Nancy L. Ertle Professor of Marketing; Faculty Director, Program on Data Analytics at Kellogg
How much of the truth should you tell about what you are trying to sell?
Check out more from The Trust Project at Northwestern University here.
That is a question faced by anyone who has ever auctioned a non-mint-condition item on eBay or elsewhere, whether a chair with a stained cushion, a vintage toy with chipped paint, or a laptop with a weak battery. Most sellers would be tempted to withhold some or all of the information about flaws, based on the conventional wisdom that it would scare off potential buyers.
But what if that wisdom was wrong?
What if disclosure of negative information could actually boost the likelihood of a sale? Florian Zettelmeyer, professor of marketing at the Kellogg School, and his collaborator Steven Tadelis of University of California Berkeley used that line of questioning as the basis for a study on automobile auctions. They found that under certain circumstances honesty on the seller’s part is indeed the best policy—even as related to the sale of a lower-quality item. Specifically, the researchers showed that an auction house made more sales when it disclosed more quality-related information about used cars up for auction, even if that information was negative.
The researchers first set out to examine the mechanisms of auction pricing. This is an area with a long history in economic theory. According to the “linkage principle” first proposed by economists in 1982, auctions in which all known information about quality is disclosed up front will drive higher bids on higher-quality items, compared with auctions in which the quality of the item is not disclosed. According to the theory, this disclosure process will also depress bids on lower-quality items. But the net benefit of pushing bids up on higher-quality items outweighs the disadvantages regarding lower-quality items.
“It seems like a very bizarre result. You tell people something that will decrease their valuations for a vehicle relative to a standard, and the next thing that happens is that, on average, you make more money.”
To test this theory empirically, the authors designed a controlled experiment with the help of an auction house with a large business in used-automobile sales. Zettelmeyer and Tadelis took advantage of the intensive inspection procedure through which the business placed its cars. The researchers randomly varied the level of information disclosure provided in 8,100 automobile auctions over a 19-week period. Roughly half the auctions included qualitative information from the inspection. For the other half, the auction house knew the inspection results but did not disclose related information to bidders as part of the sale process.
The results of the experiment showed that disclosing qualitative information about used cars for sale by auction increased the probability of a sale—and thus potential revenue for the auction house—by an average of 16 percent, even when the information was negative. “It seems like a very bizarre result,” Zettelmeyer said. “You tell people something that will decrease their valuations for a vehicle relative to a standard, and the next thing that happens is that, on average, you make more money.”
Zettelmeyer explains this counterintuitive result with a specific mechanism: the “job” the disclosed information does for bidders. Intuitively, it seems obvious that a buyer’s goal in obtaining qualitative information about an auction item is to help them evaluate the item objectively and bid accordingly. For example, the bidder may see a laptop with a cracked screen as worth only $50; but the same laptop with an intact screen might be worth many times that to the same buyer.
But what if the “job” of disclosure is not to help a bidder evaluate a particular auction item, but rather to help the buyer decide in which auction markets to participate? Zettelmeyer calls this a “matching mechanism,” and it applies to auctions in which many different items of a similar class—like the used cars in this experiment—are bid upon simultaneously. In this scenario, an individual bidder cannot participate in all auctions at once, and thus must sort through (or match) available options based on the kind of car he or she wishes to buy.
The disclosed information—whether good or bad—acts as a signal bidders can use to guide the matching process. For example, a luxury car dealer might gravitate to an auction for a lightly used, late-model BMW sedan. Meanwhile, according to Zettelmeyer, “an inner-city dealer that lives in a low-income neighborhood [might be] more interested in buying a car that is in worse condition.” He continued, “So the role of information here isn’t necessarily to tell you whether the car is good or bad. Instead, the information allows you to decide what kind of car you want to bid on in the first place. The disclosure helps people select and do what’s most relevant for them in these situations.”
The matching mechanism drives what Zettelmeyer calls “very pro-competitive effects.” Specifically, the researchers found that the probability of making a sale on any item, regardless of its quality, rises by an average of 16 percent when information is disclosed. This happens because the bidders self-sort into “lanes” where they can compete against others with similar buying interests, but only on the auctions most relevant to them. “A Mercedes dealer is not going to be able to sell an old Toyota Corolla on his lot, because nobody’s going to be looking for a Corolla on a Mercedes dealer’s lot,” Zettlemeyer said. But dealers of used Toyotas will be happy to compete against one another for that same “undesirable” vehicle.
According to Zettelmeyer, this matching mechanism applies to “virtually any online auction” and to many other economic activities. “The one assumption that needs to hold [for this mechanism to apply] is that you are constrained in how many items you can bid on,” he says. That means the matching mechanism is relevant to government procurement or job recruitment, including when vendors are bidding on projects. For example, if a municipality is seeking a contractor to improve its roads, Zettelmeyer’s findings imply that it ought to disclose significant information not just about the next road contract, but also about the road contracts planned for the next several years.
In this case, the vendors sort themselves into the project-related lanes that are most relevant or desirable for them, just as bidders do when faced with multiple potential auctions to join.
“A vendor can’t do every road contract, so once they bid on one, they will be maxed out for a given time and they can’t do another one,” Zettelmeyer explains. “And so by giving them this view of what’s coming up, you allow them to sort themselves into the contracts that they find are the most interesting for them. And as a result of that, you basically start creating groups of bidders who are very interested in a particular contract and thereby willing to put on a particularly cheap price.”
As for auction houses, the implications of the research are very clear: greater disclosure means higher participation in auctions, which means more revenue. “Give away inspections—or at least massively subsidize them,” Zettelmeyer says. “It looks like you’re doing it out of the goodness of your heart to keep everybody honest and to increase transparency in the marketplace. But it also has a direct impact on your bottom line.”
Tadelis, Steven and Florian Zettelmeyer, “Information Disclosure as a Matching Mechanism: Theory and Evidence from a Field Experiment.” (2015), American Economic Review, February, Vol. 105 (2), pp. 886-905.