The set of equilibria of first-price auctions
Paulo Klinger Monteiro
FGV-EPGE
Praia de Botafogo 190 sala 1103
Rio de Janeiro RJ Brazil
Abstract
1
Introduction
Suppose we somehow obtained a bidding functionb(·). It could, for example, be originated from some laboratory auction or be a linear interpolation of some auction data. Is this bidding function theoretically possible? If so, does it come from a model adequate to the situation at hand? In this paper I study this problem in a sealed bid first-price auction set up. To be more concrete we are in the independent private values model and suppose we have 3 bidders with signals in the interval [0,1] and that our estimated bidding function is b(x) = x
2. If the distribution of signals is uniform with three
bidders the equilibrium is b∗(x) = 2x
3 . With two bidders the equilibrium is
exactlyb∗(x) = x
2.Do we have a dummy bidder? Collusion? Here I focus on
the distribution of signals. The uniform distribution that we supposed in this example is usually used more for convenience than for theoretical reasons. In this example if we change the distribution to F (x) = √x we have that the equilibrium bidding function is exactly b(x). I show in this paper that for any number of bidders and practically any strictly increasing functionb(·) it is possible to find a strictly increasing continuous distribution function such that the equilibrium bidding function is exactlyb(·).This result is similar in spirit to the Sonneschein-Mantel-Debreu theorem on excess demand. I also analyze a second aspect of this problem. If we insist that the distribution of signals is given and vary the bidder valuation from Vi = xi to Vi = u(xi).
The conditions to find an appropriate u(·) are however harder to be met.
2
The model
We consider first-price sealed bid auctions. There are n bidders with inde-pendent private values and each bidders’ signal are in the interval [0,v]. We¯ consider distribution of bidders signals in the set
F ={F : [0,v¯]→[0,1] ;F is continuous strictly increasing and onto}.
Thus F is the set of strictly increasing distributions with support [0,v]¯ . Let F ∈ F be the distribution of bidder i≤n signal xi. If bidder ihas signal xi
he valuates the object as Vi =xi. Define bF (0) = 0 and if x∈(0,v],¯
bF(x) =x−
Rx
0 F
n−1(v)dv
Fn−1(x) . (1)
We first prove the
Proof: The continuity of bF for x′ > 0 is immediate. At x′ = 0 it follows
frombF (x)< x.Let us now prove that it is strictly increasing. Suppose that
0 ≤ x < y ≤ v.¯ If x = 0 it is immediate from Ry
0 F
n−1(v)dv < Fn−1(y)y
that bF(y)>0 =bF(x). Ifx >0 then:
bF(y)−bF (x) =y−
Ry
0 F
n−1(u)du
Fn−1(y) −x+
Rx
0 F
n−1(u)du
Fn−1(x) =
y−x−
Ry
x F
n−1(u)du
Fn−1(y) +
µ
1 Fn−1(x) −
1 Fn−1(y)
¶ Z x
0
Fn−1(u)du >0.
Note that Ry
x F
n−1(u)du≤Fn−1(y) (y−x). QED
We now show that bF is an equilibrium bidding function.
Proposition 1 Let F ∈ F. If there are n bidders with independent signals
distributed accordingly to F then bF (·) is a symmetric equilibrium of the
first-price auction.
Proof: Defineb =bF andx=xi. Suppose bidder j 6=i with signalxj bids
b(xj). We have to prove that for any y∈[0,v]¯ ,
(x−b(x)) Pr
µ
b(x)≥max
j6=i b(xj)
¶
≥(x−b(y)) Pr
µ
b(y)≥max
j6=i b(xj)
¶
.
Since b is strictly increasing and the signals are independent,
Pr
µ
b(x)≥max
j6=i b(xj)
¶
=Y
j6=i
Pr (b(x)≥b(xj)) =
Y
j6=i
Pr (x≥xj) =Fn−1(x).
Therefore
(x−b(x)) Pr
µ
b(x)≥max
j6=i b(xj)
¶
=
Z x
0
Fn−1(v)dv.
Thus we have to prove that for every y∈[0,¯v],
Z x
0
Fn−1(v)dv ≥(x−y)Fn−1(y) +
Z y
0
Fn−1(v)dv.
This is equivalent to
Z x
y
Fn−1(u)du
≥(x−y)Fn−1(y).
Considering separately the cases x > y and y≥xwe see that this inequality
is true. QED
Remark 1 Note that I do not suppose the differentiability of the distribution. This is of essence. If the bidding function is piecewise linear the distribution cannot be differentiable. See also the example below.
Define
B ={bF(·) ;F ∈ F}.
We may now prove our main theorem.
Theorem 1 Suppose b: [0,v]¯ →R . Then b∈ B if and only if:
1. b(·) is strictly increasing;
2. b(0) = 0;
3. b(x)< x if x >0;
4. limx→0+(x−b(x))e
R¯v x
dy
y−b(y) =∞.
Proof: Suppose b ∈ B. It is clear that it satisfies (1), (2) and (3). To see that it also satisfy (4) let F ∈ F be such thatb =bF. Then
1
y−b(y) =
Fn−1(y) Ry
0 Fn−1(v)dv
=
µ
log
µZ y
0
Fn−1(v)dv
¶¶′
.
Therefore
Z ¯v
x
dy
y−b(y) =
Z ¯v
x
µ
log
µZ y
0
Fn−1(v)dv ¶¶′
dy= log
à Rv¯
0 F
n−1(v)dv
Rx
0 Fn−1(v)dv !
and
(x−b(x))e
Rv¯ x
dy y−b(y) =
Rx
0 F
n−1(v)dv
Fn−1(x) R¯v
0 F
n−1(v)dv
Rx
0 Fn−1(v)dv
=
Rv¯
0 F
n−1(v)dv
Fn−1(x) .
Thus b satisfy (4). Now suppose b(·) satisfy (1), (2), (3) and (4). I show that b =bG where G(0) = 0 and if x∈(0,v],¯
G(x) = e−n1−1 Rv¯
x 1 y−b(y)dy
µ
¯
v −b(¯v) x−b(x)
¶n1 −1
.
First note that (4) imply thatGis continuous at 0.And obviouslyG(¯v) = 1. It is also clear that G is continuous if x >0.We now show that Gis strictly increasing. It is equivalent to prove that
φ(x) := logGn−1(x) =−
Z v¯
x
1
is strictly increasing. If h >0 then
φ(x+h)−φ(x)
h =
1 h
Z x+h
x
1
y−b(y)dy+ 1 hlog
µ
x−b(x)
x+h−b(x+h)
¶
.
If x+xh−−bb((xx)+h) ≥1 then
φ(x+h)−φ(x)
h ≥
1 h
Z x+h
x
1
y−b(y)dy.
If x+hx−−bb((xx+)h) >1, then
log
µ
x+h−b(x+h)
x−b(x)
¶
= log
µ
1 + h−b(x+h) +b(x) x−b(x)
¶
< h−b(x+h) +b(x) x−b(x)
and therefore
φ(x+h)−φ(x)
h >
1 h
Z x+h
x
1
y−b(y)dy− 1 h
h−b(x+h) +b(x)
x−b(x) ≥
1 h
Z x+h
x
1
y−b(y)dy− 1 x−b(x). Thus
lim inf
h→0+
φ(x+h)−φ(x)
h ≥0
which implies thatφ is increasing (see Saks, theorem 7.2 page 204). To show that it is strictly increasing suppose not. Then φ is constant in an interval (c, d) and therefore it is differentiable and henceb(·) is differentiable in (c, d) as well. Thus from
d
dxφ(x) = 1
x−b(x)−
1−b′(x)
x−b(x) =
b′(x)
x−b(x), x∈(c, d)
we have that b′(x) = 0 if x
∈(c, d) and this contradicts that b(·) is strictly increasing. It remains only to check that
bG(x) =x−
Rx
0 G
n−1(v)dv
Gn−1(x)
is equal to b(x). Now note that
Z x
0
Gn−1(u)du=
Z x
0
e−Ruv¯ 1 y−b(y)dy
µ
¯
v−b(¯v) u−b(u)
¶
du=
(¯v−b(¯v))
Z x 0 d du ³ e−
R¯v u
1 y−b(y)dy
´
du= (¯v−b(¯v))³e−
R¯v u
1 y−b(y)dy
´¯ ¯ ¯
x
0 =
(¯v−b(¯v))e−Rx¯v 1 y−b(y)dy.
Since
Gn−1(x) =e−Rv¯ x
1 y−b(y)dy
µ
¯
v −b(¯v) x−b(x)
¶
=
Rx
0 G
n−1(u)du
x−b(x) we conclude that
bG(x) = x−
Rx
0 G
n−1(u)du
Gn−1(x) =x−(x−b(x)) =b(x).
QED
Example 1 Let us consider b(x) = x/2, x ∈ [0,1]. Then if there are n
bidders,
G(x) = e−n1−1 R1
x 2 ydy 1
xn−11
=xn−11.
If n= 3 then G(x) =√x.
3
The model with a more general valuation
Suppose now that the set of signals of bidderiis an abstract probability space (X,T, P) and if bidderihas a signalxi ∈Xhis valuation isVi =u(xi) where
u:X →[0,¯v].If the distribution
Fu(l) = Pr (u(x)≤l), l ∈[0,v]¯
belongs to F then we can easily see that
bu(x) =bFu(u(x)) (2)
is a symmetric equilibrium bidding function.
Is it possible to fix a distributionF and vary the valuationu(·) to obtain a pre-specified bidding function b(·)? Suppose the set of signals isX = [0,v]¯ with distribution F (x) with a continuous density f(x) >0. Then we have the
Theorem 2 Suppose b : [0,v]¯ → R is continuously differentiable, strictly
increasing such that b(0) = 0 and
u(x) = b(x) + b
′(x)Fn−1(x)
(Fn−1)′(x) =
(b(x)Fn−1(x))′
(Fn−1)′(x) is increasing.
Then b(·) is the symmetric equilibrium of the first-price auction if bidders
Proof: First note thatFu(u(x)) =F (x) and therefore using (2) that
bu(x) = u(x)−
Ru(x)
0 F
n−1
u (l)dl
Fn−1
u (u(x))
=
u(x)−
Rx
0 F
n−1(l)u′(l)dl
Fn−1(x) =
Rx
0 u(l) (F
n−1)′ (l)dl
Fn−1(x) =
Rx
0 (b(l)F
n−1(l))′dl
Fn−1(x) =b(x).
References
1. Theory of the Integral, S. Saks, Dover Edition, 1964
! "#$% & #!' (! !)* !+, - .+'*! (# /001 - /2 34%5
6 6
- )# ! # ' '(#5 - .+ *! (# /001 - 70 34%5
8 " 9
#!' (! !+, & .# #:; :$<*& #$' (! !+, - .+ *! (# /001 - /0 34%5
80 = >? @ A
)# ! # $ & B %'# ! #5 - .+ *! (# /001 - C 34%5
87 6 /007
#!' (! !)* !+, & )$ .! %# ! #5 - .+ *! (# /001 - / 34%5
8/ 6 D 6
" .!E!"$)<! 55 # & 5*$( " *$( - %!5<! (# /001 - /2 34%5
81 9 6 $ $5<$' # & #(# $)! 5< #
'#$ ! " ( #5 - %!5<! (# /001 - /2 34%5
8 B
6 :+# (# F #+ #55! & $ $ <!5 #55! & G # !F - %!5<! (# /001
- 10 34%5
8C 9 A @ 0
- :+# (# F #+ #55! - %!5<! (# /001 - // 34%5
82 - !5 +%H'$! ( !5< - #<#:F ! (# /001 - 0
34%5
8 D 6 6 6 6 " !5 +%H'$! ( !5<
-#<#:F ! (# /001 - 7C 34%5
8 B 6 " I 9 I 6 !5 +%H'$! ( !5<
-#<#:F ! (# /001 - 10 34%5
88 9 B 6 !5 +%H'$! ( !5<
-#<#:F ! (# /001 - /2 34%5
C00 "
>J .!E! "$)<! 55 # & )*# !+<! # #$ - #<#:F ! (#
/001 - /8 34%5
C07 " )# ! # ' '(#5&
.! )*$: ::$% - #<#:F ! (# /001 - 1 34%5
C0/
C01 6 K L
M - #!' (! !)* !+, - #<#:F ! (# /001 - 7734%5
C0 6 6 6 B " $ D* '
- #<#:F ! (# /001 - 1 34%5
C0C 6 B 6
)! !'!:!& !5 "$ ' (# *! +<+F ! (# /001 -/ 34%5
C02 N !5 6 :$ <!' "
OP!& #' <! Q #5 . - +<+F ! (# /001 - / 34%5
C0 B 6 #( !
) '<$ # #$ & .!E! "$)<! 55 # & :+# (# F #+ #55! +<+F ! (# /001 -34%5
C0 6 " N '$# !<< $#F& 6+:F# <! ! #$
-+<+F ! (# /001 - /C 34%5
C08 6 #' <!
Q #5 . - +<+F ! (# /001 - 11 34%5
C70 " " " $ :
: # !& #' <! Q #5 . & #)R $! : $< - +<+F ! (# /001 - 12 34%5
C77 !R5$! OP!& .!5S P (! F )* '& $! 45)!
-! #:F -! (# /001 - 1 34%5
C7/ 6
!R5$! +P!& 6+:F# <! ! #$ - ! #:F ! (# /001 - 17 34%5
C71 D B 6 +P!& "
<$'5 ( !)* & D !'<#$ ! - ! #:F ! (# /001 - 7 34%K5
C7 " D B
!$5$! +P!& 4 )$ #!' - ! #:F ! (# /001 - C0 34%5
C7C T U V #' <!
Q #5 . - ! #:F ! (# /001 - / 34%5
C72 " " #' <!
Q #5 . - ! #:F ! (# /001 - / 34%5
C7 " > " # ' '(! (# 6! '(
F!5 - #,#:F ! (# /001 - 7 34%5
C7 = W T )# ! Q <#5 S $&
F$ '! ( $ $! '$'$ #,#:F ! (# /001 - /1 34%5
C78 6 )# ! Q <#5 S $& )# ! 9# #,
#,#:F ! (# /001 - / 34%5
C/0 B
C/7 6 6 6 6 B +F#'5
#'* ;5'# - . '#$ ! (# /00 - 72 34%5
C// 6 D )# !
# ' '(#5& + ! D $'%# !'<#$ ! - # # #$ ! (# /00 - /1 34%5
C/1 6 B " D !R5$! +P!&
6+:F# <! ! #$ & )!5 6 5+)*$( - # # #$ ! (# /00 - /7 34%5
C/ " N !R5$! +P!& 6+:F# <! ! #$ & )!5 6
5+)*$( - # # #$ ! (# /00 - /2 34%5
C/C $ 5#( : ' %# 5= ! % '$, <$!' (#5$%'= '( $')#'<$ # 3 ! $5$!' $5<$ '! !5< & '$# # #$ & 6+:F# <! ! #$ - # # #$ ! (# /00 - 77 34%5
C/2 '( < X#5 $' <$' :# $) ' )!'<#X< .+ $ '! . 55+'YE!& 6+:F# <! ! #$
-# -# -#$ ! (-# /00 78 34%5
C/ '($) (! #5 )!$')$(#'<#5 (# <$ $( (# #)!'Q:$) # +: ) !'! !%$ (# #)#55Z#5 3 !
5$ '%# ! . !'<K # '# + <#& .!E! "$)<! 55 # & '( #$ 3 )! # # #$ ! (# /00 - 7 34%5
C/ " *$P$# 9$ !& +$,
#' <! $: - Y! (# /00 - / 34%5
C/8 6 D 6 N
+$, #' <! $: & *$P$# 9$ ! - Y! (# /00 - /7 34%5
C10 B " 6 6 6 !;!+'% D$:& +$, #' <! $: - Y!
(# /00 - 12 34%5
C17 "
#( ! ) '<$ # #$ & $! ''$ ))*$'$ - Y! (# /00 /C 34%5
C1/ N 9 '%# ! .!5S
!'<K # '#& #( ! ) '<$ # #$ & 4 )$! '<Q'$! <! - Y! (# /00 - /C 34%5
C11 >? T T @ - # ' '(! (# 6! '( F!5 - Y! (# /00
- 72 34%5
C1 T 9 . [ - # ' '(! (# 6! '( F!5
- Y! (# /00 - 71 34%5
C1C 9 6 6
" )! !'!:!& #<$' <$'5 & !( $%! $'<! - Y! (# /00 - 18
34%5
C12 6 + ! D $'%# !'<#$ ! - Y! (#
/00 - 2 34%5
C1 B 6 6
9 + ! D $'%# !'<#$ ! & #' +X $<# & '\ 6 %# . - Y! (# /00