[Paper Review] Sacred Ecology

Neha Deopa, Daniele Rinaldo|arXiv (Cornell University)|Nov 9, 2023

Religion, Society, and Development4 citations

TL;DR

This paper proposes that African Traditional Religions (ATR) significantly reduce forest extraction through sacred ecological beliefs, using a mean-field non-market interaction model to show that individual ATR adherence lowers resource use, while higher population-wide adherence increases localized conservation. Empirical tests in Benin using instrumental variables and spatial discontinuity designs confirm a causal, positive impact: a 1 standard deviation increase in ATR adherence raises forest cover change by 0.4 standard deviations, with a 6.9% average forest loss in a counterfactual world without ATR beliefs.

ABSTRACT

Can religions shape ecosystems? We explore the role religious beliefs play in human-environment interactions by studying African Traditional Religions (ATR), which place forests within a sacred sphere. We focus on the unique case of Benin, whose history is deeply intertwined with traditional religions and where adherence is reliably reported. By exploiting three sources of exogenous variation in Benin's exposure to Charismatic Pentecostalism, we find that increase in ATR adherence yields positive changes in both forest and tree canopy cover. This increase is driven by sustainable land use policies rather than cooperation and shared governance mechanisms. To understand how ATR beliefs shape the way individuals combine the sacred and the ecology in their preferences, we build a theoretical framework of deforestation with heterogeneous pro-environmental attitudes driven by ATR adherence. Bringing the model to the data, we estimate that without any ATR adherence Benin would experience a loss of 10% of its tree canopy area, and would exhibit unsustainable deforestation rates. Our results show how ATR beliefs can play a fundamental role in forest and ecosystem conservation.

Motivation & Objective

• To investigate whether religious beliefs, particularly African Traditional Religions (ATR), shape environmental outcomes through sacred ecological worldviews.
• To model how heterogeneous ATR adherence and population-level belief distributions affect forest cover dynamics via non-market interactions.
• To test empirically whether ATR adherence causally influences forest cover change in Benin, using unique freely reported data on ATR adherence.
• To estimate the counterfactual impact of eliminating ATR beliefs across Benin to assess their role in contemporary forest conservation.
• To validate the theoretical model by comparing predicted and observed spatial forest cover distributions using SDE solutions and joint density estimation.

Proposed method

• Develops a continuous-time mean-field model of non-market interactions where individual forest extraction depends on personal ATR adherence and population-level belief distribution.
• Models forest cover evolution as a diffusion process driven by a stochastic differential equation (SDE), with equilibrium extraction policy derived from utility maximization under belief heterogeneity.
• Incorporates resource scarcity costs that depend on the distribution of forest cover, introducing spillover effects between individual and aggregate decisions.
• Uses a joint lognormal-Gamma distribution to model the spatial distribution of forest cover and ATR belief levels, estimated via generalized method of moments (GMM).
• Employs instrumental variables (IV) with proximity to the Benin-Nigerian border as instrument to identify causal effects of ATR on forest cover change.
• Applies a spatial regression discontinuity design using historical boundaries of the Kingdom of Dahomey to assess long-term impacts of ATR affiliation on forest cover.

Experimental results

Research questions

• RQ1 Does individual adherence to African Traditional Religions (ATR) lead to reduced forest extraction due to sacred ecological beliefs?
• RQ2 How does the distribution of ATR beliefs across a population affect collective forest conservation outcomes?
• RQ3 Is there a causal relationship between ATR adherence and observed changes in forest cover in Benin?
• RQ4 What would be the counterfactual forest cover distribution if all individuals in Benin uniformly lacked ATR adherence?
• RQ5 To what extent do historical affiliations with the Kingdom of Dahomey, the birthplace of Vodun, predict current forest cover patterns?

Key findings

• A 1 standard deviation increase in ATR adherence is associated with a 0.4 standard deviation positive impact on forest cover change, indicating a robust causal relationship.
• The model-predicted forest cover distribution closely matches the observed spatial distribution, with residual density between observed and predicted values centered near zero (mean: -0.208).
• The counterfactual scenario—where all ATR adherence is set to zero—results in a first-order dominant forest cover distribution, indicating systematic forest loss.
• On average, removing ATR beliefs uniformly across Benin would lead to a 6.9% loss in forest cover, ranging from 0.1% to 20.57% across grid cells.
• The spatial regression discontinuity design using the historical boundary of the Kingdom of Dahomey provides positive evidence that Dahomey affiliation is associated with better contemporary forest conservation.
• The joint spatial density of ATR beliefs and forest cover, as predicted by the model, fits the empirical data well, as confirmed by Q-Q plots and theoretical density comparisons.