Sealed bid auctions with ambiguity: Theory and experiments

Abstract

This study presents a theoretical model and laboratory experiment of the first and second price sealed bid auctions with independent private values, where the distribution of bidder valuations is unknown. We derive the symmetric equilibrium using the α-MEU framework. We then test the theoretical predictions in the laboratory. In our experimental setting, ambiguity aversion is rejected in favor of ambiguity loving. Our results suggest that decision makers' ambiguity attitudes are context dependent. Another departure from previous experimental studies is the use of subjects as auctioneers. We find that compared to zero reserve prices the presence of auctioneers significantly reduces revenue in first price auctions. It also significantly reduces bidder earnings and efficiency. Without knowledge of the distribution of bidder valuations and with auctioneers, the first and second price auctions generate the same amount of revenue.

The visible hand: finger ratio (2D:4D) and competitive bidding

“…Given this interaction, the coefficient for the digit ratio then measures the correlation of the digit ratio and bids if values are relatively high (i.e., when the variable "low value" is zero). 12 Table 8 in the appendix reports the results for the entire sample, by gender, and by major ethnic subsamples. We do find that our previous observations for different ethnicities hold also for high values.…”

Condorcet meets Ellsberg

“…Proposition 4.2 of that paper shows that independence requires that the set of priors consists of product measures. Formally, signals are conditionally independent if the signal of every voter i is (·|s)-independent for each s ∈ S. Many papers studying games with incomplete information under ambiguity implicitly or explicitly assume independence, such as Bose et al (2006), Chen et al (2007), and Bose and Daripa (2009). 23 With this definition in mind, consider the set of priors defined by π ∈ if and only if there exists a p ∈ [p p] and an r s in the convex hull of { i∈I r s i : r s i ∈ R s ∀i ∈ I} for each s ∈ S so that π(a t) = pr a (t) and π(b t) = (1 − p)r b (t) for all (s t) ∈ , where each R s is a closed, convex, nonempty set of probability distributions over T .…”

Common Value Auctions with Voluntary and Qualified Entry

“…In the field, bidders in highly complex auctions are generally professionals who specialize and self-select into that activity. These factors are conjectures based on a notion of ambiguity aversion (e.g., Chen, Katuscák and Ozdenoren [2007]). …”