Costly discrimination

Danish kids are happy to pay to avoid having to work with someone of a different ethnicity. In this clever field experiment, Danish kids with traditionally Danish-sounding names were willing to forego expected earnings in order to avoid being paired with someone with a Muslim-sounding name - and vice-versa. 

It's a great experimental design. Kids do a first round stuffing envelopes on their own, paid a piece rate. For the second round, they have to choose a day to come in, and they'll be partnered with another kid with a joint payoff for how much the team gets done. They're given information on what the other kid achieved in the first round. Because the framing is choice of day to come in for the task rather than choice of partner (though the day determines the partner), there's less chance that the participants would expect the experimenter to be inferring racial preference. 

Team productivity in this task depends on the productivity of each of the workers in the prior round - there's no diversity benefit or penalty. Of course it's just envelope stuffing.

Preference-based discrimination is symmetric. Participants with Danish-sounding names were willing to pay to avoid being partnered with participants with Muslim-sounding names, and vice-versa. 

The insignificant estimate on Danish-sounding indicates that the tendency to discriminate is not different across ethnic types, after controlling for differences in prices. We think that this is a remarkable result for two reasons. First, attention both in the literature and policy debates usually focuses on discrimination of the minority group by the majority group because members of the majority group are more often in the position to discriminate, and workers from the minority group tend to be disadvantaged. However, our results suggest that observing more frequent discrimination of minorities may simply be due to the fact that majority decision makers have more opportunities to discriminate rather than a stronger ethnic animus.

This too was interesting:

Second, this result highlights the importance of controlling for prices when measuring discrimination. From simply looking at discrimination percentages, a layperson may be misled to conclude that decision makers with Danish-sounding names are more likely to discriminate. In fact, decision makers with Danish-sounding names discriminate in 44 percent of the cases, while those with Muslim-sounding names do so in only 33 percent of the cases (however, p = 0.517, χ2 test). Yet, these differences do not reflect differences in animus because decision makers with Danish-sounding names face a lower price on average than decision makers with Muslim-sounding names (€5.2 versus €7.8, p = 0.078, KS). The reason is that workers with Danish-sounding names are systematically more productive (116 letters) in round 1 than participants with Muslim-sounding names (100 letters). According to regressions (2) and (4) in Table 3, these price differences explain the observed differences in taste-based discrimination across ethnic types (Danish-sounding is insignificant, but Price is significant).

They conclude:

Using a sample from Denmark, we find that discrimination is common even at a substantial price, that majority and minority groups are equally likely to discriminate for given prices, and that the demand for discrimination is highly elastic. Our best estimate is that the probability to discriminate falls by about 9 percent if the price of discrimination goes up by 10 percent.

Denmark may be more polarised than other places; would be very interesting to see this repeated elsewhere.

There are a host of interesting implications for policy, which the cautious blogger leaves as exercise for the reader. The elasticity of immigration restrictions with respect to the stringency of workplace diversity mandates or the legal status of repugnant homeowner covenants is one fun one to think about. Second best policy in a world of crooked timber....

Oh: Jean-Robert Tyran is one of the paper's authors. He also wrote my most favourite ever experimental economics paper.

Racial bias in mortgages?

Simon Collins at the Herald asked me for comment on a paper alleging racial bias in mortgage lending; his story's now up.

The paper is available here. It shows, using ordered logit regression, that people who self-identify as being more easily identified as Maori are less likely to own their own home, correcting for income and a few other variables. The paper's empirics say absolutely nothing about mortgages or banks. But the study nevertheless concludes:

"To sum it up in one sentence: results from a large national probability sample of Māori indicate that the more Māori you look, the less 'mortgage worthy' you are."

Here are a few alternative hypotheses:

• The empirics correct for current employment and current income but not past employment and past income. If Māori employment histories are more varied than non-Māori, and if this also follows the "is identified as Māori " indicator, Māori will have less accumulated wealth at any given level of income, and this is not controlled in the study.
• If those who look more Māori are given preference in state housing, then home ownership would also be attenuated.
• If parental resources are negatively correlated with looking more Māori , then that also affects ability to put together a deposit on a house. Note too the potential influence of holding household wealth under Māori land tenure.

I also think they've an error in how they described the magnitude of the effect. Remember that this is an ordered logit regression. So you can't just take the point estimate and multiply it by the number of interval steps to get an accumulated effect; you have to ask your stats package to give you a predicted value at the different values of the category. At page 11, it really looks like they linearised from the point estimate:

Some readers may be wondering how large this effect is in practical terms. One way to think about it is like this: when statistically adjusting for numerous other demographics, such as differences in income, region of residence, and education, a Māori person with a score of 5.55 on our Perceived Appearance measure of Māori identity would be twice as likely to not own their home relative to someone with a score of 1 in Perceived Appearance. This is a statistically significant association, which in our view represents a large and extremely important difference in the rate of home ownership based solely on merely appearing more Māori.

They have an odds ratio of 0.82, which ought to mean that a step change increase in perceived appearance score from the mean score reduces likelihood of owning a home by 18%. I don't think that means that if you go 5.55 steps in the other direction (1/0.18) from the mean score doubles your likelihood of home ownership, except under some pretty strong assumptions. But it's been a little while since I've played around in ordered logit.

Here's the bit where Collins quoted me - entirely fairly:

However Dr Eric Crampton of the NZ Initiative think-tank said there could be many other explanations for this besides racial bias. For example, people who looked more Maori might have parents who did not have freehold properties to use as collateral for loans, a factor that was not surveyed.

"Banks would be throwing money away if they decided to not lend to somebody simply based on looks," he said.

Mortgage brokers Bruce Patten in Auckland and Karen Essex-Mooney in Blenheim both said they had never seen a mortgage application turned down because the borrowers were Maori. They said many borrowers now applied online and never actually met the lenders.

New Zealand Bankers' Association chief executive Kirk Hope said racial stereotyping was not in the banks' or their customers interests especially within such a competitive part of the banking sector.

Neighbourhood externalities

Adam Gurri provides a moral challenge to Paretean welfare economics.

All this only works if you think one preference is just as good as another, and maximizing preferences is a wonderful goal for a moral philosophy. I, for one, think that this is a terrible goal for morality, especially taken in isolation.

My favorite example of this model’s utter failure to provide a sufficient moral compass comes from the play A Raisin in the Sun. In it, an African-American family seeks to move to a predominantly white, affluent suburban neighborhood. You can probably guess where this is going: the people who already lived there did not want this family as their neighbors. From an economics point of view, the family was imposing an externality on their racist neighbors. What’s more, the real historical phenomenon of “white flight” in such cases actually reduced demand for housing in the neighborhoods that were being fled. As a result, the housing prices fell in the neighborhoods that African-Americans moved into, making the existence of a formally-defined externality undeniable.

I challenge the committed Pigovian to explain to me how this is anything other than a clear-cut externality, and how they can avoid the conclusion that their model would have them impose a tax on the African-American family. Moreover, libertarians aren’t in a much better position, morally. In A Raisin in the Sun, we get to see an actual Coasian bargain in action—-the white neighbors pool their resources to offer to buy out the African-American family from the house. Without spoiling the plot, I can’t say that the fact that this is a voluntary bargain inclines me to believe that the preference of the bargainers are on equal moral footing.

The philosophy of maximizing preferences does not distinguish between the family seeking the opportunity to rise above their circumstances and the racists who hate them.

While I'm more than happy to condemn the racist neighbours, I don't think the example provides that strong a case against standard economics.

The baseline is liberal: the African-American family is able to buy the house. The neighbours haven't been able to appeal to some notion of "neighbourhood character" to ban them from moving next door. Despite potentially high transactions costs, the neighbours were able to buy them out, indicating that their actual willingness to pay to be horrible racists was pretty high.

Gurri suggests that the main policy conclusion would be "tax the African-American" family; Coase reminds us that Pigovean externality-taxation solutions are not the only way of enabling Paretean welfare economics, and that the direction of the externality is always a bit up for debate. We could equally note the substantial negative externality the racists place on the incoming family and tax them for it. And in this case the Coasean solution obtained without any policy imposition: the neighbours bought out the incoming family.

I haven't read the play, but based only on the snippet above, I'd have written a Sylvester McMonkey McBean style conclusion: the outgoing family would tell their friends about these crazy racists who are willing to buy you out, at a premium, if you buy a house in their neighbourhood. Eventually, the entrepreneurs would drain the racists of any continued ability to pay for racism.

And while I'll agree with Gurri that any of us coming from particular philosophical perspectives will have sets of preferences that they view as better than others' preferences, having something other than a Paretean set-up for policy requires picking winners among competing values. Even leaving aside the substantial problems Cowen points to in his chapter on non-Paretean welfare standards, we have another simpler, but important, issue. In the play, the neighbours had to demonstrate a real willingness to pay for racism.

In the alternative, in which we move away from Paretean welfare economics, we need some voting apparatus to overturn willingness-to-pay and choice as basis for assessing whether a move has improved welfare. And racism is cheaper at the ballot-box than it is in the real world. It is far easier to imagine neighbourhoods being willing to vote for measures that would ban particular types of people from being able to buy houses, or lodging objections on notified consents, than it is to imagine their actually being willing to buy those people out. If actual willingness to pay is less than the amount necessary to change the outcome, then the externality is not Pareto-relevant.

I'd written on psychic externalities a few years ago:

I'm reminded of Jennifer Roback's work showing how southern racists were able to achieve at the ballot box segregation outcomes they were unable to achieve in the market. To recap: racist southern whites wanted segregated streetcars. But it was too expensive for the streetcar companies to run segregated cars: the increased ticket revenues from white racists didn't compensate sufficiently for lost black custom and, especially, increased running costs. White racists effectively weren't willing to pay enough for tickets to segregated streetcars, so the market didn't provide them. But casting a racist ballot is individually costless. And so streetcar segregation was mandated through regulation.

When I see folks going to the ballot box to enforce their preferences over other peoples' activities, my general presumption is that transactions costs isn't what's keeping meddlers from seeking less coercive options. The ballot box is just cheaper when a majority has weakly meddlesome preferences, regardless of efficiency.

Where illiberal preferences are weakly but broadly held, overturning the results of voluntary choice through use of a non-Paretean framework risks intervening to address externalities that are not Pareto-relevant, with greater illiberalism as result. If, on the other hand, illiberal preferences were very strongly held among a very well-heeled minority, results could flip.