Using experimental auctions for marketing applications: A discussion

Journal of Agricultural and Applied Economics, Aug 2003 by Lusk, Jayson L

Market Feedback and Bidder Affiliation

As was discussed in the previous subsection, there are some potential advantages to allowing subjects to incorporate market feedback into their valuations. Specifically, a number of studies have emphasized the importance of an active market environment at generating rational behavior consistent with economic theory (e.g., Cherry, Crocker, and Shogren; Shogren; Shogren et al., 2001b). For marketing applications, the intuition is that subjects routinely use market information, such as market prices, in the "real world," and laboratory experiments should be designed with this fact in mind. With the second-price and random Nth price auctions, subjects can be provided market feedback by conducting multiple bidding rounds and posting market prices at the conclusion of each round. The English auction, by construction, permits feedback, not just about the final market price, but about all subjects' bids.

Despite the advantages of allowing market feedback with auction mechanisms, there is one potentially serious drawback. When subjects are allowed market feedback through posted prices, there is the potential for bidder's values to become affiliated (Milgrom and Webber). Bidder affiliation refers to the situation when a relatively high value for one subject implies high values for other subjects. The incentive capability property of the aforementioned auctions rests on the assumption that bidders' values are independently distributed. This assumption is clearly violated if values become affiliated. Thus, bidder affiliation has the potential to degrade the incentive compatibility of the aforementioned auctions.

List and Shogren (1999) directly tested whether bidder affiliation, caused by posting market prices, significantly influenced bidder behavior in the types of auctions commonly used in marketing applications. They found that bidder affiliation only had a very small impact on bids-posted prices increased median willingness-to-pay bids by 1%. They also found that bidder affiliation only existed for novel, but not familiar, goods. Another interesting observation in this regard is that Lusk, Feldkamp, and Schroeder and Rutstrom both found that BDM and English auctions generated statistically equivalent results. This finding is fascinating, because the English auction permits the greatest possible amount of market feedback (and thus has the greatest potential for bidder affiliation), whereas the BDM permits no market feedback (and thus has no bidder affiliation). On the surface, these results suggest that potential problems with bidder affiliation are likely outweighed by the potential advantages of market feedback ceteris paribus. Of course, this conclusion is not universal, and more theoretically minded experimentalists have reached the opposite conclusion (Harrison, Harstad, and Rutstrom).

Even if bidders' values do become affiliated over multiple bidding rounds, one interesting area of investigation that has received little, if no, attention is in determining why and how values become affiliated. For example, Lusk et al. (2002) recently conducted fifth-price auctions in a number of US and European locations, where they elicited the minimum amount of compensation subjects demanded to consume a genetically modified food. These values were elicited over 10 bidding rounds, with one of the rounds randomly selected to be binding. As is commonly the case in willingness-to-accept auctions, they found that median bids were relatively high in the first bidding round and generally trended downward as the experiment progressed. Certainly this behavior implies something about the relative importance of market prices and individuals' concern for biotechnology. That subjects demanded less compensation to consume a genetically modified food when they were informed of the market price might indicate that a portion of an individuals acceptance/rejection of biotechnology is a function of individuals' perception of acceptance in the market. In this and other similar situations, subject response to the posting of market prices might have meaningful interpretation beyond theoretical concerns with affiliation.