Economists increasingly consider language to be an important cultural trait worth studying in its own right. Recent research has catalogued a number of intriguing correlations between the grammatical structure of languages and the economic behaviour of their speakers. This includes relationships between the presence of a future tense and individual inter-temporal decision making (Chen 2013), as well as between the presence of sex-based grammatical genders and gender inequality in the home, at work, and in politics (Givati and Troiano 2012, Malaksvian 2015, Hicks et al. 2015).
Yet, a key question arises. If languages co-evolved with other cultural norms of behaviour, as is the case, then these correlations could be spurious (Roberts et al 2015) or dependent on the institutional environment itself (Gay et al 2016a). How much can we really learn from studying languages? The answer, it turns out, is a lot.
In recent research, we use the epidemiological approach to studying language. Originally used in biology to disentangle the role of genetic factors from environmental influences, the epidemiological approach in social science typically refers to studying immigrant populations in a common institutional environment in an effort to isolate the impact of specific cultural factors. This approach has recently been applied to study cultural norms of behaviour including the formation and persistence of gender roles (Fernández and Fogli 2009, Fernández 2011, Blau et al. 2011).
Since migrants travel with some aspects of culture ingrained, but shed many of the external influences found in their home country, the epidemiological approach allows researchers to study behaviour as a function of specific cultural traits, alleviating many forms of spurious correlation. We discuss several research projects employing this strategy to shed light on one of the underlying causes of gender inequality.
Case 1. The formation and persistence of gender norms: Evidence from the household behaviour of migrants.
Women generally do more housework than men, but this pattern varies immensely in degree across countries and over time. Why?
Using time-use survey data from the US, we show that female immigrants coming from countries whose dominant language relies on sex-based grammatical distinctions bear a far larger share of housework (Hicks et al. 2015). The more heavily a language’s grammar employs sex-based grammatical distinctions, the greater the burden on women. These gaps are substantial, translating to 9% more housework by women and 28% less by men (compared to other immigrant households, which are already more skewed than non-immigrant households).
Moreover, the division of tasks in these households is more pronounced along stereotypical gender roles, with, for instance, men doing more automobile maintenance and home repair, and women allocating more time to cooking and cleaning. Interestingly, these patterns of behaviour exist even among individuals living alone, suggesting that bargaining and specialisation within marriage is not driving the association.
To study whether language could play a role, we divide migrants by age of arrival and compare individuals from the same country to one another on this basis. Researchers have shown that individuals learn languages best early in life, so late-arrivers will be more likely to speak their mother tongue instead of English.
Figure 1 depicts the additional inequity in housework borne by gender-marked female speakers (in minutes per day) in comparison to other female immigrants. These starkly skewed gender roles appear in the behaviour of individuals who migrate at age eight or later, consistent with evidence on the critical period of language acquisition (Bleakley and Chen 2010).
Figure 1 Gender bias in housework by age of migration
The pattern is driven by the behaviour of both sexes. Figure 2 depicts total housework per day. Among gender-marked speakers, females devote more time to housework, while males devote even less.
Figure 2 Housework by gender, gendered language (GA), and age or migration
(Regression adjusted means with confidence intervals)
Interestingly, we find that time since migration does little to attenuate this relationship (Figure 3). This suggests that gender roles, once learned, persist through life.
Figure 3 Gender bias in housework by time since immigration
Could these patterns be driven by other cultural influences before migration which are correlated with language? Yes, although these results suggest that any other influences would need to operate in the same way over the life cycle. Regardless of whether gendered language alone bolsters these norms, these findings empirically teach us something else. That is, gender norms of behaviour are formed early in one’s life, and once developed, become ingrained.
Case 2. Evidence from the labour market behaviour of migrants.
Hicks et al. (2015) use language to provide novel evidence concerning the precise patterns, timing of development, and persistence of an individual’s gender identity. But they cannot precisely attribute these norms to any single cause because many factors could influence us early in life when gender roles solidify. To isolate language, we again turn to the epidemiological approach and study a countervailing set of gender norms, those in the workplace.
In Gay et al. (2016c), we exploit detailed US census records to compare the labour market outcomes of immigrants who come from the same country and are of a similar ancestry, but who speak languages with varying structures. Even when we restrict to variation in migrants from the same country and of the same ethnic group through fixed effects analysis, the gender gaps we observe on the basis of language remain.
The gaps are large. Among female migrants to the US, those who speak a language which makes sex-based grammatical gender distinctions exhibit significantly lower labour force participation, hours worked, and weeks worked during the year. In the case of participation, the magnitude of the difference is about 11 percentage points when considering the mean immigrant.
Our analysis confirms that the bulk of gender roles cannot be attributed to exposure to gender in language. Instead, we find at least two thirds of the correlation between language and labour market outcomes arises from other cultural forces, leaving at most one third which could be attributed to the direct impact of linguistic structures.
Case 3. A historical perspective
In Hicks et al. (2015) and in Gay et al. (2016c), we examined household labour and labour market choices among modern day immigrant populations. But there have been remarkable changes in gender inequality over time, both in the US and abroad. For instance, the gap in labour force participation rates has considerably closed throughout the 20th century among immigrants to the US (Figure 4).
Figure 4 Labour force participation rates of immigrants to the US (1910 to present)
We explore this further in Gay et al. (2016b), exploiting historical census data from 1910 to the present. What is particularly interesting is that during this period, the pool of immigrants to the US has evolved. At the beginning of the 20th century, the largest waves of migration originated from Europe, but over time, the source of migrants shifted more towards origins in Asia and Latin America.
What bearing have changes in the composition of the immigrant pool and in US labour market opportunities had on the relationship between gender in language and gender inequality? As Figure 5 shows, even after accounting for country of origin fixed effects, the answer is surprisingly little.
Figure 5 The correlation between language structure and the gender gap in labour force participation among immigrants to the US
Over the entire period from 1910 to the present, we find evidence that speaking a language with sex-based genders is associated with lower labour market participation of women. These results suggest that the observed relationship between language and gender roles is remarkably robust.
In a 2012 Vox column, we concluded that the study of migrants might be a fruitful avenue in the future to pursue the fascinating question of whether language shapes the economy (Gay et al. 2012). Our research suggests that indeed, languages can tell us much about behaviour.
First, linguistic differences can be used to uncover new evidence such as that concerning the formation and persistence of gender norms. Second, as the observed association between gender in language and gender inequality has been remarkably constant over the course of the 20th century, language can play a critical role as a cultural marker, teaching us about the origins and persistence of gender roles. Finally, the epidemiological approach also offers the possibility to disentangle the impact of language from the impact of country of origin factors. Our preliminary evidence suggests that while the lion’s share of gender norms can be attributed to other cultural and environmental influences, yet a direct role language should not be ignored.
Future work on language and economics must advance explanations regarding the origin of differences in language structure and apply the epidemiological approach to analyse features of language other than gender. Recent work by Galor et al. (2016) is an excellent step in this direction. Experimental approaches can also provide additional insights regarding this emerging field.
Blau, F D, L M Kahn and K L Papps (2011) "Gender, source country characteristics, and labor market assimilation among immigrants", The Review of Economics and Statistics, 93.1: 43-58.
Bleakley, H and A Chin (2010) “Age at arrival, English proficiency, and social assimilation among US immigrants”, Am Econ J: Appl Econ, 2(1): 165–192.
Chen, K M (2013) “The effect of language on economic behaviour: Evidence from savings rates, health behaviours, and retirement assets”, American Economic Review, 103(2): 690-731.
Fernández, R (2011) “Does culture matter?” in Handbook of Social Economics, 481-510.
Fernandez, R and A Fogli (2009) "Culture: An empirical investigation of beliefs, work, and fertility", American Economic Journal: Macroeconomics, 1.1: 146-177.
Galor, O, Ö Özak and A Sarid (2016) “Geographical origins and economic consequences of language structures”, Available at SSRN.
Gay, V, E Santacreu-Vasut and A Shoham (2012), “Does language shape our economy? Female/male grammatical distinctions and gender economics”, VoxEU.org, 29 August.
Gay, V, D Hicks and E Santacreu-Vasut (2016a) “Migration as a window into the coevolution between language and behaviour”, in The evolution of language: Proceedings of the 11th international conference.
Gay, V, D Hicks and E Santacreu-Vasut (2016b) "Language and gender roles among immigrants to the US: A historical perspective" in Studi di Genere: Il Mondo Femminile in un Percorso Interdisciplinare, R Edicusano and P Paoloni (eds).
Gay, V, D Hicks, E Santacreu-Vasut and A Shoham (2016c) "Decomposing culture: Can gendered language influence women’s economic engagement?" Fox School of Business, Research Paper No 15-046.
Givati, Y and U Troiano (2012) "Law, economics, and culture: Theory of mandated benefits and evidence from maternity leave policies", Journal of Law and Economics, 55.2: 339-364.
Hicks, D, E Santacreu-Vasut and A Shoham (2015) “Does mother tongue make for women’s work? Linguistics, household labor, and gender identity”, Journal of Economic Behaviour & Organization, 110: 19-44.
Roberts, S G, J Winters and K Chen (2015) "Future tense and economic decisions: Controlling for cultural evolution", PloS One, 10.7: e0132145.