Dear Counterfactual Enthusiasts,
Welcome to the third edition of the bi-annual SCB-IEWG Research Briefings, where we explore the latest insights in impact evaluation in conservation and beyond. Our featured studies highlight quasi-experimental methods new to our research briefings highlighting the variety of tools that exist to evaluate a conservation intervention. This edition has been timed to arrive before the summer conference season, with the aim to spark connections and discussions at the upcoming events. Enjoy!
Measuring the social cost of conservation
Trade-offs between nature and people in Ethiopia’s protected areas demonstrate challenges in translating global conservation targets into national realities uses a quasi-experimental approach to show that the Ethiopian protected area network has reduced forest loss and agricultural expansion, and helped to maintain grasslands but that this appears to have come at the cost of local food security. Interviews with national conservation stakeholders show that these challenges are widely recognized and that they prioritize improving the effectiveness of the existing network over expansion.
The impact of women’s representation on forest cover
Female legislators and forest conservation in India studies the impact of female political representatives in India’s state legislatures on forest conservation using a sharp Regression Discontinuity Design. They use the difference in vote shares of female and male politicians in close mixed gender electoral races as the running variable. Female legislators were found to have a positive impact on forest cover.
Assessing the overestimation of forest carbon offsets
Learning lessons from over-crediting to ensure additionality in forest carbon credits brings together the results from a range of different studies applying a variety of quasi-experimental approaches to estimate the impact of avoided deforestation carbon projects. It confirms that most projects did reduce deforestation relative to a credible counterfactual, however claims were 10 times higher than the true impact. This overestimation was mostly due to projects selecting reference areas exposed to lower deforestation pressures than project sites.
INSTRUMENTAL VARIABLES TO ASSESS DEFORESTATION
Separating supply from demand
When Does Higher Credit Lower Forest Area? Amazon Contexts Change Credit’s Impacts uses instrumental variables to evaluate how rural credit affects deforestation in the Brazilian Amazon. Since there is a risk of reverse causality where credit can affect the rate of deforestation, at the same time that the rate of deforestation could change the need for credit, the authors use an instrumental variable based on a shift-share approach that separates the demand component from the supply component.
Access to electricity changes incentives
Electricity, Agricultural Productivity, and Deforestation: Evidence from Brazil uses the topographic suitability of locations for hydropower generation as an instrumental variable for rural electrification to estimate how electricity access affected deforestation across Brazilian counties from 1960-2000. By making crop cultivation more productive than cattle grazing, electrification caused farmers to shift away from land-intensive pasture and into cropping, protecting native vegetation.
DIGGING INTO COUNTERFACTUAL METHODS
Knowledge toolkit
Causal Inference for Biodiversity Conservation provides a detailed overview of the state of causal inference within conservation science. The paper explores the current limitations in conservation policies and programs such as the lack of addressing assumptions and including sensitivity analyses. It covers study designs, intervention types, drivers of biodiversity loss, metrics, and more.
A framework to identify causal pathways from land to sea
Riparian vegetation reduces coastal turbidity compares a suite of causal inference methods to quantify the effects of riparian forest cover on turbidity in Costa Rica’s Golfo Dulce. It provides an observation-based method that can be applied where detailed process-based catchment or hydrodynamic models are unfeasible or where long-term empirical datasets are unavailable.
Quantifying unobservable mechanisms
Do the Effects of Social Nudges Persist? Theory and evidence from 38 natural field experiments focuses on the persistent effects of an energy conservation intervention, it highlights a method for quantifying the mechanisms through which interventions have impacts when we cannot directly observe the mechanism variables. One can decompose a treatment effect by leveraging a causal interaction between the treatment and another causal variable that is known to shut down a mediating pathway.
UPDATE ON PREPRINTS
Since our last issue, two of the three preprints we featured have been published. You can find the published versions here:
– Best practices for moving from correlation to causation in ecological research
– Unobserved confounders cannot explain over-crediting in avoided deforestation carbon projects
📚 Final Word
Our featured articles highlight the diverse range of tools available for evaluating conservation interventions. They show that robust evaluation is achievable across many contexts, and underscore the importance of defining your research questions and identifying potential data sources to aid in the selection of the most appropriate method. Equipped with the lessons they provide, we can accelerate the adoption of rigorous causal methods in our research.
Until next time,
The SCB Impact Evaluation Working Group (IEWG)
If you have any comments or suggestions about this newsletter, please email: teamcounterfactual@gmail.com.
Do you know someone who would like to receive this newsletter? Please ask them to sign up to the SCB-IEWG listserv.
Want to contribute to the newsletter? Watch for calls for input, or email us at teamcounterfactual@gmail.com with your ideas. To suggest a recent paper, include the title, DOI, and – most importantly – a two-sentence summary of what you find compelling about it.
