Enhancing Honesty in Bargaining Under Incomplete Information: An Experimental Study of the Bonus Procedure

Abstract

The sealed-bid k-double auction mechanism for two-person bargaining under incomplete information can be extended by providing a bonus for both traders if, and only if, an agreement is reached. Brams and Kilgour (1996, Group Decision and Negotiation, 5, 239-262) proved that there is a unique level of bonus, namely, half the difference between the buyer's bid and seller's ask, that induces the traders to bargain truthfully. Stein and Parco (2001, unpublished manuscript) extend the theoretical work to the case of partial bonus levels that reduce, but do not completely eliminate, misrepresentation of the reservation values. In the present study, we experimentally investigate the effects of providing different bonus values on the misrepresentation of the bargainers' reservation values. Our results show that implementation of a bonus has a significant impact on bargaining behavior but not nearly to the extent predicted by the equilibrium analysis. A reinforcement-based learning model originally proposed for the no bonus case accounts quite well for the results of all three experimental conditions.

Fair Cake-Cutting in Practice

“…all their actions are undominated). While this finding is in line with previous studies showing that human subjects often play dominated strategies (Artemov et al, 2017;Hassidim et al, 2016Hassidim et al, , 2017Rees-Jones, 2017;Parco and Rapoport, 2004), we find it somewhat surprising that so many people behave in a way that is so clearly irrational.…”

Price‐setting power and information asymmetry in sealed bidding

“…7. Designed to simultaneously test the previously reported finding of Rapoport et al (1998) Parco (2002) for a complete articulation of the method and justification of its applicability (see also Daniel et al, 1998;Rapoport et al, 1998;Seale et al, 2001). 12.…”

“…The model strives for parsimony using only four parameters comprised of both a linear function and a conjoined exponential function. This model (see also Seale et al, 2001) has repeatedly demonstrated its ability to closely approximate behavior of information-advantaged players with similar, although somewhat weaker results for information-disadvantaged players across a variety of treatments (Parco, 2002;Parco and Rapoport, 2005). The model makes no probabilistic assumptions nor does it require the estimation of initialization values for the parameters.…”

“…Subjects were individually seated and given a set of written instructions to read at his or her own pace. Once all subjects completed reading the instructions (see Parco, 2002) the experiment began. The same procedure was used for both experimental treatments reported in this study.…”

“…The original LES formulation published by Chatterjee and Samuelson (1983) contained several typographical errors, which have been verified (Parco, 2002) and correctly annotated in Equations (1)-(4) for any pair of prior probability distributions where F $ uniform½a s ; b s and…”