Factors that Distort Auctions — The Sell Side

In our first article in this series, we provided an overview of auctions history and explored some of the most common types of auctions. In the subsequent article, we showed how variance in bidder mindsets and behavior can lead to auctions producing less than optimal results. In this article, we turn our attention to sellers, and the ways that they too can alter auction outcomes in ways that make them less efficient or effective.

We’ve seen how auctions can be made less optimal by buyers — whether because of unconscious mindsets or consciously unfair behavior. But naturally, sellers are equally capable of distorting auctions away from their most efficient form, either because of a desire to reduce price uncertainty, to extract more gain for themselves at buyers’ expense, or to control what types of buyers will ultimately win the auction.

One of the more common ways in which sellers can influence auction outcomes is by setting a reserve price — a minimum price that any bid must meet (or exceed) in order to be considered. By setting a reserve price equal to their own valuation for an item, a seller can ensure that they are not forced to sell at a price so low that they would have been better off just keeping it, rather than putting it up for auction.

Of course, the reserve price doesn’t need to equal the seller’s valuation — sellers are free to set reserve prices at whatever level they want. This allows them to potentially capture a portion of the auction surplus that would otherwise be retained by the buyer. To see this, consider a Vickrey auction, where the price the seller receives is equal to the second-highest bid submitted. Suppose there are three participants — Alice, Bob, and Carol — who value the good at $60, $80, and $100, respectively. The seller herself values the good at $70.

As we learned in the previous article in this series, participants in a Vickrey auction will typically bid honestly — that is, their bids will equal their “true” valuations. Without a reserve price in place, the auction outcome is obvious: Carol will win (because she submitted the highest bid), and she will pay $80 (the second-highest bid, submitted by Bob). Because this amount exceeds the seller’s valuation ($70), the seller is better off than if she had kept the item for herself.

How would this result differ with a reserve price in place? That depends on where it’s set. Consider the following three scenarios:

1. The seller decides to set the reserve price equal to her valuation ($70). Because the two highest bids both exceed this amount, the reserve price will not come into play — the auction results are identical to the case with no reserve price.
2. Suppose instead that the seller decides to set the reserve price above her valuation — say, at $90, which falls in between the two highest bids. In this case, while Carol will still win the auction, she must pay the reserve price — $90 — rather than $80, the price submitted by Bob. Thus, the reserve price earns the seller an extra $10 in revenue.
3. Finally, suppose that the seller decides again to set the reserve price above her valuation, but this time even higher — say, $110. Because this amount exceeds even the highest bid, the auction will result in no sale.

These examples reveal two key insights: first, by setting the reserve price above her valuation, a seller can indeed increase her revenue. But this has a potential downside: Since the seller does not know the valuations of any bidders before their bids are cast, she runs the risk of setting a reserve price that exceeds the highest bid, resulting in no sale. Because, as we noted, the sale that would have occurred if there were no reserve price still left the seller better off than keeping the item for herself, the reserve price yields a worse outcome for the seller than no reserve price at all. And of course, buyers end up worse off as well, either paying more for the same item or not being able to purchase the item at all.

Sellers may also demand entry fees, paid by all participants whether they win or lose an auction, as a way to extract revenues from buyers. And in some cases sellers may try to provide special preferences to certain types of bidders — often driven by altruism, as in the case of government set-asides and subsidies for underrepresented minority groups or military veterans.

However, poorly designed set-aside strategies can create unintended consequences. In 1994, the Federal Communications Commission auctioned a band of spectrum for Personal Communications Services. In addition to the basic goals of allocating spectrum efficiently and maximizing revenues, the FCC wanted to ensure that the spectrum sales reflected a diverse set of owners. To that end, they created special rules for bidders designated as “entrepreneurs,” as “regional companies” and as “female/minority owned firms,” giving them 25% bid credit (and thus making their bids worth more relative to other bidders), requiring very low initial deposits and offering generous installment plans for full payment. The outcome of this crude application of targeting: Many purchasers from these groups simply chose to flip their spectrum to large establishment players, while others defaulted on their payment plans when spectrum proved to be worth less than they bid.

All of the above represent wholly legal ways in which sellers can introduce distortionary effects into auctions. But sellers also sometimes manipulate auctions in more illicit fashion. One common means of doing so is “shill bidding,” which is when sellers make bids on their own auctions, by themselves or through agents, to drive up prices. This is a common risk of online auctions, where bidding from multiple accounts is relatively easy and challenging to unmask. In 2001, the FBI charged three men with auction fraud for rigging 1100 eBay fine art auctions by shill-bidding using 40 different fake accounts, driving prices up in some cases by hundreds of thousands of dollars. (Ironically, a form of shill bidding is a regular practice in live-outcry art auctions conducted by major houses like Sotheby’s and Christie’s, whose auctioneers are allowed to “pull bids from the chandeliers” to stimulate competitive bidding by real participants; New York state regulators determined that the practice was legal, so long as the bids are below the publicly stated reserve price for the items being auctioned.)

As with buy-side distortions, different kinds of auctions can reduce or eliminate the possibility of shill bidding. But what if the auction design that best prevents buy-side auction distortions also happens to be the one that’s most at risk for shill bidding?

The Online Ad Market Conundrum

That’s exactly the situation currently faced by online advertising markets. When a digital ad pops up on your phone or computer while you’re browsing, chances are that you’re seeing it because the advertiser won an automated real-time auction for the right to put that ad in front of you.

The auctions that determine how much advertisers pay for that right are conducted by online giants like Google and Facebook, who together serve up the majority of ads on the internet, and by third-party platforms that aggregate and provide ad-presentation services for many other websites.

Google, Facebook and these third-party platforms have all adopted the Vickrey auction — second-price sealed bid — as their auction format of choice. Vickrey auctions maximize profit for ad servers by encouraging competition among advertisers, while incentivizing them to bid what they truly believe the exposure opportunity is worth. Because advertisers don’t pay what they bid, but rather $0.01 more than what the second-highest bidder bid, Vickrey auctions also minimize the “winner’s curse” (buyer’s remorse on the part of auction winners).

However, the fact that the ad-serving platforms are the only ones who know for sure the price of the second-highest bid means that they have the ability to claim that any sum below the winning bid was the “second price” — in effect, making a shill bid at that pricing and inserting it between the winner and the true second-highest bidder. Because all bids are sealed, no one is the wiser.

There’s enough concern about such price manipulation in online advertising auctions — an industry worth $200 billion a year — that many third-party exchanges are now experimenting with shifting to first-price clearance for their ad auctions, which ensures that advertisers know exactly what they’re going to pay for impressions, since they’re the ones setting the price. It’s a risky fix, since first-price sealed-bid auctions can harm buyers, who might overbid and suffer from the winner’s curse as a result, but many in the online ad industry claim that gaming of Vickrey prices by ad sellers has already led to the equivalent of first-price clearance, without the transparency that the latter provides.

This example, and the ones we shared in our prior post about buy-side distortions, illustrates the complicated tradeoffs that exist when trying to optimize auctions. Do you choose to maximize seller revenue, buyer satisfaction or overall utility? Do you emphasize prevention of collusion or speed of process? Ultimately, there are a range of different priorities that auctions providers must assess in designing auctions — while also determining the best means to make them profitable, efficient and scalable.

That’s especially true when the items being auctioned are unfamiliar, valuations are uncertain, supply and demand are wildly volatile and buyers and sellers have a wide range of different risk profiles — all things that are true about today’s NFT market.

We’ll take a deeper look at how auctions platforms are operating in that fast-growing industry in our next article.

Interested in contributing to our Community Economics series? We’d love to hear from you. Comment below or email us at cec@forte.io.

Follow us on Twitter @FortePlatform.