Over the past decade, widespread financial illiteracy has been found among households in many countries, with greater concentration among certain demographic groups that include women and minorities. Strong correlations have been observed between limited financial literacy and suboptimal household financial behavior or inferior economic outcomes, such as lack of saving for retirement or lower wealth levels. Establishing causality from poor financial literacy to those outcomes, however, has remained particularly challenging. A number of authors have proposed econometric instruments for financial literacy that go back in time, in order to uncover an exogenous influence on current financial behavior or outcomes. In order for this approach to work, the influence of the exogenous initial factor (the instrument) must run solely through financial literacy or, at most, through other observable factors for which researchers can control in the econometric analysis.

Unfortunately, going back in time to find important exogenous influences on current financial literacy and financial behavior militates against the requirement that all possible channels of influence be controlled. Factors that operate in early life tend to have pervasive and multi-faceted influence on behavior or outcomes, likely to extend beyond financial literacy and other factors that are observable in the data.

Our new working paper ‘Financial Literacy Externalities’ takes a different approach. Instead of examining the link of a household’s financial behavior or outcomes to its own financial literacy, it studies instead the exogenous role of financial literacy in the environment within which the household operates. If the financial literacy of neighbors can influence that of the household, the case for a causal role of financial literacy is significantly strengthened. Moreover, financial education or related programs that boost financial literacy become more cost-effective, as they are shown to affect not only those directly involved in them but also their neighbors and peers.

Now, a problem with establishing exogenous peer effects is that people choose their neighbors and the area in which they live. For example, households more likely to exhibit sound financial behavior may tend to pick a neighborhood with more financially literate neighbors and with features that facilitate financial behavior (eg, banks and other financial institutions). If this is so, a positive relationship between financial literacy of neighbors and financial outcomes of the households under study may be a mere correlation that cannot be exploited for policy design.

Our paper uses unique administrative data and a quasi-field experiment of exogenous allocation of refugees in Sweden between 1987 and 1991 to uncover and estimate effects of being exposed to financially literate neighbors on household financial behavior ten to twenty years later. The advantages of this setup for our purposes are that the initial neighborhood of the household was not a result of choice possibly reflecting unobserved factors, and that the characteristics of the greater neighborhood, observed and unobserved, can be econometrically controlled and isolated from the share of financially literate neighbors in a smaller area.

The paper contributes evidence of a causal impact of financial literacy on behavior and points to a social multiplier of financial education initiatives. We find that exposure to a large share of financially literate neighbors in the electoral district of initial placement promotes the probability of saving for retirement in the medium run (10 to 15 years later) and of holding stocks in the longer run (15 to 20 years) after initial placement. The effect is larger when neighbors have economics or business education, rather than having quantitative education broadly defined. When splitting the sample, we find that significant effects of financially literate neighbors are only observed on educated or male-headed households. Taken together, these findings imply that financial literacy externalities are not the product of mere imitation of financial behavior of the neighbors, but they involve the transfer of financially relevant information, the ability to process it, and the confidence to apply it. This view is supported further by our finding that the share of neighbors with business or economics education has a much bigger estimated effect on financial behavior than the share of neighbors who participate in private retirement saving.

All in all, the findings in this paper support a causal role of financial literacy on household financial outcomes and strengthen the case for adopting (early life) financial education programs that can improve the financial choices not only of participants, but also of their social circle. In a completely different context, the use of a refugee sample, albeit for econometric reasons, suggests that initial placement of refugees close to financially literate neighbors can have longer-term effects on the financial behavior of those who are able to absorb and use the information conveyed to them by such neighbors.

Michael Haliassos is Professor and Chair of Macroeconomics and Finance at the Goethe University Frankfurt, a Research Fellow of the Centre for Economic Policy Research and an International Research Fellow of NETSPAR, Thomas Jansson is a researcher at the Sveriges Riksbank, and Yigitcan Karabulut is an Assistant Professor of Finance at the Rotterdam School of Management and a Network Associate of the Centre for Economic Policy Research.