The recent riots in Charlottesville, the renewed question of affirmative action policies in colleges and universities, and the more general rise of ethnic nationalism associated with the Trump campaign and presidency have brought back to the forefront a seemingly never-ending argument about the potential benefits and costs of diversity. From a broader angle, the question is centered on the structure and functioning of society. Are we all better off interacting with a diverse group of individuals, or does this diversity of individuals and thoughts actually lead to clustering, splintering and ultimately group-level conflict?
To summarise the current literature: both effects appear to be at play.1 The pros of diversity are roughly due to the combination of idiosyncratic perspectives that lead to unique, and often better, solutions in problem solving. Empirically, a higher level of identity based diversity (for example, sex or ethnicity) is associated with increases in innovation in the private sector (measured by patent applications) and more citations in academia.
The cons of diversity are due to the fact that idiosyncratic views often times will not mesh and that people generally like others similar to themselves (‘birds of a feather flock together’). This causes group formation and competition that may be harmful to productivity or, in the extreme, may lead to ethnic-based violence and conflict.
processes of pros and cons, costs and benefits, may suggest a ‘Goldilocks effect’ of diversity. Too much diversity may mean that conflict could overshadow the potential for new ideas, not enough diversity could suggest too few ideas and innovation but also limited conflict. Strife versus stagnation are the two extremes, but is it possible that a middle level of diversity could lead to more innovation, wealth and development without the accompanying disadvantages?
Genetic diversity and economic development
The literature on the pros and cons of diversity dovetails with recent work by Ashraf and Galor (2013) that proposes that genetic diversity plays a causal role in economic development. In short, the primary argument is that within-group genetic diversity serves in much the same way as the previously discussed between-group diversity. To show this, the authors exploit historic migration patterns of Homo sapiens out of East Africa. Migration reduced the amount of genetic diversity available to a particular region due to emigrants containing only a sub-set of the genetic variants available to the larger domestic population. The repetition of this process, which is known as the ‘serial founder effect’, led to a linear decrease in genetic diversity for areas that are further from East Africa along the proposed migration routes (seen in Figure 1).
Figure 1 Out-of-Africa migration routes
Source: Reproduced from Figure 3 of Ashraf and Galor (2013).
Using these migration routes, which are presumably unrelated to any other measures or causes of economic development, Ashraf and Galor (2013) show that country-level genetic diversity has a Goldilocks-type relationship with economic development. Too much genetic diversity is associated with an increased level of ethnic conflict that results in losses to economic output, while too little genetic diversity leads to the loss of potential innovations that result in stagnation in output.2 This set of findings implies that an optimal level of genetic diversity could exist where economic output is maximised. Indeed, the authors show a strong statistical relationship for this Goldilocks effect on economic development, both historically (before 1500 CE) and today.
Like any research paper, the findings are not without limitations. For example, it is possible that measures of genetic diversity are ‘tagging’ (i.e. are statistically associated with) other features of countries, like natural resources or easier trading routes, that have led to economic development over time. It is also possible that measures of genetic diversity are related to race/ethnicity genetic signatures that are tied to a history of colonialism and racism throughout world history that has produced differences in the development of countries. Ashraf and Galor (2013) are attentive to these issues, but no single study can be decisive.
Genetic diversity and later-life outcomes of Wisconsin students
Our own work seeks to combine the broader diversity literature with the use of genetic diversity from Ashraf and Galor (2013) and to overcome some of the potential limitations in previous research (Cook and Fletcher 2017). To do so, we focus on population aggregations within a single racially homogenous state of a single country – Wisconsin. Data for our study come from the Wisconsin Longitudinal Study, which as the name suggests is a longitudinal survey dataset of a random one third of 1957 graduating Wisconsin high school seniors. The study contains a rich amount of data throughout the graduate’s life course, and importantly for the 2003 wave, graduates (and selected siblings) were genotyped for roughly 100 genetic variants, which are used to compute our measure of genetic diversity.
Our hypothesis is that exposure to higher levels of genetic diversity during the formative years of adolescence encouraged a set of unique perspectives that shape individual life course trajectories. In other words, exposure to a varied set of peers – measured by a school’s genetic diversity – broadened an individual’s interests, views and ultimately, the individual’s problem-solving ability. Therefore, we propose that the Wisconsin Longitudinal Study graduates that attended more genetically diverse schools will have better socio-economic outcomes.
The setting of our sample, 1950s Wisconsin, presents unique opportunities and challenges. As expected, the ethnic composition of our sample is nearly uniformly ancestral to Europe (mostly Germany, Poland, Norway, Britain and Ireland). This limiting of potential ethnic diversity is beneficial for narrowing down the effect of genetic diversity by reducing the role of ethnocentric cultures that may play a role in determining economic conditions. The limited ethnic range of our study, however, means that our findings are not as generalisable as those of Ashraf and Galor (2013), and we may not have clear distinctions in race or ethnicity to account for the negative effects of diversity, notably group conflict.
Despite our sample being composed solely of individuals of European ancestry, we find a strong positive and statistically significant association between genetic diversity at the high school level and students’ later-life years of schooling, job prestige and income. This seen in Figure 2, which plots equal-sized bins of the positive linear relationship between years of schooling and high school genetic diversity. For years of schooling, a one-standard-deviation increase in genetic diversity is associated with roughly one more month of schooling on average. The statistically significant and positive relationship between genetic diversity and each socioeconomic outcome remains unchanged after controlling for unobserved county-level factors, individual ability (namely IQ), and parental input (parents’ education and income). All of these results support our primary hypothesis.
Figure 2 High school genetic diversity and years of schooling
Notes: This figure bins high school genetic diversity into 20 equally sized bins.
Source: Reproduced from Figure 2 of Cook and Fletcher (2017).
As a way to further understand our proposed mechanisms, we explore the association between genetic diversity and personality traits tied to divergent thinking, particularly openness to experience and extraversion. In doing so, we once again find a positive and statistically significant effect of genetic diversity. On average, individuals that attended more genetically diverse schools tend to score higher on indexes of openness, extraversion and a statistical combination of the two (which are associated with divergent thinking).
A complementary strain of research also suggests that diverse populations are likely to form complementary relationships with regards to task specialisation. In other words, a varied group will choose varied occupations. This is shown for ancestral populations by Depetris-Chauvin and Özak (2016) and in the contemporary influx of immigrants by Peri and Sparber (2009). To test this idea in our setting, we create a high school measure of occupational diversity for each graduate’s first job. Once again, and in line with prior studies, we find a positive and statistically significant association between this measure of occupational diversity and our high school measure of diversity.
As a final check on our analysis, we perform a semi-replication focused on the graduates’ parents. Parents in the Wisconsin Longitudinal Study are, on average, at high school in 1920, so we are able to use Ager and Brueckner’s (2017) 1920 data on county-level genetic diversity from European immigrants to examine parent-level outcomes. In short, we find similar results as in our high school analysis. Parents from more genetically diverse counties have more years of schooling, more prestigious occupations and a higher socioeconomic status.
Our findings suggest that diversity does indeed have beneficial effects to individuals within a relatively homogenous population. This supports our hypothesis that early-life exposure to diversity is embedded within an individual and that it broadens their perspectives and increases problem-solving ability.
We see genetic diversity as a complement to other identity based diversity measures. Genetic diversity quantifies base differences both within and between populations in a way that identity based measures are unable to capture. Specifically, genetic diversity is not dependent upon respondents’ personal identity but is based on the presence of particular genetic variants. For example, the ‘one-drop’ rule may create a disconnect between ethnic identity and actual ancestry. Furthermore, current measures of ethnic fractionalisation and diversity treat ethnicities as equally distinct, so that a population composed equally of British and Irish would be measured as diverse as a population composed equally of British and Han Chinese. Genetic diversity avoids these issues by measuring base-level neutral genetic changes that capture population stratification, representing historical separation and isolation.3
Our findings support the work of Ashraf and Galor (2013). In our homogeneous sample with (relatively) limited cultural variation, we observe a consistent beneficial socioeconomic effect of exposure to genetic diversity. This finding corroborates the positive mechanism of Ashraf and Galor (2013) and, more broadly, of the diversity literature.
As a next step, we will seek to examine these effects in a more ethnically/racially diverse setting. To do so, we are in the process of using the Add Health dataset, which has both genetic data and life-course data for a representative sample of high school students across the contemporary US. Our primary goal is to examine whether there is indeed a Goldilocks-type effect of diversity (as identified in Ashraf and Galor 2013) when considering a more globally representative population and also to further explore the potential mechanisms that appear during adolescence that translate into life course effects.
Ager, P and M Brückner (2016), “Immigrants’ genes: Genetic diversity and economic development in the US”, working paper.
Arbalti, E, Q Ashraf and O Galor (2015), “The nature of conflict”, working paper.
Ashraf, Q and O Galor (2013), “The “out of Africa” hypothesis, human genetic diversity, and comparative economic development”, American Economic Review 103(1): 1-46.
Ashraf, Q and O Galor (forthcoming), “The macrogenoeconomics of comparative development”, Journal of Economic Literature.
Cook, C J and J Fletcher (2017), “High school genetic diversity and later-life student outcomes: Micro-level evidence from the Wisconsin Longitudinal Study”, NBER working Paper No. 23520.
Conley, D and J Fletcher (2017), The genome factor: What the social genomics revolution reveals about ourselves, our history and the future, Princeton University Press.
Depetris-Chauvin, E and Ö Özak (2016), “Population diversity, division of labor and comparative development”, working paper.
Kemeny, T (2017), “Immigrant diversity and economic performance in cities,” International Regional Science Review 40(2): 164-208.
Peri, G and C Sparber (2009), “Task specialization, immigration, and wages”, American Economic Journal: Applied Economics 1: 135-169.