Collusion and Tricks
-by Kate Reynolds
One in a Series of Articles from Agorics, Inc.Which auction type is the best? As discussed earlier, there is no single answer to that question, but one way of approaching that subject is to think about collusion; the extent to which incentives to collude vary under different auction types can help determine which type of auction to choose. All auction types are susceptible to collusion, but degrees of incentives vary.
It works like this: A ring of 5 people agree to band together and bid on a necklace. Frank is appointed the designated bidder, and he buys the necklace for $50. After the auction, the ring members meet and hold another auction at which the necklace is sold to Blanche for $100. Now, from the $100, Blanche reimburses Frank for the original $50 he paid out, and she pays each other person (including Frank and herself) $10. So, Blanche ends up paying $90 for a necklace worth about $100, and the others make a $10 profit. (In theory it is possible for some ring member never to win a lot and just collect payments from the others, but in practice such a person would be ejected from the ring.)
Interestingly, cartels themselves often have an enforcement problem because it is in the private interest of each member to cheat the others. They are aided in this to some extent because generally the auctioneer does not reveal the name of the winner, and this creates opportunity for ring members to cheat each other and bid privately. One theory is that open formats (English, Dutch) ensure that no ring member exploits the ring by making a side deal simply because it is so easy to monitor compliance amongst ring members. A cheater can be seen calling out his bid.
In the English format, it is thought that collusion problems diminish as the number of bidders increases. Intuitively, a ring will work best if the majority of bidders belong. Intuitively, the more bidders there are, the less likely it is they will all be ring members.
However, intuition is sometimes wrong. Consider the fact that rings are usually formed by choosing only those bidders who might be expected to know that an item has special characteristics causing it to be more valuable than the vendor knows. In this case, a ring would consist of only informed bidders, and uninformed bidders would not be included in the ring. There might indeed be vigorous bidding in the early stages and then when the highest uninformed bidder has finished, one ring member would step up and buy the item. Such an arrangement would actually help the ring members: auctioneers are constantly on the lookout for rings (or possible rings) and vigorous initial bidding would help allay fears.
Generally, it is assumed that such a cartel credibly reveals private information to each other and then chooses a "designated winner" and a bidding strategy. Usually the designated winner is the ring member who has the highest valuation of the object for sale.
Sluggish bidding is a tip-off to the possible presence of a ring. Sometimes the auctioneer refuses to recognize a known ring leader and simply ignores his bid. At other times the auctioneer holds back good merchandise.
Robert Wraight, [Wraight]an art critic, reveals that sometimes there are inner rings. The outside ring reduces competition in the public auction, and the inner ring pays off unimportant members (impoverished bidders who are knowledgeable and therefore too dangerous to ignore) for a small fee, thus leaving the major advantage to the central few.
An inner ring works like this: A ring of five men buy a painting, they then need to resell it amongst themselves. A, B, and C form an inner ring. A (by prior agreement with B and C) outbids D and E. Later the three inner members (A, B, and C) bid against each other in the third and final auction.
Nearly every regular auction attendee becomes a ring member sooner or later. If you and another antique piggy-bank collector arrive at the same auction and find two good piggies, it doesn't take a rocket scientist to figure out that you each buy one without competing.
Frequently, bidders do not want others to know they are bidding. An example might be a famous art critic who understands that the price of an unknown painting will skyrocket if he is known to be bidding on it. Some bidders hide behind pillars at an auction house, call out or (preferably) signal a bid, and then move to another location.
Sometimes bidders work out hand signals with the auctioneer before the sale. Of course, frequently this is illegal collusion, but at other times the reason can be innocent. There is a story about a man who wanted to buy a certain painting as a surprise for his wife's birthday, but she always went with him to the auctions. He arranged with the auctioneer in advance that he would be bidding as long as his glasses were held in his hand. The ruse was successful, the man won the auction, and his wife learned of her present only on the way home.
Some bidders are quite good at casting doubts on the authenticity or quality of an item. Before an auction they examine the lots and offer opinions as to the poor quality of this or that item. They will then bid and win the goods.
Another trick is to switch rare items from one lot to another. For example, a dishonest bidder could steal a rare letter from Lot 49 and hide it in a group of low-quality letters in Lot 23. He then buys Lot 23 at a low price and nobody notices (at least not until later). An alternate method is to switch the lot numbers marking the groups so that rare items are purchased well below their value.
Some people have been known to steal part of a lot and then offer it for sale to the buyer of the lot. For example, a thief could snip a button from a coat said to have been worn by George Washington and hide it. After the coat is purchased, the thief approaches the winner, says he owns a button that seems to match. The unsuspecting owner often buys the button he already rightfully owns.
Auction houses have been known to rent an old stately mansion and then advertise an auction full of antiques. There may be genuine antiques present, but also the house may have crammed it full of dealer merchandise. This is known as "salting".
There is also the phenomenon known as the "phantom bidder". Sometimes, in order to keep an auction going (or even to start one) an auctioneer will look to a dark corner of the room and nod as though accepting a bid. This technique is quite successful even in a whole roomful of people.
Another auctioneer trick is to pretend that the reserve price is higher than it really is or even to pretend that there is a reserve price when there is none. A variation on the general theme is to raise a price intentionally under the theory that some people will only bid if the price is high enough.
Some sellers arrange to have an object authenticated falsely. This is simple to accomplish because often an expert's own salary is tied to the price of the item he is authenticating. In other words, an expert promising that a painting really is a Monet will receive more money than if he establishes that the same painting is a fake.
Some auctioneers cheat by dishonest use of the "left" bid. A "left" bid is one which has been deposited by a person unable to attend the auction in person and is that person's maximum bid. For example if a book were auctioned and the people present bid a high of $50, the book would be sold to the absent person who offered a maximum of $100. In such a case, an honest auction house would knock down the book for $55--slightly over the highest in-person bid. Crooked dealers would charge the entire hundred dollars or perhaps cheat by announcing at the beginning of the auction that there was an opening bid of $100 in an effort to realize more than they otherwise might.
Which Formats are Most Susceptible to Collusion?
In 1987 Walter Mead [Mead]formulated a hypothesis that the ascending-bid open (English) formats are more susceptible to collusion than are sealed-bid auctions. This belief may explain the popularity of sealed-bid auctions even though ascending-bid formats generate greater revenue. Under the English (ascending, open outcry) scheme, it is easy to form rings.
Sealed-bid auctions, on the other hand, are vulnerable to collusion involving the auctioneer and one or more bidder. This format is less prone to rings because sealed bidding tempts participants to bid higher than the agreed-upon price in order to cheat the others. This is also true for the Dutch format, even though it is an open format since the first defecting bidder ends the auction. Paul Milgrom points out that "collusion is hardest to support when secret price concessions are possible, and easiest to support when all price offers must be made publicly."
The four major formats can be ranked from most prone to collusion to least as follows: (1) English (2) Uniform second-price (3) Discriminatory first-price (4) Dutch.
The English auction is most susceptible to collusion because there is no incentive to betray the ring--more aggressive bidding does not win the item, and also attempts to cheat are visible. For example: Louise values an object at 15, and other buyers in her ring value the object at 12. They all agree to bid no more than 5. If someone cheats and bids 7, Louise is willing to go all the way to 15, and so the cheater gains nothing.
In a first-price sealed auction it is hard to prevent buyers from cheating in a one-shot game (only one trading period). After all, by the time the cheater has been discovered, the auction is over.
By the same reasoning, some economists consider that the Dutch auction is the least susceptible to collusion because ring members will have trouble enforcing collusive behavior. When a ring member bids more aggressively than was agreed, his actions are obvious, but once again the auction is over before anyone can react.
Many auction techniques share a serious problem with trust. An auctioneer is required to be totally trustworthy, and that is not easy to attain or certify. Consider, for example, that an auctioneer has great opportunity and motivation to cheat in a second-price sealed-bid auction. He need only open the bids and insert a bid just slightly under the highest bid offered. His profits would increase.
Matthew K. Franklin and Michael K. Reiter point out that "there are numerous possibilities for corruption and misbehavior in a sealed-bid auction. Possibly the most difficult to counter are those that involve the misbehavior of agents in charge of executing and overseeing the auction (e.g., employees of the auction house), especially when this behavior involves collaboration with certain bidders."
They list several ways to cheat:
The agent opens bids prior to the close of the bidding period and informs a collaborator of the amounts bid.
An agent closes the bidding period prematurely so as to preclude certain bids.
The agent allows a bidder to withdraw (or insert) his bid after the bidding period has ended.
The agent collects money from losing bidders.
The winning bidder refuses to pay claiming insufficient funds.
In the future, as more and more auctions are conducted via computer, the above problems must be addressed.
Notice of Copyrighted MaterialWork presented here is copyrighted material belonging to Agorics, Inc. (Copyright © 1996, Agorics, Inc.) Agorics, Inc. reserves all copyrights. Agorics disclaims any warranty as to the utility, accuracy or effectiveness of the information contained in this document and specifically disclaims any liability for consequential damages that may result directly or indirectly from use of the information in this document.