The misery of diversity?

I just finished reading this 2024 NBER Working Paper by Resul Cesur (University of Connecticut) and Sadullah Yıldırım (Marmara University), provocatively titled "The Misery of Diversity". They look at whether greater genetic diversity is associated with subjective wellbeing (SWB, measured as happiness, or life satisfaction, or affect balance), and find that:

...diversity lowers human SWB, measured by cognitive life evaluations and hedonic assessments of emotional states.

Cesur and Yıldırım demonstrate these results using data on genetic diversity that comes from this 2013 article by Ashraf and Galor (ungated version here). As Cesur and Yıldırım explain:

Population geneticists demonstrate that the dispersal of anatomically modern humans via migratory routes determined within-ethnic genetic heterogeneity. As one moves away from Ethiopia via migratory tracts, genetic diversity, defined as the likelihood of two randomly picked individuals having dissimilar genetic material, decreases...

Our diversity measure impacts the outcomes of interest through social ecology, which, over many generations, likely has influenced cultural evolution. In particular, interpersonal diversity determines the endowment of genetic variation, a measure of social diversity, capturing within-group interpersonal differences across the globe...

This measure of social diversity performs better than conventional diversity indicators, such as the indices of fractionalization and polarization, in capturing the true extent of diversity... In particular, these authors show that while interpersonal population diversity has a substantial and precisely estimated impact on intrastate conflict, fractionalization, and polarization indices fail to explain it.

Underlying data for this index is the expected heterozygosity measures of 53 indigenous human populations genotyped at 780 microsatellite loci as a part of the Human Genome Diversity Project (HGDP–CEPH). It captures the probability that two randomly selected individuals within an ethnic group differ in genetic makeup. In light of the Out of Africa hypothesis, Ashraf and Galor (2013a) constructed predicted genetic diversity for each country by using the coefficient estimate of the impact of migratory distance to Addis Ababa on genetic diversity in the sample of indigenous ethnic groups across the world. Although

Using this measure, with an instrumental variables analysis, Cesur and Yıldırım show that genetic diversity causally decreases subjective wellbeing at both the country level and the individual level (using data from the World Values Survey and the World Happiness Report). Their results are robust to excluding countries that experienced large migrations after 1500 (such as countries in North America and Oceania), and to various other modelling choices. Cesur and Yıldırım dig into the mechanisms for lower subjective wellbeing, and conclude that:

...the misery of diversity is an evolutionary trap caused by the mismatch it creates between the ancestral and current social environments via reduced social cohesion, retarded state capacity, elevated mistrust, and increased inequality of economic opportunities.

So, it seems like this is good evidence that genetic diversity decreases subjective wellbeing. However, there are a couple of problems. First, when most people think about diversity, they are thinking about between-group diversity, not within-group diversity. Between-group diversity is what you get when people from different ethnic groups are together. Within-group diversity is what you get when people from the same ethnic group differ genetically from each other. Cesur and Yıldırım's measure is heavily weighted towards within-group diversity. And indeed, in one of their analyses they find that it is within-group diversity that matters the most in their analysis. When they split their measure into within-group and between-group diversity, within-group diversity has a statistically significant (and negative) effect on subjective wellbeing measures, while between-group diversity is statistically significant.

So, Cesur and Yıldırım's analysis might be correct, but at the same time kind of misses the point. Between-group diversity is something that has potential policy levers (migration policy), whereas within-group genetic diversity is not something that is amenable to policy change. At least, not without eugenics (and, to be clear, I am not advocating for that). 

The second problem comes from the analysis of first-generation and second-generation immigrants in Europe and the US, where Cesur and Yıldırım find that:

...while home country diversity continues to hurt the SWB of first-generation immigrants, such effects weaken among the second-generation, suggesting that long-run improvements in the social environment can mitigate the misery of diversity over generations.

These results are not well-explained. If a person is born in one country, and then moves to a new country, shouldn't it matter how long they are exposed to the genetic diversity in the country of birth, and how long they are exposed to the genetic diversity in the destination country, in terms of the impact on subjective wellbeing? Cesur and Yıldırım don't show any dose-response relationship here. And there should be no effects at all on the second generation (which is what they find), because for the second-generation immigrants, the genetic diversity they have been exposed to is the country of their own birth, not the country of birth of their parents. However, that is only a small problem in an otherwise interesting paper.

Overall, I think Cesur and Yıldırım need to engage a bit more with why anyone should care about genetic diversity, given that it is not amenable to policy change. Until they can do that, this paper can be filed under the interesting, but unhelpful category.

[HT: Marginal Revolution, last year]

Population diversity and economic growth

Population diversity has a theoretically ambiguous effect on economic growth. On the one hand, having a more diverse population makes it more difficult for people to agree on things like spending on public goods (e.g. see this post), it can open the door to policies that favour certain ethnic groups, and lead to conflict over resources and the management of public services. On the other hand, having a more diverse population brings people together with different (and complementary) skills, experiences, and ways of thinking, which can boost innovation and productivity, as well as fostering connections with different communities (and other countries), which may increase international trade and investment.

Many studies have tested the relationship between population diversity and economic growth, with little consensus. That makes the literature ripe for meta-analysis, where the results of many studies are combined in order to estimate an overall relationship. That is the approach in this new article by Andreas Sintos (University of Luxembourg), published in the Journal of Economic Surveys (open access). Sintos collates the results from 83 studies, with 1537 estimates of the relationship between some measure of population diversity and some measure of economic growth.

First, Sintos establishes that there is a small publication bias overall, with studies that find a negative relationship between diversity and growth being more likely to be published than would be expected given the overall distribution of results. Then, after adjusting for publication bias and methodological quality of the studies, he finds that:

...while ethnic and linguistic diversity demonstrates a small and statistically insignificant positive effect on economic growth, the remaining dimensions of diversity—religious, genetic, birthplace, and the residual category—demonstrate a significant positive impact on economic growth, with effect sizes spanning from moderate to large.

So, population diversity (specifically religious, genetic, and birthplace diversity, as well as a residual category that captures other forms of diversity) is positively associated with economic growth. Places that have more diversity of those types (but not places that have more ethnic or linguistic diversity) grow faster. A 'moderate to large' effect here means that each standard deviation higher diversity is associated with 0.1 to 0.4 standard deviations higher economic growth. That is not to be sneezed at.

What Sintos isn't able to do, though, is explore the mechanisms that underlie that positive relationship. So, while meta-analysis can give us an overall estimate of the relationship, it can't tell us why that relationship exists. To do that, we would need to go and look at the individual studies, particularly those that found a positive relationship between diversity and growth, and see if they explored the mechanisms.

Finally, this article made me chuckle, as it is clear that substantial portions of it were written by generative AI. No human uses the word "elucidate" 14 times in a research paper, and quantitative papers rarely refer to the "scholarly discourse". I should really have been alerted to this when the first paragraph included the LLM-ese sentence: "The significance of population diversity within the economic sphere is multifaceted". Perhaps diversity's significance is multifaceted. This article doesn't tell us that though. All it tells us is that the relationship between diversity (by some measures) and economic growth is positive. More diverse places tend to grow faster.

Ethnic discrimination among young children

In yesterday's post, I discussed the idea of taste-based discrimination. Such discrimination is effectively a prejudice against members of particular population groups (such as a particular ethnic group or gender). On that theme, I recently read two research articles that looked at ethnic discrimination among children.

The first article was this one by Jane Friesen (Simon Fraser University) and co-authors, published in the Journal of Economic Psychology in 2012 (ungated earlier version here). Their sample was 430 Grade 1 and 2 children (aged 5-8 years) from Vancouver. Friesen et al. ran an experiment that involved two tasks. The second task (a 'sharing task') was the most relevant to this post. The sharing task was a variant of the dictator game, which is a common experimental game used to examine norms of fairness within a population. In this case, each child got to share 12 stickers among four people (themselves, a White child from their class, an East Asian child from their class, and a South Asian child from their class). The task was repeated three times for each child. 

Looking at the results of the sharing task, Friesen et al. find that:

Overall, participants share on average 13.6/36 stickers or 38% of their endowment. South Asian children share fewer stickers overall (11.7) compared to Whites (13.8) and East Asians (13.7)...

...50% of participants chose a non-discriminatory allocation (including 5.2% of participants who shared zero stickers)... White participants were substantially more likely (55%) than East Asian (45%) and South Asian (40%) participants to choose a non-discriminatory allocation, and girls were more likely (54%) than boys (42%) to do so...

...on average, White participants shared slightly more stickers with the recipient from their own ethnic category than with the other two; East Asians participants shared slightly fewer with their own category than with the other two; and South Asian participant shared slightly fewer with the East Asian recipient than with the other two.

That provides some evidence in favour of discrimination. Going a bit further, Friesen et al. then run a regression model on their data, and find that:

...White subjects share slightly less than one-third of a sticker more on average with the White target than with either of the other two recipients... However, East Asian participants show no in-group bias; if anything, they share fewer stickers with their own group with other groups... This difference in patterns of in-group bias between Whites and East Asians is statistically significant... the relevant point estimates indicate that South Asians may show somewhat less in-group bias than Whites in terms of the number of stickers shared... but this difference is not statistically significant.

In other words, there is evidence for ethnic discrimination among White and South Asian children, but not among East Asian children. There were no differences in the degree of discrimination between girls and boys.

The second article was this one by Annika List, John List (both University of Chicago), and Anya Samek (University of Southern California), published in 2017 in the journal Economics Letters (ungated earlier version here). List et al. report on a field experiment among children aged three to five years, which also involved an application of the dictator game. In this case:

...children were matched to teddy bears or other students and decided how many of their marshmallows to send them. We unobtrusively indicated the race of the match by showing pictures of hands (lighter or darker skin color) or pictures of teddy bear paws (light or dark brown).

The choice to use teddy bears as well as human hands is novel, and explained as:

...we use the teddy bears as a control in order to rule out that preferences are driven solely by dislike for darker or lighter colors. By comparing the aversion to giving to a darker color hand person relative to darker paw teddy bear, we disentangle the role of racial discrimination from preferences for colors in children’s choices.

Based on their sample of 117 children, each completing four rounds of the dictator game, List et al. find that:

On average, white children send 0.97 marshmallows to white recipients and 1.47 marshmallows to black recipients. Similarly, Hispanic children send 1.18 marshmallows to white recipients and 1.55 marshmallows to black recipients. Alternatively, black children send more marshmallows to white recipients (1.33) relative to other blacks (0.99)...

Contrary to our expectation, we do not see a difference when comparing giving to teddy bears versus to human children. 

In other words, this study provides little evidence of ethnic discrimination among the youngest children. 

These two studies piqued my curiosity because it seems somewhat obvious that ethnic discrimination develops over the course of a person's childhood. Taken together, these two studies provide some support for that view. I was a little surprised that there wasn't already a well-established literature in this space (see this 2008 article in the journal Nature for more). However, these two studies tell us little about what contributes to discrimination among children, or how it can be prevented. That gives a lot of scope for future research in this space.

Your difficult-to-pronounce name could hold you back in the labour market

Both of my children have common first names. Their mother was in favour of weird names, but my feeling was that giving your child an unusual name consigned them to a lifetime of having to spell their name for people, or having to deal with mispronunciations of their name. My preference for common names was based purely on reducing the direct costs to my children, and not on how it might affect their labour market outcomes.

However, I have probably underestimated the negative impacts of having difficult-to-pronounce names. This recent working paper by Qi Ge (Vassar College) and Stephen Wu (Hamilton College) shows some significant negative effects, in the form of labour market discrimination. And those effects are independent of ethnicity. Ge and Wu note that:

Although many ethnic sounding names are also difficult to pronounce, particularly for those outside of that particular racial or ethnic group, there is still variation in the fluency of names within particular groups. For example, most non-Chinese speakers would consider Chen to be more familiar and easier to pronounce than Xiang; people without a Polish background will generally have much more trouble trying to pronounce the surname Przybylko than they will with Nowak.

Even after controlling for the ethnic or racial origin of one’s name, there are a few reasons that individuals with hard-to-pronounce names may experience worse outcomes in the labor market. There may be subconscious bias against those with difficult-to-pronounce names, leading potential employers to have more negative evaluations for these applicants and be more critical of their profiles. Recruiters will also have an easier time processing and remembering names that are more fluent and/or familiar sounding. Some individuals on hiring committees may undertake the mental effort to remember difficult sounding names, but others may not.

Ge and Wu conduct three analyses to demonstrate the labour market effects of name fluency (how easy a name is to pronounce). First:

...we utilize observational data from the academic labor market by assembling curriculum vitae (CV’s) of over 1, 500 economics job market candidates from 96 top ranked economics PhD programs from the 2016-2017 and 2017-2018 job market cycles and find that name fluency is significantly related to job market outcomes. Specifically, candidates who have difficult-to-pronounce names are much less likely to be initially placed into an academic job or obtain a tenure track position, and they are placed in jobs at institutions with lower research productivity, as ranked by the Research Papers in Economics (RePEc) database. Our results are consistent and robust across three separate ways of measuring pronunciation difficulty: an algorithmic ranking based on commonality of letter and phoneme combinations, a survey-based measure that records the average time it takes individuals to pronounce a name, and a purely subjective measure based on individual ratings.

Ge and Wu then re-analyse data from two seminal studies on labour market discrimination: (1) this 2004 study (ungated version here) by Marianne Bertrand and Sendhil Mullainathan, where the researchers sent out CVs that had either an African American name or a non-African-American name; and (2) this 2011 Canadian study (ungated version here) by Philip Oreopoulos, where CVs were sent out with Indian, Pakistani, or Chinese names, and compared with those sent out with English names. In these re-analyses, Ge and Wu find that:

In analysis of data from Bertrand and Mullainathan (2004), we find that job applicants with less fluent names have lower callback rates, even after accounting for the implied race of the candidate. What is particularly striking is the fact that within the sample of resumes with distinctly African-American names, name fluency is still strongly correlated with callback rates. We also document similar results using data from another prior audit study by Oreopoulos (2011). Once again, job applicants are less likely to be called back when they have names that are difficult to pronounce, and even when restricting the sample to immigrants with ethnically Indian, Pakistani, and Chinese names, those whose names are less fluent are significantly less likely to be called for a job interview.

Ge and Wu used three different measures of name fluency in their three analyses:

...a computer-generated algorithm that assesses the difficulty of pronouncing various words, a rating based on the median time it takes for people to pronounce a particular name, and a subjective measure based on three independent raters.

They receive similar results regardless of how they measure name fluency. Ge and Wu also explore which mechanisms may be behind the apparent discrimination, concluding that:

For job searches at academic, governmental, and research institutions, an initial screening generally involves committees getting together to discuss names of potential candidates, which may lead to some subconscious discrimination against names that are harder to pronounce and/or remember. This may also occur in the settings of prior audit studies, where recruiters must decide which applicants to call for potential interviews. Another possibility is that there are mental costs to processing and remembering less fluent names, and that these mental costs are only worth spending on higher quality candidates or when the jobs carry significant stakes. Additional analysis from both observational and experimental data seems to support this explanation. In particular, we find that PhD job market candidates with relatively weak profiles or from non-top PhD programs are more likely to suffer from name penalty in their search for academic positions. Likewise, we document similar patterns from separate analyses of Black... and Indian, Pakistani, and Chinese... applicants: those who are less qualified and have weaker resumes tend to encounter much greater discrimination due to name pronunciation than those who are more qualified and have stronger resumes.

All in all, there is some strong evidence that having difficult-to-pronounce names is negative for labour market outcomes, on top of any ethnicity-based labour market discrimination. While the study context is the US academic job market (at least for the primary analysis), it seems likely that this effect would appear more broadly across the labour market (and the secondary analyses of the two prior studies support this). It is little wonder, then, that some people choose to change their names in employment applications, as noted in a new report on Pacific workers in New Zealand prepared by the Human Rights Commission (as reported in the New Zealand Herald this week).

The solution is somehow to reduce the mental costs associated with difficult-to-pronounce names. Could that be as simple as greater exposure of people making hiring decisions to a range of different ethnic groups, or specific training in cultural intelligence, or something else? We will need some further (experimental) studies on this to find out.

[HT: Marginal Revolution]

Read more:

• What's in a (first) name?
• Uncovering labour market discrimination against foreign-born and native-born minority workers

Ethnic diversity of local government and decision-making gridlock

Is it better to have more (ethnically) diverse local government, or less diverse local government? There are valid theoretical arguments in both directions. If local government leaders (e.g. elected council members) are more diverse, then they will have a diversity of opinions and preferences, possibly leading to more disagreements and less consensus decision-making, and government may become 'gridlocked', unable to make key decisions. On the other hand, local government leaders do feel electoral pressure, including the pressure to conform, and to the extent that there is effective electoral pressure, gridlock would not be a problem.

Whether more diverse local governments spend less on public goods (or not) is the subject of this 2017 article by Brian Beach (College of William & Mary) and Daniel Jones (University of South Carolina), published in the American Economic Journal: Economic Policy (ungated version here). They use data from the:

...California Election Data Archive (CEDA), which provides the names and number of votes for every candidate in every local government election occurring between 1995 and 2011.

For the 5177 candidates who won elections (or were close but lost) over the period from 2005 to 2011, they collect data on ethnicity, either directly from city councils, or by asking workers on Amazon Mechanical Turk (mTurk) to classify the candidates' ethnicities. They got 10 mTurk workers to classify each candidate, and had 94 percent agreement overall (they drop the 31 candidates who had low agreement from their sample). This was similar to the level of agreement between mTurk workers and city council data (95 percent).

Beach and Jones then measure the ethnic diversity of each city council, using indices of fractionalisation and polarisation. As they explain:

Both indices range from zero to one, where zero corresponds to a situation where there is no diversity. Fractionalization is maximized when each council member is of a different ethnicity. Polarization, on the other hand, is maximized when the seats are distributed into two ethnic groups.

The outcome variable that Beach and Jones are most interested in is public goods expenditure, which they calculate:

...by taking a city’s total expenditures for the year and removing expenditures on “government administration” and debt repayment. The “public goods” category therefore includes all spending on roads, parks, police protection, sewerage, public transportation, etc.

Now, a simple regression approach would be to look at the relationship between diversity (fractionalisation and polarisation) and public goods spending. The problem with that approach is the potential for endogeneity - maybe there are city-level factors that affect both the diversity of local government and local public goods spending. For example, perhaps having a more diverse population requires a greater variety of public goods, and more public goods spending, but also tends to lead to a more diverse city council. In that case, the relationship between diversity of the council and public goods spending is biased because of the relationship of both variables to the diversity of the population overall. Beach and Jones deal with that problem by looking at what happens subsequent to close elections, where one of the candidates is the majority ethnicity, and one is a minority. In sufficiently close elections, it is close to random which candidate is ultimately elected, provided some random variation in the diversity of the council, that doesn't depend on any other variable.

Beach and Jones identify 684 close elections with candidates of different ethnicities over the period from 2006 to 2009. Using that data, they find that:

Regardless of whether we measure diversity with fractionalization... or polarization... there is a strong and positive relationship between the election of a non-modal candidate and the diversity of the city council.

No surprises there. Electing a minority candidate increases the diversity of the council. Moving onto the effect on public goods, they find that:

...per capita spending on public goods falls by approximately 13 percent (significant at the 1 percent level) following the election of a non-modal candidate. The effect on nonpublic goods spending remains positive (roughly 14 percent) but is not significant at conventional levels.

So, more diverse local governments spend less on public goods. Beach and Jones then drill down into potential mechanisms that explain their results, and the consequences, and show that:

Our results indicate that diversity leads to gridlock. Cities reduce the amount they spend on public goods as their city council becomes increasingly diverse. These effects are largest for segregated cities and cities with more income inequality (where the potential for disagreement may be largest). We also find that all members of a council that experienced an exogenous shock to diversity receive fewer votes when they run for reelection. This latter point suggests that the city’s population is dissatisfied with the decline in public goods, ruling out the possibility that diverse councils simply achieve greater efficiency in public good provision.

So, ethnic diversity of local government appears to encourage gridlock, reducing local public goods spending, and it isn't an outcome that citizens favour. However, one thing that Beach and Jones didn't consider, is whether (or to what extent) a match between the majority ethnicity of local government and the majority ethnicity of the population matters. Or whether the effect is different at different levels of ethnic representativeness. Those would be interesting follow-up questions.

Also, the negative implications of this research need to be juxtaposed with the problem of groupthink. Groupthink occurs when there is too much consensus, leading to decision-making that lacks critical evaluation. This is more likely when the group of decision-makers is less diverse. So, perhaps the quantity of public goods spending is higher when there is less diversity in local government, but perhaps the quality of that spending is lower?

Genetic diversity and its enduring effect on economic development in the US

How important (or otherwise) is ethnic diversity for economic development? This is a genuinely difficult question to answer. However, this 2017 article by Philipp Ager (University of Southern Denmark) and Markus Brueckner (Australian National University), published in the journal Economic Inquiry (ungated earlier version here), makes a good attempt in one context. Ager and Brueckner use US county-level data over the period from 1870 to 2020, looking at how genetic diversity in 1870 (measured using the number of European-born people from different countries) was related to the change in output per capita (which is the sum of agricultural output and manufacturing output, and is a proxy for GDP). In their simplest regression models, they find that:

The point estimates on genetic diversity are positive and significantly different from zero at the 1% level in all regressions; quantitatively, they range between 0.09 and 0.12... the coefficient of 0.09... suggests that, on average, a one unit increase in genetic diversity increased U.S. counties’ output by around 10%. An alternative interpretation is that a one standard deviation increase in immigrants’ genetic diversity increased output per capita growth during the 1870–1920 period by around 20% (equivalent to around 0.25 standard deviations; or alternatively, 0.4% per annum).per capita during the 1870–1920 period.

That is quite a sizeable positive effect. Ager and Brueckner then go on to show that there are enduring effects, even more than a century later. The impact on 2010 income per capita is:

...around 0.04 and significantly different from zero at the 1% level.

There is also a statistically significant and positive (and somewhat larger) relationship with income per capita in 2000, 1990, 1980, and 1970. Ager and Brueckner then go even further, looking at the relationship between genetic diversity in 1790 and income per capita in 2000, finding that:

The estimated coefficient on 1790 genetic diversity is around 3.1... and statistically significant at the 1% level. Quantitatively, the estimated coefficient on immigrants’ 1790 genetic diversity suggests that a one standard deviation higher genetic diversity in 1790 was associated with a higher value of income per capita in the year 2000 of around 0.3 standard deviations.

Their results are also robust to a variety of alternative measures of development, non-linearity, controlling for pre-trends, and excluding outlier observations. However, they fall a bit short of demonstrating a causal impact, even when accounting for pre-trends, since counties that are initially more diverse may differ in meaningful ways from counties that are less diverse, not least in the types and variety of immigrants they attract. Nevertheless, the results are interesting, and especially when Ager and Brueckner consider a potential mechanism, finding that, at the state level:

...the genetic diversity variable has a significant positive effect on patents. On the other hand, there is no significant effect of genetic diversity on conflict for the sample at hand...

They don't dwell on these results too much, but I think they demonstrate that diversity is not associated with greater conflict (which would reduce productivity), but is associated with more innovation (which would increase productivity). Greater diversity leading to more innovation and therefore greater development does seem like a plausible (and good news) story. However, it would be interesting to see some more research corroborating these results in other (and perhaps more contemporary) contexts.

Price and prejudice

Economists distinguish between two different types of discrimination:

1. Taste-based discrimination, which involves bias against members of a particular group (this discrimination arises because of people's preferences for or against particular groups); and
2. Statistical discrimination, which involves treating people differently based on the group they belong to, because of differences in average characteristics between groups (this discrimination arises because of imperfect information about people, leading them to be treated as if all members of a particular identifiable group are the same).

As I noted in my recent review of Thomas Sowell's book Applied Economics, Sowell carefully explained that discrimination often imposes a cost on the person doing the discriminating. For example, if an employer discriminates against employees of a particular type, they may choose to employ others with lower productivity (and lower profitability for the employer) instead. The cost comes in the form of lower profits.

How much of a cost are discriminators willing to bear? That is the research question that is addressed in this 2018 article by Morten Hedegaard (University of Copenhagen) and Jean-Robert Tyran (University of Vienna), published in the American Economic Journal: Applied Economics (appears to be open access, but just in case there is an ungated version here). Hedegaard and Tyran use a field experiment among Danish high school students to estimate the willingness to pay to work with someone of the same ethnicity. Specifically, in the field experiment:

We hire 162 juveniles from secondary schools in Copenhagen, Denmark, with Danish-sounding and Muslim-sounding names to prepare letters for a large mailing and pay a piece rate. Workers are requested to show up for work twice in two consecutive weeks. In the first round, they work by themselves and we measure their individual productivity on the job. Before they come back for the second round, we call randomly selected workers on the phone and inform them that they will again do the same job but now have to work in teams of two. They are informed that they are paid the same piece rate as in round 1 and share earnings from team output in round 2 with the coworker. These randomly selected workers can choose whom to work with. The choice is between a candidate from the ethnic majority group and a candidate from an ethnic minority group. In treatment Info, we provide the decision maker with information about the individual productivity of the two candidates, i.e., the number of letters they prepared in round 1, and their first names as a marker of ethnicity... Rational decision makers who choose the less productive worker of the same ethnic type thus discriminate knowingly and deliberately.

Importantly, because Hedegaard and Tyran know the productivity of the workers from the first round. The sample is essentially split into threes. One person in each group of three is a decision-maker, and chooses which of the other two workers that they will work with in the second round. Hedegaard and Tyran ensure that the choice is between someone of the same ethnicity as the decision-maker, who has lower productivity, and someone of the opposite ethnicity, who has higher productivity. The 'price' of discrimination varies between decision-makers, depending on how more productive the opposite-ethnicity worker is than the same-ethnic worker that each decision-maker is offered. Hedegaard and Tyran can then test how much discrimination varies between decision-makers with higher and lower prices of discrimination. Based on their sample of 140 workers who completed both rounds, they find that:

...discrimination is common even at a substantial cost and that the tendency to discriminate is not different across ethnic types. We estimate that discriminators on average are willing to forego 8 percent of their earnings in round 2 to avoid a coworker of the other ethnic type. Our main result from treatment Info is that discrimination is highly responsive to the price of prejudice. Our best estimate is an elasticity of −0.9, i.e., we find that the probability to discriminate falls by about 9 percent if the price of discrimination goes up by 10 percent.

There is a lot to unpack there. First, people are willing to discriminate even if it costs them (which is consistent with Sowell's argument). Second, and a result that would surprise many people, ethnic majority and ethnic minority workers are equally likely to discriminate. Third, the elasticity is quite high - increasing the cost of discrimination reduces discrimination significantly.

Could it be that Hedegaard and Tyran are picking up statistical discrimination? That is, are the workers basing their decision on the average expected productivity of the workers of different ethnic groups? This seems unlikely, since both ethnicities are engaging in discrimination, and by definition both groups can't be less productive than each other (in fact, the Danish group is statistically significantly more productive). However, Hedegaard and Tyran explicitly test how important taste-based discrimination using a different treatment group, where the decision-makers were not provided with information about the round 1 productivity of the workers they could choose between (Hedegaard and Tyran refer to this as the 'NoInfo' treatment). They find that:

...statistical discrimination does not explain observed outcomes in NoInfo well. We find a large gap between observed earnings and earnings predicted by statistical discrimination (about 4 percent of total output). To account for taste-based discrimination, we use our estimate from treatment Info and find that it predicts well out of sample; about 40 percent of that gap is explained by animus-driven prejudice. Thus, our results suggest that prejudice is an important cause of ethnic discrimination in the workplace, and that it needs to be taken into account above and beyond the theory of statistical discrimination.

So, clearly there is a substantial amount of taste-based discrimination in this sample. To see just how much, consider that in the NoInfo experiment, 78 percent of decision-makers chose the same-ethnic worker, compared with just 38 percent in the Info experiment. Simply providing information about the price of discrimination was enough to reduce discrimination substantially.

Other than an interesting test of the relative important of the two types of discrimination, does this research provide some policy implications? It clearly demonstrates the existence of substantial ethnic bias or prejudice between workers. However, in terms of addressing the problem of discrimination this research suggests that, if there is some way to explicitly estimate the costs of discrimination, and make those explicit to decision-makers, discrimination could be reduced. Unfortunately, it is not clear how that would work as a solution in other real-world contexts.

The effect of migrant children on native-born childrens' academic performance

Back in July, I wrote a post about how the exposure to foreign-born students affected the academic performance of students in Florida. Unsurprisingly, there is a lot of related research on the impact of immigrant students on native-born students. The effect will depend on the characteristics of the immigrant students (e.g. whether they speak the language of instruction; or the education level of their parents) and how teachers respond (e.g. do they change the way that they present material) and how schools respond (e.g. are immigrant students clustered into particular classes).

Identifying the effect of school peers (including immigrant students) on academic performance is quite difficult, mainly because there is a problem of selection bias. Parents often get to choose what school to send their children to. Schools usually get to choose how to allocate students across classrooms, sometimes on the basis of academic merit (often referred to as 'streaming'). These selection effects mean that the observed relationship between academic performance and the number (or proportion) of immigrant peers is going to be biased. Researchers must find some way to deal with the selection bias.

In this 2020 article (open access) by Kelvin Seah (National University of Singapore), published in the journal Australian Economic Review, selection bias is reduced by comparing student performance between two consecutive cohorts of students at the same school. That deals with any selection bias in relation to schools (because the comparison is within schools). Using data from entire school cohorts might deal with class selection issues (although I am not convinced - it simply means that the bias might be positive for some students, and negative for others, but there is no guarantee that it averages out to zero).

Seah uses data from the 1995 TIMSS study, for Australia, Canada, and the United States. This study collected data on maths and science performance for students in seventh and eighth grades. The data set includes over 40,000 students from over 700 schools across the three countries. Measuring exposure to immigrant students as the proportion of non-native-born students in each school cohort, Seah finds that:

There are marked differences in the share of immigrant students to which native students are exposed in the three countries. On average, natives in the Australian sample have the highest share of immigrant peers in their grade level in school while natives in the Canadian sample have the lowest.

Looking at the effects on maths achievement in the TIMSS test, Seah finds that:

For Australia... the maths achievement of native students is positively associated with the share of immigrant peers in the grade... a 10‐percentage point rise in the grade share of immigrants increases native maths achievement by about 0.093 standard deviations (significant at the 5 per cent level)...

For Canada... the share of immigrant students in the grade is negatively associated with natives’ maths achievement (this relationship is statistically significant at the 5 per cent level)... A 10‐percentage point increase in the share of immigrant grade peers is estimated to reduce native maths achievement by 0.048 standard deviations...

For the United States... Once non-random sorting of immigrant and native students across schools is taken into account, the relationship between these variables disappears and the coefficient falls to essentially zero. The result is unaltered when controls for individual, family, and school‐grade characteristics are added...

The results for Australia might seem a little surprising at first, but Seah shows that immigrant children in Australia perform better in maths than native-born children. The results are similar (but the size of the effects are much smaller) for performance in science. Digging a bit deeper into the maths results, Seah finds that:

...the peer effects of immigrant students are more adverse when immigrant students are non‐native speakers of the test language and when they have less‐educated parents. Further, the estimated peer effects of immigrant students attenuate when subject achievement of immigrant students is controlled for, suggesting that peer effects are at least partially working through immigrant students’ achievement.

There's nothing too surprising there. Seah then tries to tie the disparate results for Canada and Australia to the degree of autonomy that teachers have in setting the curriculum, and finds that:

...the results for maths achievement indicate that the peer effects of immigrants are more positive in schools where teachers have a high degree of influence in determining curriculum.

That suggests that, if we want to limit any negative effect of immigrant students on native-born students' achievement, we should allow teachers to better tailor their course offerings for their class. Unfortunately, that is almost the opposite what we might conclude from this 2018 article by Hu Feng (University of Science and Technology Beijing), published in the Journal of Comparative Economics (sorry, I don't see an ungated version online). Feng uses data from the China Education Panel Survey (CEPS), and instead of international migrants, the focus here is on internal migrants. So, at the least, there is unlikely to be much of a language effect in Feng's sample. Interestingly, Chinese middle school students (who are the focus of Feng's study) are mostly allocated to classes either randomly, or using a "balanced assignment rule" (which means that the best and worst students go in one class, the second-best and second-worst in another class, and so on). This random allocation allows Feng to deal with the class selection issue (but I'm not entirely convinced about the absence of school selection, because wealthy parents could opt their children out of public schools).

Feng looks at the effect of migrant peers on performance in maths, Chinese, and English, and finds that:

...the presence of migrant peers in the classroom has negative and statistically significant effects on math scores of local students... a ten-percentage-point increase in the proportion of migrant students in the classroom reduces local students' math test scores by 1.06 points, which is equivalent to 0.11 standard deviations...

...migrant peers have large and negative effects on local students’ Chinese test scores... a ten-percentage-point increase in the proportion of migrant students in the classroom reduces local students’ Chinese test scores by 1.06 points, which is equivalent to 0.11 standard deviations... Finally... migrant peers have negative but relatively small effects on local students’ English test scores.

Feng also finds that the results are larger for male than for female students. They then go on to look at how teachers respond to migrant students. They find that:

...in the classes with higher proportions of migrant students teachers are less likely to use the methods of group discussion and interaction with students, which are usually assumed to better improve students’ cognitive abilities... On the other hand... the presence of migrant students in the classroom has negative effects on the use of relatively advanced teaching media like multi-media projector, Internet, and pictures, models, or posters, which are important for teaching effectiveness.

So, it appears that when teachers have the ability to change the mode of instruction, they do so in ways that make local students worse off, when there are more migrant children in the classroom. This is probably exacerbated by teachers' attitudes, because Feng reports that teachers "prefer to teach classes with fewer migrant students". Unfortunately, Feng's study is silent on whether teachers modify the curriculum in response to the presence of more migrant students.

It is worth noting that the negative effects in these two studies are contrary to the findings of Figlio et al., which I referred to in my earlier post. They found no effect of immigrant exposure on native-born students. This literature is crying out for a meta-analysis at some point. We also need some further studies to unpack the mechanisms, since it is important to better understand whether changing curriculum or teaching modes, or both, to suit immigrant students has a net negative or positive effect overall.

Read more:

• Exposure to foreign-born students and the academic performance of U.S.-born students

Uncovering labour market discrimination against foreign-born and native-born minority workers

There is an interesting strand of research in economics (especially labour economics) that seeks to uncover evidence of discrimination using what are referred to as correspondence tests. Essentially, the researchers send out pairs of job applications to a bunch of firms, where each pair of applications differs only in some characteristic that the researchers want to test for discrimination against. So, for example, in the original research using this method (ungated version here) by Marianne Bertrand and Sendhil Mullainathan, the CVs they sent out had either an African American name or a non-African-American name. They then compared the number of each type that were invited to interviews.

The literature that applies this approach is large and growing. However, this short 2017 article by Nick Drydakis (Anglia Ruskin University), published in the journal Economics Letters (ungated earlier version here) caught my eye. In the paper, Drydakis compares invitations to interviews at 344 Greek firms between CVs that were designed to be interpreted as being for Greek-born applicants with Greek names (natives), foreign-born applicants with foreign names (non-natives, where the names were obviously Albanian, Ukrainian, or Georgian), and Greek-born applicants with foreign names (natives with an ethnic minority background). He finds that:

...natives with an ethnic-minority background have a 17.5 percentage points lower chance of receiving an invitation for interview than natives. Also, it is observed that non-natives have a 20.1 percentage points lower chance of receiving an invitation for interview than natives. Both estimates are statistically significant at the 1% level. However, the two estimates are not statistically significantly different...

In other words, Greek-born applicants with non-native names and foreign-born applicants are essentially discriminated against to the same extent. That suggests that discrimination in this context is to some extent taste-based discrimination (where employers have preferences not to employ the minority group), rather than statistical discrimination (where belonging to a particular group is statistically associated with lower productivity, which might arise for instance when that group has lower education). We can infer that statistical discrimination is less likely, because the CVs of the Greek-born applicants with non-native names showed the same educational and labour market backgrounds as the CVs of the native applicants.

Also interesting is that the discrimination extends to wages as well:

...natives with an ethnic-minority background are invited for interviews for vacancies that offer 5.5 percentage points lower wages compared to natives. Moreover, the estimates suggest that non-natives are invited for interviews for vacancies that offer 6.4 percentage points lower wages compared to natives. Both estimates are statistically significant at the 1% level. However, the two estimates are not statistically significantly different...

So again, there is no difference between the Greek-born applicants with non-native names and foreign-born applicants. This should be especially disappointing for children of immigrants. I wonder if these results hold in other countries besides Greece? I also wonder to what extent these effects might be moderated by contact with minority groups, as per the contact hypothesis I discussed yesterday. Those suggest some potential avenues for future research.

Read more:

• What's in a (first) name?

The contact hypothesis, and African American GIs in Britain

The contact hypothesis posits that contact between members of majority and minority groups (under certain conditions) can reduce prejudice towards the minority group. In a new test of this hypothesis, this article by David Schindler (Tilburg University) and Mark Westcott (Munich Graduate School of Economics), published in the journal Review of Economic Studies (ungated earlier version here), uses data on African American GIs in Britain. They introduce their paper as:

In this paper, we show that the temporary presence of African American G.I.s... in the UK during World War II persistently reduced anti-minority prejudice amongst the British population. As the base of the US military’s European operation, the UK played host to over one and a half million US troops during World War II. Around 150,000 of these troops were black, serving in segregated units with non-combat support duties such as transport and supply... Many Britons thereby saw and interacted with non-whites for the very first time. Despite pervasive racist attitudes before the war, we show evidence from surveys that these interactions were positive experiences for both the local population and for black G.I.s.

More specifically, they derive a measure of which local areas in England and Wales had military bases where African American support units were stationed. They focus on support units because:

As in previous wars, black soldiers served in racially segregated units, normally under command of white officers. With few exceptions, black troops were limited to non-combat “labour” or “service” roles, most often supply and quartermaster services, transport, food preparation, and sanitation...

Schindler and Westcott then compare areas in England and Wales that were more, or less, exposed to African American troops (controlling for the presence of other support troops) during World War II. They make this comparison in terms of a number of measures that are plausibly associated with racial bias in more recent times, including: (1) membership of the far-right BNP political party (using a membership list published online in 2008); (2) local election results (in terms of the share of votes for the BNP in elections from 2006-2012; or the share of votes for the Conservative Party from 1973-2012); (3) online data from Implicit Attitudes Tests taken by UK residents between 2004 and 2013; and (4) online survey data (from the same source as the IAT data) on 'warmth of feelings' towards African Americans. They find that:

...individuals in areas of the UK where more black troops were posted are more tolerant towards minorities 60 years after the last troops left. First, we show that such areas contain fewer members of the British National Party (BNP), a far-right political party with racist policy positions. Next, we demonstrate that voters in affected areas were less likely to vote Conservative in local elections during times when far-right parties were not widely fielding candidates (until the early 2000s). This effect disappears as the BNP emerged as the strongest far-right party and BNP candidates subsequently received fewer votes in locations where black G.I.s were posted. Finally, we show that there is less implicit anti-black bias in these areas, as measured by Implicit Association Test (IAT) scores, and that those living in locations where black G.I.s were posted report warmer feelings towards black people.

The size of the effects are small. Schindler and Westcott provide some additional analyses where they mathematically manipulate their results to account for the number of generations that have passed between World War II and the present day, and the proportion of the population that might have been exposed to African American GIs. Those adjustments rely on some heroic assumptions, and I don't find them compelling. However, in terms of the contact hypothesis, their analysis separated into rural and urban areas is potentially more important, where they find that:

...the effect of black troops on BNP membership is about twice as large in rural areas as in urban areas. In fact, the effect in urban areas is not statistically significant at any conventional level, despite a larger sample size.

That's consistent with the contact hypothesis because urban areas in England and Wales already had minority populations prior to World War II, so exposure to African American GIs is likely to have a much smaller (in this case, effectively zero) impact. Schindler and Westcott only provide this disaggregated analysis for the BNP membership data - in the other analyses, they simply limit their results to the rural population. Perhaps we are to take from that that the results only hold for rural populations overall.

At the start of this post, I noted that the contact hypothesis requires certain conditions. Schindler and Westcott come back to those conditions in their conclusion:

Taken as a whole, our results provide support for the “contact hypothesis” (Allport, 1954), which postulates that contact between groups can reduce animosity towards the minority group, and show that such effects can persist in geographies across time.

It is interesting to note that the contact which we describe meets many of the conditions that Allport postulated were necessary for intergroup contact to lead to improved relations: equal status, common goals, intergroup cooperation, and personal interaction. Black G.I.s were in the UK for a relatively short period of time, to support the war effort, and did not compete for jobs or public goods with the local population.

Of course, those conditions are not always met. For instance, in many Western countries in recent times, some migrant groups or refugee groups would struggle to meet any of the four conditions. That suggests a role for government or non-government organisations to improve the conditions of contact between these groups and the majority population.

[HT: Marginal Revolution, last year]

Experimental results on diversity and team performance in Kenya

In David Epstein's 2019 book, Range (which I reviewed here), he outlines a positive case in favour of diversity in teams. Epstein noted that bringing a variety of viewpoints to a problem enhances the chance of a creative and positive solution. However, there is a trade-off. Greater diversity increases the costs of agreeing on a decision. Different viewpoints create conflict, and increase the costs of communication between team members. Epstein doesn't outline that counter-argument at all, or that a balance must be drawn between the benefits of diverse viewpoints and costs of more challenging communication.

The research literature isn't much help either. Almost all of the research is based on observational studies that suffer from problems of establishing causality (as was the case with the two studies I discussed in my most recent posts, here and here). It is difficult to generate a randomised experiment that would allow researchers to attribute causality to the differences in diversity in teams. Difficult, but not impossible. In a recent article published in the Journal of Public Economics (ungated earlier version here, and a non-technical summary by Richard Holden of University of New South Wales was published in The Conversation back in April), Benjamin Marx (Sciences Po), Vincent Pons (Harvard), and Tavneet Suri (MIT) report on a randomised experiment they conducted in Kenya. As they explain:

The context of this study was a door-to-door canvassing campaign conducted in collaboration with the Independent Electoral and Boundaries Commission of Kenya (IEBC) in 2012. The canvassing experiment aimed to increase voter registration in Nairobi’s largest informal settlement, called Kibera...

Subsumed within this registration experiment, we designed a field experiment to study how diversity within teams and along the organizational hierarchy affects effort and performance. This involved three levels of randomization. First, we randomly allocated the canvassers to teams of two. Second, we randomly allocated each team to a supervisor responsible for monitoring them in the field. This created variation in vertical homogeneity, i.e., in the degree of heterogeneity between managers and the workers they supervised. Third, each team was randomly allocated a set of enumeration areas (EAs) to visit in a mandatory random order. This introduced variation in the degree of similarity between canvassers and the residents which staff members were expected to canvass.

This ingenious research design allows Marx et al. to capture three dimensions of diversity: (1) horizontal diversity (between team members); (2) vertical diversity (between teams and supervisors); and (3) external diversity (between teams and the residents they are canvassing). The randomisation was successful in generating horizontal and vertical diversity:

23% of the canvassing pairs (7 teams out of 30) were ethnically homogeneous on the horizontal dimension. The ethnicity of the manager matches that of one of the team members in 23% of cases, and there is no instance where the ethnicity of the manager matches that of the two team members.

They then compare horizontally diverse teams with horizontally homogeneous teams in terms of the number of completed visits conducted by the team, and the duration of visits. They also compare those outcomes for teams that are vertically homogeneous (where the supervisor shares an ethnicity with one of the team members) and teams that are vertically diverse. Marx et al. find that:

...ethnically homogeneous teams on the horizontal dimension are approximately 8 percentage points more likely to complete a canvassing visit - about a 20% effect size. Such teams also conduct visits that are 0.8 to 1.3 min longer on average... Overall, a team conducts more and longer visits if it is composed of two co-ethnic canvassers...

...we find some evidence that the effects of vertical homogeneity are opposite in sign to those of horizontal homogeneity. If a manager and any one team member belong to the same ethnicity, the probability that the canvassing visit is completed decreases by about 3 percentage points, but this estimate is not statistically significant... The duration of visits decreases by 1.4 min...

In contrast to horizontal and vertical diversity, there are no statistically significant effects of external diversity. The results also contrast with gender diversity, where:

The results we find are much more nuanced than those we find for horizontal and vertical ethnic homogeneity. The effect of gender horizontal homogeneity is positive in all specifications, but is relatively small in magnitude and significant in only three out of eight specifications. We also show that gender vertical homogeneity does not significantly affect either measure of performance.

What could be driving the ethnic diversity results? Looking at time use data, as well as data from a staff survey conducted after the experiment, Marx et al. conclude that:

Overall, we interpret these results as evidence that the horizontal homogeneity acted as a monitoring and disciplining device, and improved work organization within the team, which facilitated learning and possibly led to increased socialization after the experiment. Meanwhile, vertical homogeneity led to less stringent norms and discipline and hence, lower effort and performance.

Marx et al. conclude that:

Our findings suggest much of the trade-off between diversity and homogeneity in organizations may come from the different effects diversity has along different dimensions of organizational structure. On the one hand, diversity may reduce efficiency within teams of workers, by creating communication costs or other frictions, leading to a worse division of tasks and lower performance. On the other hand, diversity along the organization’s hierarchy has the opposite effect in our context, since it improves both effort and performance.

Overall, in this experiment horizontally homogeneous teams were good, but in the vertical dimension diversity was better than homogeneity. The gains from vertical diversity are intuitive though, as is the mechanism. You don't want supervisors giving some teams, with which they share characteristics, an easier ride. 

The negative impact of horizontal ethnic diversity suggests that the negative communication effects of diversity dominate within teams in this context. However, it would be easy to over-generalise these results, especially because we generally see randomised experiments as a 'gold standard' for research in identifying causal effects. We should recognise that the teams in this experiment consisted of only two members. I would hesitate to extrapolate to larger teams. We shouldn't rush off and immediately reconstitute organisations into teams that are ethnically homogeneous. Benefiting from diverse viewpoints is likely to be more of a factor when the teams are larger. And the context was very specific. This wasn't the case of a team brought together to solve a problem, design a solution, etc.

The takeaway message from this paper is that diversity within organisations is a complex mix of horizontal and vertical diversity, and the effects across those two dimensions may differ in important ways. Organisations need to recognise that there is a balance to be achieved in operating with the optimal mix of diversity within and between teams.

[HT: Richard Holden on The Conversation, although I swear I had read about this research somewhere earlier]

Exposure to foreign-born students and the academic performance of U.S.-born students

Following on from yesterday's post about international programmes in the Netherlands (which was really about exposure to foreign students rather than international programmes per se), earlier this week the Economics Discussion Group at Waikato discussed this NBER Working Paper by David Figlio (Northwestern University) and co-authors. Figlio et al. looked at the effect of foreign-born high school students on the academic performance of U.S. born students. As with the paper by Wang et al. I wrote about yesterday, there is a selection problem and you can't simply compare U.S.-born students in schools with more foreign-born students with U.S.-born students in schools with more foreign-born students. As Figlio et al. explain:

First, immigrant students are not randomly assigned to schools, and are more likely to enroll in schools educating students from disadvantaged backgrounds... Second, US-born students, especially those from comparatively affluent families, may decide to leave when a large share of immigrant students move into their school district. Indeed, evidence shows that in the US, following an influx of disadvantaged students and immigrants, affluent, especially White, students move to private schools or districts with higher socio-economic status (SES) families, a phenomenon which has been labeled “white flight”... Both of these factors imply that immigrant exposure is negatively correlated with the SES of US-born students. Therefore, research that does not address the non-random selection of US-born students is likely to estimate a correlation between immigrant exposure and US-born student outcomes that is more negative than the true relationship.

Figlio et al. have access to very detailed data from the Florida Department of Education, which they link to birth records. This allows them to compare siblings, which is a smart way of dealing with the selection problem. Since both siblings tend to go to the same school, comparing siblings deals with the first selection issue above, since school-specific effects will be the same for both siblings. And because parents would likely move both siblings to a new school if they move one of them, then that deals with the second issue. Importantly, because each sibling is in a different grade, and the number of foreign-born students is changing over time, that provides variation between the siblings in their exposure to foreign-born students, and it is that variation that Figlio et al. test, for its association with academic performance.

Their dataset contains information on all K-12 students in Florida from 2003-03 to 2011-12. The most restrictive analysis (of siblings) has a sample size of more than 1.3 million (out of a total sample of over 6.3 million U.S.-born students who speak English at home). The key outcome variables are student performance on standardised reading and mathematics tests, which they standardise (so all results are measured in terms of standard deviations). Comparing foreign-born (which they label 'immigrant') students with U.S. born students in terms of test results, they find that:

Immigrant students’ performance in math (-0.097) and reading (-0.206) is lower than the one of US-born students (0.044 and 0.052).

Not too surprising then. In their main analysis, they find that:

...once selection is accounted for with family fixed effects, the correlation between cumulative immigrant exposure and academic achievement of US-born students is positive and significant. Moving from the 10th to the 90th percentile in the distribution of cumulative exposure (1% and 13%, respectively) increases the score in mathematics and reading by 2.8% and 1.7% of a standard deviation, respectively. The effect is double in size for disadvantaged students (Black and FRPL [free-or-reduced-price-lunch] eligible students). For affluent students the effect is very small, suggesting that immigrant students do not negatively affect US-born students, even when immigrants’ academic achievement is lower than the US-born schoolmates.

As with the Wang et al. paper from yesterday's post though, these results do not definitively demonstrate causality. However, Figlio et al. do a bit more digging, and find:

...suggestive evidence that the effect on US-born students is larger when the immigrants systematically outperform US-born students. Overall, these results suggest that immigrant students do not affect negatively US-born students, even when the immigrants’ academic achievement is lower than the US-born students, and may have a positive impact on US-born students when immigrants outperform them.

These are important results, and clearly run counter to the narrative that underlies the phenomenon of 'white flight' - that immigrant students make domestic-born students worse off academically. Although more research is needed to identify whether these results are causal, and to further explore the mechanisms that drive them, it is clear that diversity is better than (or at least not as bad as) many people expect.

[HT: Marginal Revolution]

The wage premium from studying in international programmes in the Netherlands

The Bologna process was an attempt to harmonise the higher education systems across Europe by adopting some common standards, and began in 1999. One of the consequences was a large increase in international programmes at postgraduate level, and many of those programmes are taught in English, which reduces to some extent the language barriers (at least, to the extent that English is a common language). How do graduates of these international programmes fare after graduation? That is the research question that this recent discussion paper by Zhiling Wang (Erasmus University Rotterdam), Francesco Pastore (University of Campania Luigi Vanvitelli), Bas Karreman, and Frank van Oort (both Erasmus University Rotterdam) addresses.

They compare Dutch students who studied in international programmes (in the Netherlands) with Dutch students who studied in domestic programmes. A simple comparison of those groups would be problematic though, because of self-selection into programmes. If better than average students tend to choose international programmes, then this would bias upwards the estimated effect of those programmes.

Wang et al. try to get around this by using a matching estimator - specifically coarsened exact matching (which I hadn't heard of before). Essentially, they match students in the international programme with students that fit into the same category, where categories are based on a combination of gender, year of graduation, field of study and university, cultural diversity in the bachelor programme, neighbourhood-level education, and father's income. They then compare students in the international programme with their matched groups who studied in a domestic programme.

Wang et al. use data from the graduates of Masters programmes of thirteen Dutch universities over the period from 2006 to 2014, with education data linked to detailed administrative data by Statistics Netherlands. The full sample has over 29,000 students, while the matched sample reduces this to around 8000). Comparing the labour market outcomes of the matched samples, they find that:

...students from international programmes obtain a wage premium of 2.3% starting from the 1st year after graduation, ceteris paribus. The wage premium keeps increasing by about 1% every year.

That's quite a large effect. So, what explains it? Wang et al. dig a little bit into the mechanisms underlying the difference, and find that:

...the wage premium is largely driven by differential choices in the first firm upon graduation, rather than cross-firm mobility or faster upward mobility within firm. Upon graduation, Graduates from international programmes are much more likely to choose large firms that have a higher share of international employees and have business of trade for their first jobs. [They] get a head start in wage level and the initial wage advantages persist in the long-run.

Joining an international programme appears to be an excellent investment. However, there are a couple of limitations with this study. First, the matching estimator does ensure that the comparisons are between students that are very similar in terms of their observable characteristics. However, that doesn't ensure that they are similar in terms of unobserved characteristics. Clearly, the two groups are different in some way that induces some students to choose an international programme, and others to choose a domestic programme. That choice is not random. That means that these results do not demonstrate causality - we can't say for certain that an international programme causes the wage premium.

Second, the definition of international programmes was a little odd to me. Wang et al. use a data driven approach. As they explain:

...we compute proxies for programme-level internationalization based on detailed information of students’ ethnic composition. We classify all students into different cells by year of graduation and study programme. For each cell, we make a count of “most likely English-taught foreign students” that satisfy three criteria: First, they are first-generation immigrants... Second, they have never lived in the Netherlands before they start master programmes... Third, they are originally from non-Dutch-colony areas, non-Dutch-speaking countries or non-German-speaking countries... Our preferred measure is that when the number of these students exceeds 4 or the share of these students exceeds 25% for the first time, the study programme is regarded as an international one from that year onwards.

Their proxy for an international programme is really proxying the exposure to foreign-born or foreign-educated students. It isn't so much about international programmes at all. It surprised me that they didn't simply investigate which of the university programmes are described as international programmes, taught in English, etc. and use that to construct a measure instead.

Nevertheless, exposure to diversity is important (more on that in my next post), and this research is at least suggestive that such exposure may lead to better jobs in larger and more internationally-linked (and, importantly, higher paying) firms.

[HT: Jacques Poot]

Ethnic segregation in Sao Paulo schools, and its relationship with employment and wages

Following Thursday's post about ethnic segregation and spatial inequality in Europe, I was interested to dig out this 2017 article from my to-be-read pile, by Gustavo Fernandes (Fundacao Getulio Vargas, Brazil), published in the journal World Development (sorry, I don't see an ungated version online). Fernandes used data from the 2005 School Census and the 2010 Population Census for the city of Sao Paulo, and looked at the association between segregation within public and private schools, and employment and wages for those aged 18-35. As motivation, he notes that:

The belief that Brazil has benefitted from an absence of racial and ethnic problems has been widely accepted over the last century. Brazil has often been described as a racial democracy.

Part of the motivation, then, is to debunk this 'myth'. I'm not quite sure that this counts as debunking though:

...our results show that Sao Paulo is not a city with a high degree of segregation, especially when compared to the U.S. In the city, approximately 21.29% of students would have to change schools to a new institution in order to achieve an equal composition of students by color among the entire student population of the city.

That's a fairly low level of segregation, compared not just with the U.S., but with many other countries (for instance, there's a lot of concern about segregation in the New Zealand school system). But is segregation related to inequality? Fernandes finds that, for Sao Paulo:

...segregation is correlated with the level of development in the region, which positively affects the expected returns of brancos and amarelos and negatively affects those of pretos e pardos. This result appears to be explained by the predominance of brancos and amarelos in private schools, despite the fact that most of the population of white students attends public schools. However, the effect of segregation becomes negligible when analyzing only the outcomes of students within the public school system.

The predominance of whites in private schools may be the main reason for the deep economic inequality found in Sao Paulo among races. Those schools provide a higher quality of education in comparison to public schools. They may also offer access to social networks that lead to better jobs. Both factors can exponentially increase the average income of the entire white population, resulting in large disparities between the wages of whites and the wages of pardos and pretos.

The brancos and amarelos (whites and Asians, respectively) tend to make up the majority of the class in private schools, and it is private school segregation (and not public school segregation) that is most associated with young adult employment and wages.

Ultimately, this paper demonstrates a result that is the opposite of the paper I discussed last Thursday, where greater segregation was associated with lower spatial inequality. It is impossible to reconcile the results given the wide difference in methods (not least the difference between cross-country analysis at the regional level, and small-area analysis of neighbourhoods within a single city in Brazil). However, this does demonstrate that more research on this topic is needed.

Ethnic segregation and spatial inequality

For the last few years one of my PhD students, Mohana Mondal, has been looking into ethnic segregation in Auckland (see this earlier post on some of her work). I've also maintained an interest in income inequality. So, I was really interested to read this 2017 article by Roberto Ezcurra (Universidad Publica de Navarra) and Andres Rodriguez-Pose (London School of Economics), published in the Journal of Economic Geography (ungated earlier version here), which links those two ideas. Specifically, Ezcurra and Rodriguez-Pose look at whether ethnic segregation (the concentrate of different ethnic groups within a country) matters for spatial inequality (income inequality between regions or areas of a country).

They use data on a cross-section of 71 countries where they have regional-level data on ethnic groups and region-level GDP per capita. After controlling for various factors known to affect spatial inequality such as the average size of regions, the degree of ethnic fractionalisation of the population (which is basically a measure of how many different ethnic groups there are in a country), the stage of economic development, trade openness, country size and whether a country is a transition country, they find that:

The coefficient of the index of ethnic segregation... is in all cases positive and statistically significant at the 1% level. This implies that more ethnically segregated countries have on average higher levels of spatial inequality...

This holds both for a basic regression specification, but also for an instrumental variables regression, where they attempt to demonstrate a causal effect of ethnic segregation on spatial inequality (as an instrument, they use segregation predicted using the ethnic composition of neighbouring countries). They also show that their results are robust to using alternative measures of segregation and inequality.

Ezcurra and Rodriguez-Pose then go on to investigate potential transmission channels that might explain this relationship. They find that:

...once political decentralisation and government quality are controlled for, the coefficient of the index of ethnic segregation still remains positive, but its effect on spatial inequality is statistically significant only at the 10% level... While not conclusive, these findings suggest the possibility that political decentralisation and government quality could be possible transmission channels linking ethnic segregation and spatial inequality.

The argument is that countries with more ethnic segregation are more likely to decentralise authority to their regions, which increases inequality between the regions.

This is a nice paper, but there are a couple of aspects of the research where some further work is needed. First, this research was based only on cross-sectional data. I would like to see some analysis that included a time dimension before I would conclude definitively that this relationship is causal. Second, the instrumental variables analysis seems fine on the surface, but only until you read this bit:

...the instrument used in the article predicts zero segregation for island countries...

It's pretty difficult to defend an instrument that results in such a wildly off-the-mark prediction. Certainly, you wouldn't want to predict zero ethnic segregation in New Zealand or Australia. I wouldn't expect an alternative conception of the instrument to change the results by a lot, but I think it is worth exploring. So, while this article is interesting, there is definitely more research required in this area.

 

Classroom ethnic diversity and grades

When parents try to get their children into good schools, there are typically two reasons: (1) they hope that the signal provided by the good school will lead to better outcomes for their children; and (2) they hope that their children will benefit from interacting with 'higher quality' peers. Peer effects have been subject to a large research literature (see this 2011 review by Bruce Sacerdote (ungated version here). For instance, there is evidence that being peered with high achieving students leads to higher achievement (although, interestingly, the book I reviewed yesterday suggested that the peer effects literature was less than robust).

The composition of peers may matter other than through their ability. This 2018 article by Angela Dills (Western Carolina University), published in the journal Economic Inquiry (sorry, I don't see an ungated version online) looked at how the ethnic make-up of the classroom affects grades. Specifically, she used data from 4435 non-honours students from 2009 to 2013, who were automatically enrolled into a compulsory course (Development of Western Civilization 101), and randomly allocated to a section. So, the students had no control over who their peer group in the class was, or the racial composition of the section. This randomisation (and the compulsory nature of the course) overcomes any selection bias.

Looking at how students' grades in the course are affected by the fraction of their classmates who are people of colour, she found nothing much significant in a linear model. However:

The effect of the percent of classmates of color on grades appears strongly nonlinear... Allowing for the quadratic term in percent minority, the estimates show no statistical difference in effects for whites and for nonwhites. At low levels of diversity, the effect of increasing the percent minority is negative; the sign of this effect turns positive with about 18% of classmates being students of color. For a class with 25% students of color, the effect of a 10 percentage point increase is a positive and statistically significant increase of about 0.1 grade points (p value=.003).

There is also some evidence that the effects differ between high-ability students and low-ability students (particularly for black students), but I don't find those results to be as convincing. The range of the proportion non-white students was rather narrow, with sections ranging from 0 percent to 30 percent non-white. It would have also been interesting to see some measures of ethnic diversity (rather than the fairly coarse measure proportion non-white) used in the analysis. Overall, we can consider this paper as providing some suggestive evidence that having diverse peers in the classroom is a good thing.

We count Pitcairn Islanders in New Zealand. Why not Timorese?

New Zealand has a standard classification of ethnicity that allows Statistics New Zealand (or other government agencies) to describe the ethnicity of people in a way that is consistent and comparable across all government agencies. The classification has four levels. The top level has essentially six categories: (1) European; (2) Māori; (3) Pacific Peoples; (4) Asian; (5) Middle Eastern, Latin American, and African (MELAA); and (6) Other ethnicities. For a long time, I've been critical of the Pacific, Asian, and MELAA groups as merging some extremely heterogeneous populations into a single category. For instance, the Asian group includes Japanese, Fijian Indian, and Afghani people, as if they were all in some way similar. The MELAA group is an even bigger nonsense.

Many researchers use data at the top level of the classification in a fairly uncritical way. Even those who view the data with a healthy dose of scepticism are forced to use it when it is the only level of data that is available. The level of ethnic disaggregation often depends on whether you are looking at the national level, the regional level, or more locally (and relates to the point I discussed in my post yesterday - in that research we used Level 2 ethnicity data, but previous researchers had focused on Level 1). For instance, below the national level, the data produced by Statistics New Zealand is often aggregated to Level 1 or Level 2 of the ethnic classification (for example, the Census data available at NZ.Stat is only available at Level 2 for territorial authorities, and for most of the cross-tabulations it is only available at Level 1). Consequently, most of the research (especially health research) I have seen uses the Level 1 classification.

The New Zealand Herald picked up on this issue in a story this morning:

An immigration expert has slammed what he sees as New Zealand's systemic failure to recognise minority ethnic and religious communities.

This comes after Statistics New Zealand revealed that one in seven failed to fully complete Census 2018.

Massey University sociologist Professor Paul Spoonley said the use of "crude categorisations" like Asian and Pasifika by authorities hid important differences and the true diversity of the nation.

Census participants were asked to list their race and ethnic origin. But those who identified with an "infrequent" or "unanticipated ethnic group" were put under a "not elsewhere classified" group.

Of course, the classification works that way for good statistical reasons. The smaller the number of people in an ethnic group, the greater the statistical error in any research that involves them, or in any table of data or cross-tabulation involving that group. And the smaller the group, the more likely that publishing numbers related to that group inadvertently compromises the confidentiality of the data. So, even if data were published related to small ethnic groups, there would be a lot of data suppressed due to small numbers. The standard classification of ethnicity tries to minimise this problem by grouping ethnic groups that have small numbers into the "not elsewhere classified" groups.

However, Paul makes an excellent point. So, let's have a look at those "not elsewhere classified" (NEC) ethnic groups, focusing on the Asian Level 1 ethnic group.

At Level 4 of the standard classification of ethnicity, the following groups fit under the Asian category [*]: Filipino; Cambodian; Vietnamese; Burmese; Indonesian; Lao; Malay; Thai; Karen; Chin; Southeast Asian NEC; Hong Kong Chinese; Cambodian Chinese; Malaysian Chinese; Singaporean Chinese; Vietnamese Chinese; Taiwanese; Chinese NEC; Bengali; Fijian Indian; Indian Tamil; Punjabi; Sikh; Anglo Indian; Malaysian Indian; South African Indian; Indian NEC; Sinhalese; Sri Lankan Tamil; Sri Lankan NEC; Japanese; Korean; Afghani; Bangladeshi; Nepalese; Pakistani; Tibetan; Eurasian; Bhutanese; Maldivian; Mongolian; and Asian NEC.

That is a lot of groups, but there are some notable omissions. Essentially it would be impossible to find any data from Statistics New Zealand or any government agency on any Asian group that does not appear on that list. The New Zealand Herald highlighted the omission of the Peranakan group in an article today, but we could add Timorese, Uighur, Rohingya, and any number of other small (and marginalised) groups, for which having access to data would be useful. I expect that Timorese and Rohingya are included in the Southeast Asian NEC category, but it would be hard to know (there is no detail in the classification to tell us).

Uighur might be included in the catch-all "Eurasian" group, which presumably also includes the Uzbek, Kazakh, Turkic, and Kyrgyz ethnic groups, etc. You might argue that these groups are small, and similar enough to group together, but I expect they would disagree. Or maybe the Uighur are in the Chinese NEC category. Again, it is hard to know.

The Pacific Peoples category has similar cases (for example, no category for Marshall Islanders, Palauans, or even Micronesians more generally). Don't get me started on the MELAA group - this post might end up thousands of words long.

The argument for 'ignoring' these groups, or merging them into other groups or into an NEC category is, as I noted above, in order to ensure that the merged groups are large enough to count. However, what constitutes an ethnic group that is 'too small to count' in Level 4 of the classification must be fairly arbitrary. Level 4 includes a category for Pitcairn Islander. Pitcairn Island has a permanent population of 50, so how many of them could possibly be living in New Zealand?

If we can count Pitcairn Islanders (population in the home location of 50), then surely we can count Timorese (1.3 million, in Timor-Leste, plus more in the western half of the island), Rohingya (around 1-1.3 million), or Uighurs (over 12 million). Not to mention the hundreds of other ethnic groups with small numbers in New Zealand. It's time to refresh New Zealand's standard classification of ethnicities to ensure that all minority groups count, and are counted.

*****

[*] I'm ignoring the "not further defined" category, which is where we know what category the person belongs to based on a higher level of the classification, but not at Level 4. That might occur, for instance, if someone says they are "Southeast Asian", which is a category at Level 2 of the classification, but not at Level 4.