An assessment of four popular auction mechanisms in the siting of NIMBY facilities: some experimental evidence

Abstract

The issue of locating locally unfriendly but socially beneficial facilities such as landfills and power stations is an important public policy concern in many countries. Local residents in the area where such facilities are to be located tend to exhibit strong opposition, no doubt due to the asymmetric distribution of the costs and benefits of such not-in-my-backyard (NIMBY) facilities. A potentially useful mechanism for the siting of such facilities is by compensation auctions, which attempt to incorporate the market mechanism into the decision making process. In such auctions, communities name the compensation they require to host such facilities, and the community demanding the least amount of compensation gets to host the facility. This research attempts to evaluate the performance of such compensation auctions using laboratory exepriments. Four popular auction formats are evaluated: first- and second-price and all-pay first-and second-price sealed-bid auctions. The latter two formats correspond to the compensation auctions with penalty payments proposed by Kunreuther and Kleindorfer (1986) and Quah and Tan (1998), who claim that these auctions are more efficient as they restrains strategic (or over) bidding. Our results, however, contradict this claim. We show that the first-and second-price auctions without penalty payments are in fact more efficient, in that they tend to minimize social costs, and truthful bidding is more likely.

Acknowledgement

We thank James Ang for his very able research assistance, and Alvin Eng, Lee Wei-Bing and Lee Weilin for their help in conducting the experiments. Research funding from the National University of Singapore is gratefully acknowledged.

Notes

¹ Subjects in the all-pay auction formats took significantly longer to submit their bids, as a result fewer rounds were completed within the alloted time.

² For details, see Appendix A which lists the complete instructions for the second-price auction.

³ Ties were broken randomly by the program using a random number generator.

⁴ Both outliers came from the same player, bidder number 6 in rounds 9 and 11 in Session 4. His costs in these rounds were e$1297 and e$1294. Frustrated by the high costs assigned to him, he put in bids of e$1600 and e$12940, respectively.

⁵ With the exception of one outlier in session 4, round 13, where Bidder 6 submitted a bid of 1 when his cost is very low, all the other five outliers involve extremely high bids when the bidders’ costs were high.

⁶ The test statistic follows a Chi-square distribution with 2 degrees of freedom.

⁷ The classic correction for heteroscedasticity is the HC0 estimator proposed by White (1980). MacKinnon and White (1985) discussed three improvements, HC1, HC2 and HC3. Long and Ervin (2000), in a comprehensive evaluation, conclude that HC3 is the best among all HC estimators they evaluated.

⁸ The robust SEs also provide a remedy for any potential nonnormality problem of the error terms, since normality assumption is required only for hypothesis testing. The robust SE estimates thus ensures that the t-tests we perform below using the SE are valid.