The stakes are especially high when the quality of your purchase is important but also difficult to sign into a contract.
“You can have this problem when a person or an organization tries to buy a good whose quality is unknown to them and they can’t verify it until later in the act of consumption,” he says. Nikola Persikoprofessor of managerial economics and decision sciences at the Kellogg School.
Think of a government that wants to build a new road, a warehouse that wants to install a cooling system, or even a person that needs a decent haircut. Until the buyer can reasonably measure the quality of what he bought, congratulations! It’s already theirs, and there’s no going back.
“Maybe in two years the road will start to crack,” says Persico. “And then it’s too late.”
In situations like these, many common procurement practices—such as holding auctions where the highest bidder gets the job—are not ideal. “If you awarded the contract to the highest bidder, the highest bidder will likely do a poor job and that will cost you a lot more later,” says Persico.
In new research with Giuseppe Lopomo of Duke University and Villa Alessandro of the Federal Reserve Bank of Chicago, Persico shows that the optimal way to find a supplier when you have quality problems is to set a price floor below which bids will not be considered. This prevents higher quality suppliers from being outbid in the bidding process and in turn reduces the chance of the contract going to a shady or unqualified supplier.
The problem with lemons
In economics, the “lemon problem” is widespread and vexing. Whenever sellers know more about the quality of a good or service than buyers do, there is a risk that the market will be flooded with lemons: poor quality products that the buyer would never knowingly buy. After all, why would a seller bother to go the extra mile (and spend the extra money) to create a superior product when they can sell a cheap one for the same price?
Many solutions to this conundrum have been proposed. For example, some buyers require strong seller guarantees. Other buyers rely on a seller’s strong reputation or relationships of trust developed over many transactions to feel confident they are making a quality purchase.
But these options are not always possible. Indeed, says Persico, “even if you have a long-term relationship with a supplier, there may be other constraints that prevent you from acting on that information.” For example, many governments, including the European Union, must adhere to anti-corruption or anti-collusion policies that prevent from making procurement decisions based on any term not specified in a contract.
“You have to award the contract to the highest bidder for transparency,” says Persico. “And you can imagine the disasters that follow from that! You’ll have fly by night starting with a post office box and bidding on contracts.”
So, in the absence of any great way to ensure quality – and with the lowest bids likely to come from the most dubious suppliers – what’s a buyer to do?
Quality-price trade-off
In their paper, Lopomo, Persico and Villa propose that buyers use a lowball lottery auction, which they affectionately call LoLA. At first, LoLA is similar to another popular type of auction, where the buyer sets the price high and then gradually lowers it over time, with bidders leaving the auction until only one final bidder remains.
“So it works the same way, except the price won’t drop to zero,” says Persico. Instead, the buyer will set a minimum price—a floor—below which they will not accept offers. The winner will then be randomly selected from any bidders still in the auction.
This floor price used in the auction is calculated on a case-by-case basis, allowing the buyer to optimally balance their own desire to save money (which would lower the floor price) against their need for a high-quality supplier (which would increase the floor price). .
“There is a trade-off,” says Persico. “The entity bidding on the contract wants to pay less. they want an aggressive bid, a low bid. But at the same time, they worry that if the bid is too low, the highest bidder could be an airline. And so the analysis asks, “How do you optimally resolve this contract?”
Italian roads
So adopting a LoLA seems like a smarter procurement strategy, at least in theory. But how much do shoppers really stand to gain from the switch?
To show the real-world impact of LoLA, the researchers accessed data from Italian government procurement auctions, looking specifically at those contracts awarded to the lowest bidder. They then conducted a back-and-forth experiment, assessing how these municipalities would have fared if they had used a LoLA. “So we say, ‘OK, let’s say instead of giving the contract to the highest bidder, you had used our mechanism,'” says Persico.
They did this by analyzing data about the quality of goods or services delivered by suppliers—in this case, whether there were delays or cost overruns—and then estimating how the same scenario would have played out if governments had used the optimal floor. price for each auction.
Their finding? Setting a price floor would cost the government a little more money up front. But in the long run, once you factor in the costs associated with overruns and delays, the government would have ended up saving money. The savings weren’t huge—a few percentage points—but cumulatively and across many large expenditures, “even modest savings make a difference,” Persico says.
Additional benefits
In addition to being the optimal strategy for buyers concerned with balancing quality and cost, LoLA may also be an optimal strategy if all you care about quality. The difference is that the price floor would be set higher, reflecting less cost consciousness.
LoLA also has another advantage. In the absence of a price floor, attracting more bidders to an auction tends to exacerbate the lemon problem, as the extra competition drives down prices further and discourages quality sellers. But with a price floor, “you can always take advantage of the positive aspects of competition without the negatives,” says Persico. “This floor prevents these very poor bidders from getting the contracts.”
And LoLA isn’t just useful for public sector procurement.
“If you are sourcing ingredients for your private food company, you can imagine that there will be cheap suppliers and expensive suppliers, and of course the expensive suppliers will say that the cheap suppliers will be of lower quality. But if you don’t know that,” Persico says, “you might want to use the auction format that we recommend.”
In fact, Persico and his colleagues are actively looking for companies that fit that bill. In exchange for sharing data, they would be willing to offer free advice.
“Anyone reading this at a private company that has this problem should contact me,” he says.
