is the bidder submitting the
highest bid, but he/she must pay the second highest bid.
3 Third price auction: the winner is the bidder submitting the
highest bid, but he/she must pay the third highest bid.
4 All-pay auction: the winner is the bidder submitting the
highest bid, but every single bidder must pay the price he/she
submitted.
Auctions
All auctions can be interpreted as allocation mechanisms with
the following ingredients:
1 an allocation rule (who gets the object):
1 The allocation rule for most auctions determines the
object is allocated to the individual submitting the
highest bid.
2 However, we could assign the object by a lottery, where
prob (win ) = b1 +b2 b+1...+bN as in "Chinese auctions".
2 a payment rule (how much every bidder must pay):
1 The payment rule in the FPA determines that the
individual submitting the highest bid pays his bid, while
everybody else pays zero.
2 The payment rule in the SPA determines that the
individual submitting the highest bid pays the second
highest bid, while everybody else pays zero.
3 The payment rule in the APA determines that every
individual must pay the bid he/she submitted.
Private valuations

I know my own valuation for the object, vi .

I don’t know your valuation for the object, vj , but I know that
it is drawn from a distribution function.
1 Easiest case:
10 with probability 0.4, or
vj =
5 with probability 0.6

2 More generally,
F (v ) = prob (vj < v )
3 We will assume that every bidder’s valuation for the object is
drawn from a uniform distribution function between 0 and 1.
Private valuations
Uniform distribution function U [0, 1]

If bidder i ’s valuation is v , then all points in the horizontal axis

where vj < v , entail...
Probability prob (vj < v ) = F (v ) in the vertical axis.
Private valuations

Uniform distribution function U [0, 1]

Similarly, valuations where vj > v (horizontal axis) entail:

Probability prob (vj > v ) = 1 F (v ) in the vertical axis.
Under a uniform distribution, implies 1 F (v ) = 1 v.
Private valuations

Since all bidders are ex-ante symmetric...

They will all be using the same bidding function:

bi : [0, 1] ! R+ for every bidder i

They might, howver, submit di¤erent bids, depending on their

privately observed valuation.
Example:
1 A valuation of vi = 0.4 inserted into a bidding function
bi (vi ) = v2i , implies a bid of bi (0.4) = $0.2.
2 A bidder with a higher valuation of vi = 0.9 implies, in
contrast, a bid of bi (0.9) = 0.9 2 = $0.45.
3 Even if bidders are symmetric in the bidding function they use,
they can be asymmetric in the actual bid they submit.
First-price auctions

Let us start by ruling out bidding strategies that yield negative

(or zero) payo¤s, regardless of what your opponent does,
i.e., deleting dominated bidding strategies.
Never bid above your value, bi > vi , since it yields a
negative payo¤ if winning.

EUi (bi jvi ) = prob (win) (vi bi ) + prob (lose ) 0 < 0

| {z }

Never bid your value, bi = vi , since it yields a zero payo¤ if

winning.

EUi (bi jvi ) = prob (win) (vi bi ) + prob (lose ) 0 = 0

| {z }
0
First-price auctions

Therefore, the only bidding strategies that can arise in

equilibrium imply “bid shading,”
That is, bi < vi .
More speci…cally, bi (vi ) = a vi , where a 2 (0, 1).
First-price auctions

But, what is the precise value of parameter a 2 (0, 1).

That is, how much bid shadding?
Before answering that question...
we must provide a more speci…c expression for the probability
of winning in bidder i’s expected utility of submitting a bid x,

EUi (x jvi ) = prob (win) (vi x)

First-price auctions
Given symmetry in the bidding function, bidder j can
"recover" the valuation that produces a bid of exactly $x.
From the vertical to the horizontal axis,
Solving for vi in function x = a vi , yields vi = x
a
First-price auctions
What is, then, the probability of winning when submitting a
bid x is...
prob (bi > bj ) in the vertical axis, or
prob ( xa > vj ) in the horizontal axis.
First-price auctions

And since valuations are uniformly distributed...

prob ( xa > vj ) = xa
which implies that the expected utility of submitting a bid x
is...
x
EUi (x jvi ) = ( vi x )
a
|{z}
prob (win )

And simplifying...
xvi x2
=
a
First-price auctions

2
Taking …rst-order conditions of xvi a x with respect to x, we
obtain
vi 2x
=0
a
and solving for x yields an optimal bidding function of
1
x (vi ) = vi .
2
Optimal bidding function in FPA

x (vi ) = 12 vi .

Bid shadding in half :

for instance, when vi = 0.75, his optimal bid is 12 0.75 = 0.375.
FPA with N bidders

The expected utility is similar, but the probability of winning

di¤ers...
x x x x
prob (win) = ... ...
a a a a
x N 1
=
a
Hence, the expected utility of submitting a bid x is...

x N 1 x N 1
EUi (x jvi ) = (vi x) + 1 0
a a
FPA with N bidders

Taking …rst-order conditions with respect to his bid, x, we

obtain
x N 1 x N 2 1
+ (vi x) = 0
a a a

Rearranging,
x N a
[(N 1)vi nx ] = 0,
a x2
and solving for x, we …nd bidder i’s optimal bidding function,
N 1
x (vi ) = vi
N
FPA with N bidders
N 1
Optimal bidding function x (vi ) = N vi

Comparative statics:
Bid shadding diminishes as N increases.
Bidding function approaches 450 line.
FPA - Generalization

Let us now allow for valuations to be drawn from any cdf

F (vi ) (not necessarily uniform).
First, note that, for a given bidding strategy s : [0, 1] ! R+ ,
i.e., s (vi ) = xi , we can de…ne its inverse s 1 (xi ) = vi ,
implying that the cdf can be rewritten as
1
F (vi ) = F (s (xi )).

Then bidder i’s UMP becomes

1 n 1
max F (s (xi )) (vi xi )
xi | {z }
prob(win)
FPA - Generalization

Taking …rst-order conditions with respect to x yields

1 n 1
F (s (xi )) +

ds 1 (x )
1 n 2 1 i
(n 1) F (s (xi )) f (s (xi )) (vi xi ) = 0
dxi
ds 1 (x )
1
Since s 1 (xi ) = vi and dxi
i
= s 0 (s 1 (x , the above
i ))
expression becomes
1
[F (vi )]n 1
+ (n 1) [F (vi )]
n 2
f (vi ) (vi xi ) = 0
s 0 (vi )
FPA - Generalization

Further rearranging, we obtain

(n 1) [F (vi )]n 2
f (vi )vi (n 1) [F (vi )]n 2
f (vi )xi
n 1 0
= [F (vi )] s (vi )

or

[F (vi )]n 1 0
s (vi ) + (n 1) [F (vi )]n 2
f (vi )vi
n 2
= (n 1) [F (vi )] f (vi )xi

d [[F (vi )]n 1 s (vi )]

The LHS is dvi . Hence,
h i
d [F (vi )]n 1
s (vi )
= (n 1) [F (vi )]n 2
f (vi )xi
dvi
FPA - Generalization

Integrating both sides yields

Z vi
n 1
[F (vi )] s ( vi ) = (n 1) [F (vi )]n 2
f (vi )vi dvi (1)
0

Applying integration by parts on the RHS, we obtain

Z vi
(n 1) [F (vi )]n 2
f (vi )vi dvi (2)
0
Z vi
= [F (vi )]n 1
vi [F (vi )]n 1
dvi (3)
0

Plugging that into the RHS of (1) yields

Z vi
n 1 n 1
[F (vi )] s (vi ) = [F (vi )] vi [F (vi )]n 1
dvi (4)
0

A note on integration by parts (next slide)

FPA - Generalization

Recall integration by parts: You start from two functions g

and h, so that (gh)0 = g 0 h + gh0 . Then, integrating both
sides yields
Z Z
0
g (x )h (x ) = g (x )h(x )dx + g (x )h0 (x )dx

We can then reorder the terms in the above expression as

follows
Z Z
g 0 (x )h(x )dx = g (x )h(x ) g (x )h0 (x )dx
FPA - Generalization

In order to apply integration by parts in our auction setting,

n 2
let g 0 (x ) (n 1) [F (vi )] f (vi ) and h(x ) vi . That is
Z vi
(n 1) [F (vi )]n 2
f (vi ) vi dvi
0 | {z }|{z}
g 0 (x ) h (x )
Z vi
= [F (vi )]n 1 vi [F (vi )]n 1 |{z}
1 dvi
| {z }|{z} 0 | {z }
g (x ) h (x ) g (x ) h 0 (x )
FPA - Generalization

We can now rearrange expression (3). In particular, dividing

both sides by [F (vi )]n 1 yields
R vi
0 [
F (vi )]n 1 dvi
s (vi ) = vi
[F (vi )]n 1
which is bidder i’s optimal bidding function, s (vi ).
Intuitively,
R
he shades his bid by the amount of ratio
vi n 1
0 [F (v i )] dvi
n 1 .
[F (vi )]
As a practice, note that when F (vi ) is uniform, F (vi ) = vi
n 1
implying that [F (vi )] = vin 1 . Hence,
1 n
n vi vin n 1
s (vi ) = vi = vi = vi
vin 1 nvin 1 n
FPA with risk-averse bidders

Utility function is concave in income, x, e.g., u (x ) = x α ,

where 0 < α 1 denotes bidder i’s risk-aversion parameter.
[Note that when α = 1, the bidder is risk neutral.]
Hence, the expected utility of submitting a bid x is
x
EUi (x jvi ) = (vi x )α
a
|{z}
prob (win )
FPA with risk-averse bidders

Taking …rst-order conditions with respect to his bid, x,

1 x 1
(vi x )α α ( vi x )α = 0,
a a
and solving for x, we …nd the optimal bidding function,
vi
x (vi ) = .
1+α
Under risk-neutral bidders, α = 1, this function becomes
x (vi ) = v2i .
But, what happens when α decreases (more risk aversion)?
FPA with risk-averse bidders
vi
Optimal bidding function x (vi ) = 1 +α .

Bid shading is ameliorated as bidders’risk aversion increases:

That is, the bidding function approaches the 450 line when α
approaches zero.
FPA with risk-averse bidders

Intuition: for a risk-averse bidder:

the positive e¤ect of slightly lowering his bid, arising from
getting the object at a cheaper price, is o¤set by...
the negative e¤ect of increasing the probability that he loses
the auction.
Ultimately, the bidder’s incentives to shade his bid are
diminished.
Second-price auctions

Let’s now move to second-price auctions.

Second-price auctions

Bidding your own valuation, bi (vi ) = vi , is a weakly dominant

strategy,
i.e., it yields a larger (or the same) payo¤ than submitting any
other bid.
In order to show this, let us …nd the expected payo¤ from
submitting...
A bid that coincides with your own valuation, bi (vi ) = vi ,
A bid that lies below your own valuation, bi (vi ) < vi , and
A bid that lies above your own valuation, bi (vi ) > vi .
We can then compare which bidding strategy yields the
largest expected payo¤.
Second-price auctions

Bidding your own valuation, bi (vi ) = vi ...

Case 1a: If his bid lies below the highest competing bid, i.e.,
bi < hi where hi = maxfbj g,
j 6 =i
then bidder i loses the auction, obtaining a zero payo¤.
Second-price auctions

Bidding your own valuation, bi (vi ) = vi ...

Case 1b: If his bid lies above the highest competing bid, i.e.,
bi > hi , then bidder i wins.
He obtains a net payo¤ of vi hi .
Second-price auctions
Bidding your own valuation, bi (vi ) = vi ...

Case 1c: If, instead, his bid coincides with the highest competing
bid, i.e., bi = hi , then a tie occurs.
For simplicity, ties are solved by randomly assigning the object
to the bidders who submitted the highest bids.
As a consequence, bidder i’s expected payo¤ becomes
1
2 ( vi hi ).
Second-price auctions

Bidding below your valuation, bi (vi ) < vi ...

Case 2a: If his bid lies below the highest competing bid, i.e.,
bi < hi ,
then bidder i loses, obtaining a zero payo¤.
Second-price auctions

Bidding below your valuation, bi (vi ) < vi ...

Case 2b: if his bid lies above the highest competing bid, i.e.,
bi > hi ,
then bidder i wins, obtaining a net payo¤ of vi hi .
Second-price auctions

Bidding below your valuation, bi (vi ) < vi ...

Case 2c: If, instead, his bid coincides with the highest competing
bid, i.e., bi = hi , then a tie occurs,
and the object is randomly assigned, yielding an expected
payo¤ of 12 (vi hi ).
Second-price auctions

Bidding above your valuation, bi (vi ) > vi ...

Case 3a: if his bid lies below the highest competing bid, i.e.,
bi < hi ,
then bidder i loses, obtaining a zero payo¤.
Second-price auctions
Bidding above your valuation, bi (vi ) > vi ...

Case 3b: if his bid lies above the highest competing bid, i.e.,
bi > hi , then bidder i wins.
His payo¤ becomes vi hi , which is positive if vi > hi , or
negative otherwise.
Second-price auctions
Bidding above your valuation, bi (vi ) > vi ...

Case 3c: If, instead, his bid coincides with the highest competing
bid, i.e., bi = hi , then a tie occurs.
The object is randomly assigned, yielding an expected payo¤ of
1 (v hi ), which is positive only if vi > hi .
2 i
Second-price auctions

Summary:
Bidder i’s payo¤ from submitting a bid above his valuation:
either coincides with his payo¤ from submitting his own value
for the object, or
becomes strictly lower, thus nullifying his incentives to deviate
from his equilibrium bid of bi (vi ) = vi .

Hence, there is no bidding strategy that provides a strictly

higher payo¤ than bi (vi ) = vi in the SPA.
All players bid their own valuation, without shading their bids,
unlike in the optimal bidding function in FPA.
Second-price auctions

Remark:
The above equilibrium bidding strategy in the SPA is
una¤ected by:
the number of bidders who participate in the auction, N, or
their risk-aversion preferences.
E¢ ciency in auctions

The object is assigned to the bidder with the highest

valuation.
Otherwise, the outcome of the auction cannot be e¢ cient...
since there exist alternative reassignments that would still
improve welfare.
FPA and SPA are, hence, e¢ cient, since:
The player with the highest valuation submits the highest bid
and wins the auction.
Lottery auctions are not necessarily e¢ cient.
Common value auctions

In some auctions all bidders assign the same value to the

object for sale.
Example: Oil lease
Same pro…ts to be made from the oil reservoir.
Common value auctions

Firms, however, do not precisely observe the value of the

object (pro…ts to be made from the reservoir).
Instead, they only observe an estimate of these potential
pro…ts:
from a consulting company, a bidder/…rm’s own estimates, etc.
Common value auctions

Consider the auction of an oil lease.

The true value of the oil lease (in millions of dollars) is
v 2 [10, 11, ..., 20]
Firm A hires a consultant, and gets a signal s

v + 2 with prob 12 (overestimate)

s=
v 2 with prob 12 (underestimate)

That is, the probability that the true value of the oil lease is v ,
given that the …rm receives a signal s , is
1
prob (v js ) = 2 if v = s 2 (overestimate)
1
2 if v = s + 2 (underestimate)
Common value auctions

If …rm A was not participating in an auction, then the

expected value of the oil lease would be
1 1 s 2+s +2 2s
(s 2) + (s + 2) = = =s
|2 {z } |2 {z } 2 2
if overestimation if underestimation

Hence, the …rm would pay for the oil lease a price p < s,
making a positive expected pro…t.
Common value auctions

What if the …rm participates in a FPA for the oil lease against
…rm B?
Every …rm uses a di¤erent consultant...
but they don’t know if their consultant systematically
overestimates or underestimates the value of the oil lease.
Every …rm receives a signal s from its consultant,
observing its own signal, but not observing the signal the other
…rm receives, every …rm submits a bid from f1, 2, ..., 20g.
Common value auctions

We want to show that bidding b = s 1 cannot be optimal

for any …rm.
Notice that this bidding strategy seems sensible at …rst glance:
Bidding less than the signal, b < s.
So, if the true value of the oil lease was s , the …rm would
get some positive expected pro…t from winning.
Bidding is increasing in the signal that the …rm receives.
Common value auctions

Let us assume that …rm A receives a signal of s = 10.

Then it bids b = s 1 = 10 1 = $9.
Given such a signal, the true value of the oil lease is

s + 2 = 12 with prob 12
v=
s 2 = 8 with prob 12

In the …rst case (true value of 12)

…rm A receives a signal of sA = 10 (underestimation), and
…rm B receives a signal of sB = 14 (overestimation).
Then, …rms bid bA = 10 1 = 9, and bB = 14 1 = 13, and
…rm A loses the auction.
Common value auctions

In the second case, when the true value of the oil lease is
v = 8,
…rm A receives a signal of sA = 10 (overestimation), and
…rm B receives a signal of sB = 6 (underestimation).
Then, …rms bid bA = 10 1 = 9, and bB = 6 1 = 5, and
…rm A wins the auction.
However, the winner’s expected pro…t becomes
1 1 1
(8 9) + 0 =
2 2 2
Negative pro…ts from winning.
Winning is a curse!!
Winner’s curse

In auctions where all bidders assign the same valuation to the

object (common value auctions),
and where every bidder receives an inexact signal of the
object’s true value...
The fact that you won...
just means that you received an overestimated signal of the
true value of the object for sale (oil lease).
How to avoid the winner’s curse?
Bid b = s 2 or less,
take into account the possibility that you might be receiving
overestimated signals.
Winner’s curse - Experiments I

In the classroom: Your instructor shows up with a jar of

nickels,
which every student can look at for a few minutes.

Paying too much for it!

Winner’s curse - Experiments II

In the …eld: Texaco in auctions selling the mineral rights to

o¤-shore properties owned by the US government.
All …rms avoided the winner’s curse (their average bids were
about 1/3 of their signal)...
Expect for Texaco:
Not only their executives fall prey of the winner’s curse,
They submitted bids above their own signal!
They needed some remedial auction theory!