| Vitae (in pdf) | VLAD MARES |
| Visiting Assistant Professor of Economics | |
| Research | |
| Address | |
Articles Ex-post full surplus extraction, straightforwardly (with Ron Harstad) |
|
| Kellogg School of Management | |
| MEDS | |
| Northwestern University | |
| 2001 Sheridan Rd | |
| Evanston, IL 60208 | |
| Email: | |
| v-mares@kellogg.northwestern.edu | |
| Recent papers Procurement Auctions (with Jeroen Swinkels) In a number of settings, of which procurement is a key example, there will be natural asymmetries in costs (or values). One supplier may be viewed as more reliable, or one supplier may currently be better equipped to handle a certain job while another needs to make specific investments. In such settings, we study the question of which simple mechanisms are best, and how they relate to optimal mechanisms. We find a surprisingly strong result. Under a wide variety of conditions on cost distributions, a variation on a second price auction with bonuses always outperforms (on an outcome by outcome basis) any of a class of first price mechanisms that includes on one extreme the Request for Proposal mechanism commonly used and on the other, a standard first price auction. Asymmetric First-Price Auctions (with Jeroen Swinkels) We study the implications of rho-concavity on the shape of the surplus functions in asymmetric first prices auctions. This paper generalizes a known result of Prekopa and Borell and explores the applications to auction theory and mechanism design. On the Near-Optimality of Second Price Auctions (with Jeroen Swinkels) (Coming in Nov. 2009) Consider a setting with n sellers having independently distributed costs. We show that in an environment where buyer preferences matter a simple second price mechanism with appropriately chosen bonuses comes surprisingly close to the performance of the optimal mechanism. Further Results in Asymmetric Auctions (with Jeroen Swinkels) (Coming in Dec. 2009) We analyze the comparative statics of handicaps in asymmetric auction environments where the seller's preferences matter. We find that optimal distortions never reflect the true preferences of the designer. We show that for an entire class of problems a simple second price mechanism outperforms on an ex-post basis all distorted first-price mechanisms. The Competitive Nature of Syndicates (with Mike Shor) Firms commonly form syndicates to bid jointly for financial assets. Recently, this practice has come under legal scrutiny out of concern that syndicates decrease competition by having firms bid with rather than against each other. Recent research indicates that joint bidding in common-value auctions is anticompetitive. However, these models do not account for one important feature of financial markets: bidders' estimates of the value of the item for sale are likely to be correlated with each other. We show that correlated valuations make it likely that joint bidding is pro-competitive. The result holds even if syndicates lead to a highly concentrated market. Preemption and Jump Bidding This paper constructs an integrated model of preemption and joint bidding in open auctions. The decision to discretely increase the price is endogenous and relies purely managing the flow of information among a bidder's competitors. Bidders' differentiated sensitivities to the information revealed in the auction process provides incentives to manipulate and censor the price formation process in their favor by using simple devices like opening bids or discrete jumps. In doing so they affect both the expected price and the allocation of the asset. Increased competition can actually strengthen these incentives. The paper characterizes optimal bidding plans and derives comparative statics. Information Concentration (with Mike Shor) We consider how information concentration impacts the revenues of a seller of a common value object. The common value is a function of n random variables partitioned among m ≤ n bidders. For each partition, the seller devises an optimal mechanism. We show that, whenever the value function allows scalar sufficient statistics for each player's signals, the mechanism design problem is well-defined. Additionally, whenever a common regularity condition is satisfied, a coarser partition always leads to lower revenues than a less coarse partition. In particular, any merger or collusion among bidders reduces revenue. Relative Performance Auctions (with Octavian Carare) We study a multi-unit auction environment in which the agents who carry out the bidding are rewarded according to the ranking of their surpluses. We show that in a symmetric equilibrium of our auction game if one exists an efficient allocation may not be attainable. Our auction mechanism may be viewed as combining the features of a discriminatory auction and of a tournament. The inefficient outcome of our auction is remarkable given the efficiency of each of its components |
|
| Phone: (847)491-8813 Fax: (847)467-1220 |
|
| Work in Progress Monotonicity and Selection “Monotonicity and Selection”- examines the set of equilibria of asymmetric common value auctions. It is a well-known fact that the set of monotone equilibria is infinite but this raises two additional questions. Do common value auctions admit non-monotone equilibria? The answer in general is yes, if bidders have multi-dimensional signals. Even in the case where three or more bidders with single-dimensional signals compete non- monotone equilibria can not be ruled out. On the other hand, the paper establishes that in two-bidder auctions with single-dimensional signals only monotone equilibria survive. For practical applications this raises the issue of selection. While in symmetric contexts, the unique symmetric equilibrium is the natural candidate, what is its counterpart in asymmetric situations? The paper argues that the revenue maximizing equilibrium is the appropriate choice for asymmetric auctions. This is a consequence of the fact that, when it comes to equilibrium selection, symmetry is equivalent to revenue maximization in symmetric auctions Bundling in Low Competition Environments This paper looks at possible explanations for deviations from the optimum when the key design variable is the decision to bundle. An impressive host of literature suggests that the bundling decision is straight-forwardly linked to the level of competition. For two bidders the revenue form bundled auctions exceed those for separate ones, while for more bidders (the number is distribution specific but for most commonly used distributions four is a good benchmark) the revenues from separate auctions will outweigh the resulting equilibrium prices from bundled auctions. The paper suggests two common causes of reversal for these relatively robust results. When bidders are uncertain about the distribution about their opponents values, an informed auctioneer will find it in his best interest to credibly communicate his private information through the auction design, in particular bundling in high competition environments and unbundling in low competition ones. Furthermore the presence of multiple sellers can have similar effects, where competition among them can dissipate rents sufficiently fast in order to lead to the adoption of the least preferred design |
|
| Slides Preemtion and Jump Bidding Procurement Auction |
|
| Teaching Winter '10 |
|
| . |