Money talks: investigating the relationship between linguistic diversity and financial inclusion

Differences in languages spoken within a population can be thought of as transaction costs that make economic activities more difficult. This perspective has motivated a host of academic literature analyzing the linguistic profile of countries in relation to different socio-economic variables. Among these studies, financial inclusion is rarely one of the variables of interest. Language and financial inclusion are sometimes analyzed together in more granular studies of a single country, or even of individuals, but never in a cross-sectional, country-level analysis. However economic growth, which is generally considered to be positively related to financial inclusion, has frequently been studied, with mixed results. Earlier researchers of the question identified negative relationships between economic growth and linguistic diversity, in what became known as the “Fishman-Pool Hypothesis”. Later researchers determined that such a relationship did not exist, or that, in certain contexts, linguistic diversity and economic growth could even be positively related. This study departs from the intuition that financial inclusion's relationship to linguistic diversity may parallel that of economic growth – a relationship that seems intuitively negative but is more ambiguous after analysis. To overcome the broad interpretability of the concepts of interest, this study constructed two dependent variables representing financial inclusion, and four independent variables representing linguistic diversity with cross-sectional data for a sample of 61 countries. The models were estimated by accounting for multicollinearity of the regressors, as well as heteroskedasticity and non-normality in the error terms using the Seemingly Unrelated Regressions models and ordinary least squares estimation techniques. The results indicate that linguistic diversity indicators were all nearly zero, and highly insignificant, despite the strong specification of the models. This suggests that linguistic diversity has no significant relationship – positive or negative – to financial inclusion at a country level. This result was consistent across all the possible combinations of the operationalized variables for both concepts