Do religious congregations provide informal insurance against the risk of income loss for their congregants? A new paper examines the impact of an expansion in crop insurance subsidies over three decades and finds that as crop insurance take-up increased, religious adherence fell off. 



Experiencing a loss in income is a major risk confronting many households. While the government offers limited protection against a decline in consumption caused by unemployment, private insurance companies are reluctant to offer insurance against a loss of income because of adverse selection and moral hazard. For example, private insurance companies cannot easily detect which individuals will diligently search for employment to replace a loss of income. Therefore, premiums for income insurance are prohibitively high. The lack of cost-effective insurance to mitigate negative income shocks creates an opportunity for informal risk-sharing agreements to arise within religious congregations.   


Our study focuses on agricultural communities where incomes are sensitive to weather conditions. Information shared within the congregation can determine whether a farmer’s loss of income was caused by poor weather or low effort. An informal risk-sharing agreement within a religious congregation acts as an informal credit market to assist farmers whose poor harvest was caused by poor weather. For example, by temporarily delaying or reducing payments, the consumption of farmers experiencing a bad harvest due to poor weather is less dependent on their low current income. The risk-sharing agreement then requires these farmers to compensate their fellow congregation members in periods of good weather. Provided members of a religious congregation are employed in a variety of different occupations besides farming, the informal risk-sharing agreement within a religious congregation can diversify its exposure to agricultural risk. For example, a local dentist can extend credit to farmers experiencing a poor harvest via the risk-sharing agreement facilitated by their congregation.


Compared to formal insurance contracts, an informal risk-sharing agreement within a religious congregation can mitigate adverse selection since its rituals and prohibitions screen out uncommitted members. Access to the informal insurance can also be suspended for members that exert insufficient effort or fail to reciprocate, thereby mitigating moral hazard and ensuring the risk-sharing agreement is enforced. Furthermore, information sharing within a congregation is well established since the behavior of members is already monitored to ensure compliance with the congregation’s religious requirements. This monitoring can be extended to information regarding the effort of individual farmers within the congregation. Relative to other social networks, the ability of religious congregations to mitigate adverse selection and moral hazard, while enforcing their social contract is advantageous when facilitating risk sharing. 


Higher rates of religious adherence in counties with greater agricultural intensity provide empirical support for the use of informal risk-sharing agreements as a mechanism to mitigate agricultural risk. Agricultural intensity is measured as the proportion of a county’s land area that is cultivated by field crops such as wheat, corn, and barley. We limit our study to field crops that are produced almost exclusively by family farms.


Soil quality offers another proxy for exposure to agricultural risk since counties whose soil has lower water retention are more exposed to drought and flood risk.  Consistent with religious adherence offering informal insurance against agricultural risk, counties with poor soil and high agricultural intensity have the highest rates of religious adherence. Moreover, counties where agriculture is riskier due to poor soil have more religious adherents from “strict” religious denominations that are more adept at facilitating risk-sharing agreements by mitigating adverse selection and enforcing social contracts.


In response to natural disasters that prompted the federal government to offer disaster assistance to farmers, the government introduced subsidies to encourage the purchase of crop insurance. By subsidizing crop insurance, the government’s intention was to reduce the dependence of farmers on “free” disaster assistance. Crop insurance compensates farmers whose crop yields are lower than expected. Thus, crop insurance protects income, thereby protecting consumption, when harvests are poor (output risk). This protection is consistent with the increased use of crop insurance in a county being associated with lower income volatility in the county.


Although crop insurance has been available since FDR’s presidency, the introduction of government subsidies in 1980 increased the use of this formal financial contract. In response to a devastating flood in 1993 that caused widespread damage to agriculture in the Midwest, the Federal Crop Insurance Reform Act of 1994 dramatically increased government subsidies for crop insurance. As expected, these subsidies consequently increased the use of crop insurance.


What was the impact of government-subsidized crop insurance on religious congregations? Panel data at the county level between 1980 and 2010 indicates that religious congregations became smaller as religious adherence declined. Therefore, the use of crop insurance appears to have displaced religiosity as a mechanism for farmers to manage their exposure to agricultural risk. In other words, formal financial contracts displaced informal risk-sharing agreements within religious congregations in response to government subsidies that lowered the cost financial contracts.  In terms of economic significance, the use of financial contracts to mitigate household risk explains approximately one third of the time-series variation in religiosity at the county level.


The displacement of religiosity by crop insurance continues to occur in counties that did not experience any increase in their average income following the largest increase in subsidies for crop insurance. Consequently, lower-cost financial contracting rather than a higher income is responsible for the decline in religiosity we attribute to crop insurance.  Besides controlling for income, our empirical specifications also control for education and other demographic information such as marital status and ethnicity.


As expected, the decline in religiosity attributed to government subsidies for crop insurance does not occur in urban counties with low agricultural intensity. Moreover, the decline in religiosity occurs in counties that were not directly affected by the flood of 1993 that motivated the increase in government subsidies for crop insurance. These placebo tests provide additional support for the inverse relation between religiosity and low-cost financial contracts to manage household risk.


Crop diversification is a natural metric to measure the moral hazard associated with formal financial contracts since this determinant of agricultural risk is directly attributable to the actions of farmers. Consistent with an increase in moral hazard, the use of crop insurance decreases crop diversification without increasing crop yields. Past studies have documented the decline in crop yields associated with crop insurance and attributed this negative impact on productivity to a reduction in fertilizer usage by insured farmers. In contrast, the decline in crop yields associated with crop insurance is not explained by an expansion into lower productivity land. Instead, the displacement of informal risk-sharing agreements within religious congregations by formal financial contracts appears to increase moral hazard.


Overall, we conclude that a cost-benefit analysis leads households to replace the informal risk-sharing agreement within a religious congregation with a formal financial contract in the form of crop insurance. Despite improving our study’s identification, government subsidies for crop insurance do not limit its implications. Several technologies known collectively as FinTech are also lowering the cost of financial services by enhancing competition, transparency, and efficiency. An emerging literature confirms the informational advantages of Fintech relative to traditional credit scores when extending credit to households. In our setting, the use of “soft” information by FinTech poses a challenge to risk-sharing agreements within social networks. Indeed, as FinTech enables lenders to condition on non-traditional sources of information, including social media, FinTech may erode the advantage of using social networks to manage the risk associated with a temporary loss of income.


Henrik Cronqvist is Professor of Finance at University of Miami Business School. He is also Vice Dean for Faculty and Research as well as Director of Ph.D. Programs.


Mitch Warachka received his Ph.D. from Cornell University in 2000 and is currently a Professor of Finance at the University of San Diego’s School of Business Administration. 


Frank Yu is Associate Professor of Finance at China Europe International Business School (CEIBS).  Dr. Yu received his Ph.D. degree in finance from the University of Chicago and was a faculty member at CEIBS from 2009.


Disclaimer: The ProMarket blog is dedicated to discussing how competition tends to be subverted by special interests. The posts represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty. For more information, please visit ProMarket Blog Policy