Risk attitudes, investment behavior and linguistic variation

This paper explores the relationship between linguistic variation and individual attitudes toward risk and uncertainty. Linguistic variation refers to differences in linguistic forms across languages. According to the linguistic relativity hypothesis, differences in grammatical structures and the vocabulary may affect how speakers of distinct lan- guages perceive and think about the world. We develop a specific linguistic marker that classiffes languages according to the number of non-indicative moods in irrealis contexts in their respective grammars. These grammatical categories express situations involving uncertainty, and the frequency of their use may be closely related to the overall degree of uncertainty perceived by individuals. Using data from the Survey of Health, Aging and Retirement in Europe (SHARE) and World Value Survey (WVS), we show that speakers of languages where non-indicative moods are used more intensively are on average more risk averse. This evidence holds both across countries and within linguistically heterogeneous countries. The results are robust to the inclusion of additional set of regressors and several fixed-effect controls for individual characteristics. Finally, we use our linguistic marker to instrument individual attitudes toward risk in the structural model for financial assets accumulation.