Priceless: The Myth of Fair Value (and How to Take Advantage of It) (34 page)

It’s important for women to learn that they can and should negotiate more, Solnick suggested. They should start with a strong anchor, consider offers critically, and accept that there will be difficult moments in a negotiation. It’s not the woman’s unique responsibility to make everyone feel comfortable at all times.

Contentious divorces are one example of an ultimatum game complicated by gender. The Welch divorce was not atypical of its kind. The spouse with greater earning power played proposer, demanding more than a 50 percent share. The other side’s power rested largely in its ability to veto any proposal (and keep the attorney meter running). What Jane Welch did was conceivably the most effective thing she could have done. By sacrificing perks worth millions a year to
both
the Welches, Jane became Neutron Jane. She demonstrated her willingness to reject unfair offers, overriding any preconceptions that Jack or his legal team might have had. It may have worked. Just before an October 2002 hearing on the temporary alimony, Jack said, “Let’s talk.” They came to an agreement within hours. According to
The Wall Street Journal
, “both sides say the amount is far more than the $35,000 a month Mrs. Welch has been receiving.”

Hotties, male and female, get all the breaks. Economists have been slow to sneak up on what everyone else already knew. In recent years, labor economists have determined that better-looking employees are paid more. This seems to be true regardless of occupation—whether they’re a model on a runway or a coder in a cubicle. There is a “beauty premium” for the congenitally fabulous. For everyone else, there’s a plainness penalty.

Sara Solnick thought the ultimatum game might be a way to investigate the effects of physical appearance on prices and salaries. This is normally a complicated matter, because there are many reasons an employer might pay attractive people more. In sales or waiting tables, appearance is part of the total package. An employer can reason that the public likes an attractive face. The ultimatum game eliminates at least some of these factors. “There are no productivity issues, no expectations, and no contact between subjects,” Solnick and collaborator Maurice Schweitzer wrote. If looks matter even in the ultimatum game, they probably matter whenever people set a price or negotiate a salary.

In Solnick and Schweitzer’s experiment, seventy student volunteers agreed to be photographed. To spare the subjects’ feelings, the photographs were sent to another university where a jury of complete strangers rated each photo on an 11-point category scale ranging from –5 (very unattractive) to +5 (very attractive). Solnick and Schweitzer took the six most and six least attractive people of each gender. The resulting twenty-four photos were compiled into an album.

A group at the second university played the ultimatum game. Each was shown a picture from the album and told that that person was their partner. There was no meaningful difference in how the attractive and unattractive people played. The difference was in how
other
people reacted to them.

Proposers offered attractive people slightly more money than they did unattractive people ($4.72 vs. $4.61). Responders, however, demanded more
from
attractive people. The average minimum demand was $3.53 when the partner was in the most attractive group, and $3.32 when he or she was in the least attractive set. Because most offers are accepted, the latter disadvantage was less important than the former advantage. Overall, attractive people made more money.

This experiment might modestly support the value of looks in occupations that involve soliciting prices from the public (real estate agents, car dealers, auctioneers, salespeople). But Solnick and Schweitzer also looked at gender and found it mattered more than looks did. For the purposes of this experiment, it was more profitable to be a man than to be very attractive. Both genders offered men more and expected less of them. The men thereby earned 15 percent more than the women. Being very attractive was a double-edged sword. Overall the most attractive group earned 10 percent more than the least attractive group. The latter figure exaggerates the effect of looks, though. It contrasts only the best- and worst-looking groups, leaving out the average Joes and Janes.

The belief that women are poor bargainers can be a self-fulfilling prophecy. This idea appears to be rampant among car dealers. Former car salesman Darrell Parrish recalled,

Salesmen . . . categorize people into “typical” buyer categories. During my time as a salesman I termed the most common of these the “typical uninformed buyer” . . . As a rule they were indecisive, wary, impulsive, and, as a result, were easily misled. Now take a guess as to which gender of the species placed at the top of this “typically easy to mislead” category? You guessed it—women.

The argot of car dealers has a revealingly sexual term for a woman who pays sticker price: a “lay-down.”

The best-known experiment on gender, ethnicity, and car prices may still be a controversial 1991–1995 study conducted by Ian Ayres of the Yale Law School and Peter Siegelman of the American Bar Foundation. They sent a small army of 38 volunteers to some 153 randomly selected car dealerships in the Chicago area. The volunteers were all between twenty-eight and thirty-two years old, with three or four years of college education. They were instructed on how to dress: for men, polo or button-down shirts, slacks, and loafers; for women, little makeup, blouses, straight skirts, and flats. All arrived at the dealerships in rented cars, waited for a dealer to approach them, and began to negotiate for a new car.

Like a candidate preparing for a big debate, each volunteer had been coached on a negotiation “script” for two days. Each was to wait five minutes for the dealer to make an initial offer; should none be forthcoming, the volunteer was to prompt the dealer to make one. The volunteer was to respond with a counteroffer equaling the dealer’s marginal cost (options included), computed from
Consumer Reports
Auto Price Service and
Edmund’s New Car Prices
.

They then followed a rote negotiation strategy. One was “split the difference.” Whatever the dealer offered, the volunteer raised his previous offer by half the difference between the dealer’s new offer and the volunteer’s previous offer. This was continued until the dealer accepted or refused to negotiate any further. In the former event, the volunteer said he or she wanted to think about it, and in both cases, the volunteer left without buying.

The experiment found striking evidence of racial bias. On average the final offer quoted to black men was $1,100 more than for white men. This was for the same model, at the same dealership, at nearly the same time. In fact, in about 44 percent of the cases, the white males got an initial offer that was lower than the final negotiated price achieved by female or black counterparts.

The women volunteers required a high tolerance for being called “honey” and “cutie.” “You are a pretty girl, so I’ll give you a great deal,” said one dealer who didn’t. Despite that, the evidence for gender bias was inconclusive. White women paid a bit more than men, but this wasn’t statistically significant. Black women cut slightly better deals than black men.

In designing a field experiment like this, the devil is in the details. There are so many subtle things that can bias the results (even when the experiment is
about
bias). Ayres and Siegelman’s experiment was compelling because of its many well-thought-out safeguards. It was a true double-blind experiment in that neither the dealers nor the volunteers knew what was going on. Ayres and Siegelman had told the volunteers nothing about gender or race, only that they were “studying how sellers negotiate car sales.” Each volunteer was part of a pair (though they didn’t know it). One member of the pair was a white man, and the other was black or
female or both. It was arranged that both members of a pair visited the same dealership within a few days of each other, to bargain for the same model of car.

Ayres and Siegelman’s results occasioned much indignation in the media. Evidence of price discrimination is a complicated thing, though, difficult to reduce to a sound bite. As Ayres pointed out, his results did not necessarily imply what many were assuming—that dealers were prejudiced and wanted to exploit blacks and women. The dealers often steered the volunteers to salespeople of their own race and gender “who then proceeded to give them the worst deals.” Blacks actually got better deals from white dealers, and women got better deals from men.

In 1996 Pinelopi Koujianou Goldberg published another price discrimination study that appeared to overturn any conclusions drawn from Ayres and Siegelman. Instead of doing an experiment, Goldberg used the Consumer Expenditure Survey to check what buyers nationwide had paid for new cars from 1983 to 1987. Goldberg found no statistically significant price differences between blacks and whites, or between men and women. This had liberals pointing to Ayres-Siegelman as proof of the continuing burden of discrimination, conservatives pointing to Goldberg as demonstration that things aren’t so bad, and mainly a lot of people scratching their heads at yet another case of dueling scientific studies.

Goldberg argued that the contradiction can be reconciled. It is necessary, first of all, to understand why car dealers bargain in the first place. Cars are interchangeable, mass-produced products with factory warranties. There’s no reason one new car should sell for a different price than the identical one parked next to it. The vast majority of American buyers insist they hate bargaining. One popular make, Scion, caters to that distaste with no-haggle pricing.

According to a dealer quoted by Ayres, the reason for bargaining is simple. Car selling is a “search for suckers.” Some customers will pay full sticker price, out of ignorance or neurotic aversion to bargaining. There aren’t many customers like this, but they account for a disproportionate share of a dealership’s profits. Ayres reported that some dealerships earn half their profit on 10 percent of their sales.

Earlier this year, I asked a car dealer during an interview whether the bulk of his profits are concentrated in a few sales. He told me that his dealership made a substantial number of “sucker” and “non-sucker” sales. He added, “My cousin, however, owns a dealership in a black neighborhood. He doesn’t sell nearly as many [cars], but he hits an awful lot of home runs. You know, sometimes it seems like the people who can least afford it have to pay the most.”

Goldberg found more variance in the prices paid by blacks and women. There were more outliers paying high prices among these groups than among white men, even though the average sales price was nearly the same for all groups. This could account for the difference between the two studies. In Ayres and Siegelman’s experiment, everyone was required to use the same negotiation strategy. This was intended to reveal whether dealers treated minorities differently, and it did. But the experiment wasn’t designed to test whether black and female buyers bargain differently than white males.

One plausible guess is that many dealers believed in the “sucker theory.” Therefore, they quoted many high initial prices to minorities (in Ayres-Siegelman). Buyers quoted a high initial price tended to bargain longer and harder than buyers quoted a good price. This erased most of the evidence of racial and gender bias (in Goldberg).

If nothing else, this shows how complex price discrimination can be. It’s possible that some dealers weren’t even aware of a sucker theory. Their price quotes may have been statistically biased by race and gender without any conscious intention.

Ayres found that one bit of information was worth $319 to buyers across genders and races. Volunteers who said they had already taken a test drive paid an average of $319 less than those who didn’t, and this was statistically significant. It’s not hard to understand why this makes dealers anxious to close a deal.

A group including Sendhil Mullainathan and Eldar Shafir conducted a particularly ambitious experiment in the fall of 2003. They got permission from a large consumer lender in South Africa to test a grab bag of psychological tricks in its junk-mail pitches for loans. The lender was offering the equivalent of American payday loans—short-term cash for the working poor, at loan shark rates.

The lender sent letters offering a special interest rate to 53,194 past customers. Among other factors, Mullainathan and Shafir’s team tested the effect of having a photograph in the mailing. They found stock photos of pleasant, smiling faces and put them in the lower right corner of the letter, near the signature. This implicitly suggested that the person depicted was a bank employee, maybe the one who had written the letter.

Half the photos were of men, and half were of women. Some recipients got a photo of someone of their own gender and some of the opposite gender. Since race affects everything in South African society, they tested that too. They used photos of blacks, whites, Indians, and mixed-race people.

An economist would say that a photograph should have no bearing on a reasonable person’s decision to take out a costly loan. An advertiser or a con artist would differ, insisting on the value of a pretty face or an ethnic shill. Mullainathan and Shafir were interested in putting an exact value on any effect the photos had. In order to do that, each letter offered a randomly assigned interest rate. In South Africa’s consumer loan business,
it is customary to quote a monthly rate of simple (not compounded) interest. The tested rates ranged from 3.25 to 11.75 percent a month. For these customers, the 3.25 percent was a true bargain, less than half the company’s usual minimum. The 11.75 percent was the customary maximum rate offered to the least creditworthy borrowers.