Independent scholarship · Open accessThe monthly digest

Indexed text attachment

games-and-information-in-higher-categories.tex

Source for Games and Information in Higher Categories

View paper · Download original file

Lines 4821–4920 of 6559. Source text is displayed literally.

4821argument does not involve probabilities. In the first-price auction a higher bid buys insurance against4822losing, and risk-averse bidders bid more than risk-neutral ones would. So the first-price auction yields4823strictly higher expected revenue \cite{MR84b}. A risk-averse \emph{seller}, by contrast, prefers the4824first-price auction for a different reason: its revenue, being a function of the highest value alone, is4825less variable.4826 4827\section{Common values}4828 4829Now let the object have a common value $V$ and let each bidder observe a signal $x_i$ correlated with it.4830 4831\begin{example}[The wallet game]\label{ex:wallet}\Scl[\cite{Kle98}]4832Two bidders each know the amount $x_i$ in their own wallet, independent and uniform on $[0,1]$; the prize is4833the sum $V=x_1+x_2$, sold by second-price auction. A bidder's unconditional estimate of the prize is4834$x_i+\tfrac12$. But if she wins against an opponent who bids increasingly in her signal, the opponent's4835signal is below her own, and the estimate \emph{conditional on winning} is $\tfrac32x_i$, which is lower4836unless $x_i=1$. The symmetric equilibrium is $b(x)=2x$: each bids what the prize would be worth if the4837opponent's signal were equal to her own. Then the winner pays $2x_j$ for a prize worth $x_i+x_j$ and earns4838$x_i-x_j>0$, and the loser has no regret. Expected revenue is $\tfrac23$, out of an expected value of $1$.4839\end{example}4840 4841The \term{winner's curse} is the failure to condition on the event of winning. The bidder who wins is the4842one with the most optimistic signal, so winning is bad news about the value; a bidder who bids her4843unconditional estimate will, on average, regret winning. Rational bidders shade, and shade more the more4844competitors there are, since beating more rivals is worse news. The curse is a statement about selection4845of exactly the kind met in \cref{ch:adverse}: the event on which the payoff is realized is correlated with4846the unknown, and the relevant expectation is the one conditional on that event.4847 4848\begin{theorem}[Milgrom--Weber]\label{thm:MW}\Scl[\cite{MW82}]4849Suppose the bidders' signals and the value are \emph{affiliated}: their joint density $g$ satisfies4850$g(z\vee z')\,g(z\wedge z')\ge g(z)\,g(z')$. Then4851\begin{enumerate}[label=(\roman*)]4852\item the second-price auction has the symmetric equilibrium4853  $b(x)=\E\bigl[V\mid x_i=x,\ \max_{j\ne i}x_j=x\bigr]$;4854\item expected revenues are ranked: English $\ge$ second-price $\ge$ first-price;4855\item a seller with private information affiliated with the value raises expected revenue by committing to4856  disclose it.4857\end{enumerate}4858\end{theorem}4859 4860Affiliation says that $\log g$ is supermodular on the lattice of signal profiles: it is the complementarity4861condition of \cref{ch:order} applied to a density, and it means that high values of some variables make4862high values of the others more likely. Parts (ii) and (iii) are instances of the \emph{linkage principle}:4863the more the price paid by the winner is linked to information affiliated with her own signal---the4864runner-up's bid, the dropout prices of others, the seller's announcement---the less of an informational4865rent her private signal commands, and the higher the revenue. This is the informativeness principle of4866\cref{thm:informativeness} in the service of the other side: there, tying pay to informative signals4867reduced the cost of incentives; here, tying price to them reduces the cost of eliciting values.4868 4869Affiliation also breaks revenue equivalence, and more drastically than (ii) suggests. With correlated4870private values the seller can extract the \emph{entire} surplus by a mechanism that offers each bidder,4871alongside a second-price auction, a side bet on the others' reports that is fair for her true type and4872ruinous for any other \cite{CM88}. That such mechanisms are never seen is a standing reminder that the4873common-prior, risk-neutral, unlimited-liability model is an idealization, and that robustness to its4874failure is a design criterion in its own right.4875 4876\section{Asymmetric information among bidders}4877 4878If one bidder knows the common value and another has only public information, the uninformed bidder cannot4879profit: any bid that wins against an informed rival wins precisely when the object is worth less. In the4880first-price equilibrium the uninformed bidder randomizes, earns zero in expectation, and by her presence4881forces the informed bidder to pay a price that averages the value over the worse states \cite{EMW83}. The4882informed bidder's rent is the value of her information, and it comes out of the seller's revenue. Oil-lease auctions in which one firm owns the neighbouring tract are the4883canonical application, and one of the cleaner empirical confirmations of auction theory.4884 4885\section*{Notes}4886 4887The theory begins with Vickrey \cite{Vic61}, who introduced the second-price auction and proved revenue4888equivalence for uniform values. The general theorem and the optimal auction are due to Myerson4889\cite{Mye81} and to Riley and Samuelson \cite{RS81}. The comparison of an extra bidder with an optimal4890reserve is from Bulow and Klemperer \cite{BK96}. Risk aversion is treated in \cite{MR84b}. The general4891symmetric model with affiliated values, and the linkage principle, are from Milgrom and Weber4892\cite{MW82}; the wallet game is Klemperer's \cite{Kle98}. Full surplus extraction is due to Cr\'emer and4893McLean \cite{CM88}. The model with one informed bidder is from \cite{EMW83}. A standard text is4894\cite{Kri10}.4895 4896\section*{Exercises}4897 4898\begin{exercise}4899Derive the first-price equilibrium bid for $n$ bidders with values uniform on $[0,1]$ both from the4900first-order condition and from revenue equivalence, and verify that expected revenue is $\frac{n-1}{n+1}$.4901\end{exercise}4902\begin{exercise}4903For two bidders with uniform values, compute expected revenue of a second-price auction with reserve $r$,4904and verify the maximum $\tfrac5{12}$ at $r=\tfrac12$.4905\end{exercise}4906\begin{exercise}4907Use revenue equivalence to find the symmetric equilibrium of a \emph{third-price} auction with three4908uniform bidders. Do bidders bid above or below their values?4909\end{exercise}4910\begin{exercise}4911In the wallet game, verify that $b(x)=2x$ is an equilibrium and that it is an ex post equilibrium: neither4912bidder would change her bid after learning the other's signal. Compute the expected revenue of the4913first-price auction, whose symmetric equilibrium is $b(x)=x$, and compare.4914\end{exercise}4915\begin{exercise}4916Show that independent signals are affiliated, and that if $(x,y)$ are affiliated and $h$ is increasing then4917$\E[h(y)\mid x]$ is increasing in $x$.4918\end{exercise}4919\begin{exercise}4920Explain, using \cref{sec:thompson}, why the Dutch and first-price auctions are identified by Thompson moves