Abstract
We compare the expected revenue in first- and second-price auctions with asymmetric bidders. We consider “close to uniform” distributions with identical supports and show that in the case of identical supports the expected revenue in second-price auctions may exceed that in first-price auctions. We also show that asymmetry over lower valuations has a stronger negative impact on the expected revenue in first-price auctions than in second-price auctions. However, asymmetry over high valuations always increases the revenue in first-price auctions.
Similar content being viewed by others
Notes
The assumptions are: bidders are risk neutral, identical distributions over values (symmetry), anonymity of auction rules with respect to bidders and allocation of the prize to the bidder with the highest bid.
A recent study by Fibich et al. (2012) proves that this result holds under the assumptions of interchangeability and differentiability for any asymmetric model, not necessarily auctions or even economic models.
For the numerical simulations we used both Matlab and Maple to verify our calculations. The errors in the numerical calculations shown in the current study are less than the least significant digit we present. In other words, if \(\Delta R=-0.00045\), the error is \(10^{-6}\) or less. Further discussion can be found in Fibich and Gavious (2003) and Fibich and Gavish (2012). A Matlab code by Fibich and Gavish (2012) for solving first-price auctions with two bidders can be found in http://www.math.tau.ac.il/~fibich/publications.html#_Matlab_codes_for.
As we noted earlier, although the difference is in the sixth digit, the simulation error is less than \(10^{-7}\).
By Lebrun (2009), this assumption is satisfied up to order \(\varepsilon \). Fibich and Gavish (2012) proved (see Theorem 3.6) that when distributions have the structure \(v^{a+\varepsilon h}\), equilibrium bid functions are smooth with respect to \(\varepsilon \). In other words, expansion exists for all orders of \(\varepsilon \).
Cheng (2011) considers a first-price auction with two bidders and a discrete probability distribution, and finds that \(R_{sym}<R^{first}\) where the symmetric benchmark is the geometric average.
A recent study by Hubbard et al. (2012) shows a similar result numerically for two bidders, one with uniform distributions and the second with polynomial distributions.
The expansion of \(z(w)\) and \(z^{\prime }(w)\) at \(w=0\) gives \(z(w)=z^{\prime }(0)w+O\left( w^{2}\right) \) and \(z^{\prime }(w)=z^{\prime }(0)+O\left( w\right) \).
The second-order condition for optimality is not satisfied for the problem \( \min _{z(w)}\Delta R\), which indicates that there is no minimum.
References
Cantillon E (2008) The effect of bidders’ asymmetries on expected revenue in auctions. Games Econ Behav 62:1–25
Cheng H (2006) Ranking sealed high-bid and open asymmetric auctions. J Math Econ 42:471–498
Cheng H (2011) Asymmetry and revenue in first-price auctions. Econ Lett 111:78–80
Fibich G, Gavious A (2003) Asymmetric first-price auctions—a perturbation approach. Math Oper Res 28:836–852
Fibich G, Gavious A (2010) Asymptotic revenue equivalence of asymmetric auctions with interdependent values. Eur J Oper Res 206:496–507
Fibich G, Gavious A, Sela A (2004) Revenue equivalence in asymmetric auctions. J Econ Theory 115:309–321
Fibich G, Gavious A, Solan E (2012) An averaging principle for second-order approximation of heterogeneous models with homogeneous models. Proc Natl Acad Sci USA 109(48):19545–19550
Fibich G, Gavish N (2012) Asymmetric first-price auctions: a dynamical-systems approach. Math Oper Res 37(2):219–243
Gavious A, Minchuk M (2012) The effect of asymmetry on revenue in second-price auctions. Int Game Theory Rev 14(3):1–8
Gayle WR, Richard JF (2008) Numerical solutions of asymmetric, first-price, independent private values auctions. Comput Econ 32:245–278
Hubbard TP, Kirkegaard R, Paarsch HJ (2012) Using economic theory to guide numerical analysis: solving for equilibria in models of asymmetric first-price auctions. Comput Econ. doi:10.1007/s10614-012-9333-z
Kaplan T, Zamilr S (2012) Asymmetric first-price auctions with uniform distributions: analytic solutions to the general case. Econ Theory 50(2):269–302
Kirkegaard R (2009) Asymmetric first-price auctions. J Econ Theory 144:1617–1635
Kirkegaard R (2012) A mechanism design approach to ranking asymmetric auctions. Econometrica 80:2349–2364
Klemperer P (1999) Auction theory: a guide to the literature. J Econ Surv 13:227–286
Lebrun B (1996) Revenue Comparison Between the First and Second Price Auctions in A Class of Asymmetric Examples. Mimo (unpublished paper)
Lebrun B (2006) Uniqueness of the equilibrium in first-price auctions. Games Econ Behav 55:131–151
Lebrun B (2009) Auctions with almost homogeneous bidders. J Econ Theory 144:1341–1351
Li H, Riely J (2007) Auction choice. Int J Ind Organ 25:1269–1298
Marshall RC, Meurer MJ, Richard J-F, Stromquist W (1994) Numerical analysis of asymmetric first price auctions. Games Econ Behav 7:193–220
Maskin E, Riley J (2000a) Asymmetric auctions. Rev Econ Stud 67:413–438
Maskin E, Riley J (2000b) Equilibrium in sealed high bid auctions. Rev Econ Stud 67:439–454
Maskin E, Riley J (2003) Uniqueness of equilibrium in sealed high-bid auctions. Games Econ Behav 45:395–409
Myerson RB (1981) Optimal auction design. Math Oper Res 6:58–73
Plum M (1992) Characterization and computation of Nash-equilibria for auctions with incomplete information. Int J Game Theory 20(4):393–418
Riley JG, Samuelson WF (1981) Optimal auctions. Am Econ Rev 71:381–392
Vickrey W (1961) Counterspeculation, auctions, and competitive sealed tenders. J Finance 16:8–37
Author information
Authors and Affiliations
Corresponding author
Appendix
Appendix
1.1 A: Auxiliary Lemma
Lemma 2
Proof
Substituting \(v_{i}(b)=v_{sym}(b)+\varepsilon V_{i}(b)+\varepsilon ^{2}U_{i}(b)+O(\varepsilon ^{3})\) in \(F_{i}(v_{i}(b))\), expanding the series near \(v_{sym}(b)\) and collecting the terms in \( \varepsilon \) gives
where \(\varepsilon H_{i}((v_{sym}(b)+\varepsilon V_{i}(b)+\varepsilon ^{2}u_{i}(b)+O(\varepsilon ^{3}))=\varepsilon H_{i}(v_{sym}(b))+\varepsilon ^{2}V_{i}(b)H_{i}^{\prime }(v_{sym}\) \((b))+O(\varepsilon ^{3})\). Substituting \( v_{sym}(b)=2b\) (14) completes the proof. A similar calculation proves (24).\(\square \)
1.2 B: Proof of Theorem 1
The following notations are used later on in this proof. Let
where \(V(b)\) is given by (16) and
Substituting (12, 14, 23) in \(R^{first}\) (given by 3) we get
(Observe that we substitute \(\overline{b}=\overline{b}_{sym}+\varepsilon ^{2} \overline{b}_{2}+O(\varepsilon ^{3})\) only in the second integral’s upper limit where the missing \(O(\varepsilon ^{3})\) in the integral’s limit is part of the \(O(\varepsilon ^{3})\) at the end of the equation). Substituting \(\overline{b}-\int \nolimits _{0}^{\overline{b} }4b^{2}db=R_{sym}^{first}+O(\varepsilon ^{3})\) given by Lemma 3 below, \(\overline{b}_{sym}=0.5\) (see Eq. 4) gives (Observe that the terms with the \(\overline{b}_{2}\) are \(O(\varepsilon ^{3})\) )
Denote \(u(b)=u_{1}(b)+u_{2}(b)\). Rearranging yields
Substituting \(2\displaystyle \int \nolimits _{0}^{0.5}{bu(b)db}\) given by Lemma 5 below gives
Substituting \(V(b)\) given by ( 25) in (27) we get
where \(y(b)\) is given by (26). Observe that
and
Substituting (29, 30) in (28) gives
Integration by parts of \(\int \nolimits _{0}^{0.5}(1-2b){H^{\prime }(2b)y(b)db}\) gives
Substituting (32) in (31) yields
Simple calculation shows that
and thus
Substituting (34) and (35) in (33) we get
By integration by parts (observe that \(y(1)=0\)) we have
and
Substituting (37) and (38) in (36) gives
Transformation of \(w=2b\) and substituting \(z(w)=y(w/2)\) given by (19 ) yields the result. \(\square \)
Lemma 3
\(\overline{b}-\int \nolimits _{0}^{\overline{b} }4b^{2}db=R_{sym}^{first}+O(\varepsilon ^{3}).\)
Proof
Since \(\bar{b}_{1}=0\) and \(\overline{b}_{sym}=0.5\) we have \( \overline{b}=\overline{b}_{sym}+\varepsilon \bar{b}_{1}+\varepsilon ^{2} \overline{b}_{2}+O(\varepsilon ^{3})=0.5+\varepsilon ^{2}\overline{b} _{2}+O(\varepsilon ^{3})\). Thus,
\(\square \)
Lemma 4
where \(u(b)=u_{1}(b)+u_{2}(b).\)
Proof
The equilibrium equations are given by
Substituting (13, 24, 23), \(v_{sym}(b)=2b\) and \(v_{i}^{\prime }(b)\) (obtained by differentiating 13) in (41) we get
To simplify the calculation we define
when \(\varepsilon \) is sufficiently small, \(\varepsilon H_{i}^{\prime }(2b)+\varepsilon ^{2}V_{i}(b)H_{i}^{\prime \prime }(2b)<1\). Then, we can use Taylor’s expansion \(\frac{1}{1+\varepsilon C}=1-\varepsilon C+(\varepsilon C)^{2}+O(\varepsilon ^{3})\) in \(A\) and \(B\) above. The expansion for A is given by
Rearranging gives
Since \(0\le {\frac{{\varepsilon V_{j}(b)+\varepsilon ^{2}u_{j}(b)}}{{b}}<<1} \) for small \(\varepsilon \) ( it easy to see that \(\frac{{ V_{j}(b)}}{{b}}\) is bounded for small \(b\) ), the Taylor expansion of B is given by
Multiplying A and B we get
Rearranging (42), comparing the \(O\left( \varepsilon ^{2}\right) \) order coefficient (\(O(1)\) and \(O(\varepsilon )\) are known as shown in Sect. 3) and substituting \(V_{1}=-V_{2}\) (15) we get
Let \(u(b)=u_{1}(b)+u_{2}(b)\). Then
Equation (43) is a first order ODE, \(u^{\prime }+P(b)u=g(b)\). By (2) the boundary condition is \(u(0)=0\). The solution of (43) is given by :
where \({P(s)=}\frac{1}{s}\) ,
and \(m\) is a constant. Since
we have
By the boundary conditions \(u(0)=0\) it follows that
Substituting \(m\) in \(u\) gives
Substituting (44) in (45) yields the result. \(\square \)
Lemma 5
Proof
Let us find \(\int _{0}^{0.5}2bu(b)db\) where \(u(b)\) is given by Lemma 4.
Integrating by parts, multiplying by 2 and using the boundary condition \( u(0)=0\) we get
Since \({V(b)H^{\prime }(2b)}+{2bH^{\prime \prime }(2b)V(b)=V(b)(H}^{\prime }{ (2b)b)}^{\prime }\), we have
Substituting \(V^{^{\prime }}(b)={\frac{3{V(b)+H(2b)}}{b}}-2{H^{^{\prime }}(2b)}\) (obtained by differentiating \(V(b)\) defined in Eq. 16) and integrating by parts gives
The substitution of (47) in (46) completes the proof. \(\square \)
Rights and permissions
About this article
Cite this article
Gavious, A., Minchuk, Y. Ranking asymmetric auctions. Int J Game Theory 43, 369–393 (2014). https://doi.org/10.1007/s00182-013-0383-9
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00182-013-0383-9