Home » Projects » Praying in the Desert

Praying in the Desert

Praying in the Mediterranean

There is a long tradition of research investigating the relationship between geography/climate and economic outcomes (geographical determinism). More recently, this theory has been revised and the new research argues that the effect of geography on economic development might be indirect via institutions: geography affects institutions and institutions affect economic development. Our research shows that geography, and in particular the local climate, affects religious institutions and religious beliefs. Religious institutions then affect both secular institutions and economic development directly.

Our previous research has focused on Mediterranean regions and the Christian faith, but we believe the same mechanisms apply more broadly. The goal of this project is to extend the analysis of this mechanism to the rest of the world.
The Ethnographic Atlas is a dataset created by the anthropologist George Peter Murdock, and then extended by Patrick Gray. It contains information on all ethnic groups in the planet, historically. Although it contains detailed information regarding the culture of these societies, it only contains limited information regarding religious beliefs. Our goal is to extend the information in the atlas by adding information on whether each particular ethnicity has a ritual or a ceremony to ask the Divine for intervention by changing the weather, e.g., whether they pray for rain.

Experimental Persuasion

Motivated by our findings, we are interested on whether weather variability was used by religious authorities to increase religiosity and their legitimacy. Bayesian Persuasion (BP, Kamenica and Gentzkow, 2011) provides a natural framework. However, in the original formulation they assume that the priest (sender) could choose any experiment to persuade the peasants (receiver). In other words, the model does not allow variation in the experiments available to the sender, and therefore is not well suited to test our model over different regions with different climates.
In this article, we extend the original BP by looking at cases when the sender is limited on the set of experiments that he can perform. We show that the BP optimal is not achievable in general, and would require very strong assumptions in most applications. Moreover, we provide results in settings where either the experiment or the signal (or both) is not directly observed by the receiver. These settings are irrelevant in the original BP framework.

Skip to toolbar