Sumit Agarwal, Pulak Ghosh, Francesco Scarazzato (p), Sofie Waltl

Mathias Weidinger

Do extreme weather events adversely affect educational outcomes of kids in India? To address this question, we link school records across India with information on local weather conditions with a special focus on extreme heat. Both, cumulative heat exposure and exposure to higher temperatures during examinations adversely affect education and students’ performance. Such effects are likely amplified for kids attending school with low-quality infrastructure.
By making use of the address of schools in combination with fine-scaled weather information over four schools years, we thus can track the success of students over several years keeping variation other than weather-impacts constant.
Preliminary results find that a constant increase in temperature by merely 0.5°C means a drop in the number of students passing the exam by 1% and a drop in the number of highest grades (“distinctions”) of almost 4%. The effect on the probability of passing the exam is increasingly negative for higher temperature brackets, and the effect is largest for days with maximum temperature above 40°C. Moreover, we exploit canopy-height and deforestation data to provide preliminary evidence that vegetation in the proximity of the schools has a mitigating effect that increases with canopy height and forest cover density.

Lukas Vashold (p), Nikolas Kuschnig

Francesco Scarazzato

Demand for agricultural products is a major driver of deforestation in the Brazilian Amazon and tropical rainforests beyond. However, it is well known that newly deforested land is relatively unproductive, and agricultural products are barred from the predominant agriculture supply chains. An alternative, intermediate channel for deforestation in the context of agricultural land is the appropriation of unclaimed public land, where forest is cut and agricultural activity feigned to justify claims of ownership. These channels are entangled, yet require distinct policy responses to reign in deforestation rates. In this paper, we propose an approach to identify the deforestation impacts of expanding agricultural production, differentiating it from other channels with different implications for economic and environmental policy. We use a shift-share design, exploiting international changes in beef consumption, to identify causal effects of the demand-induced agricultural expansion, and find that pasture and cattle herd expansions are major drivers of deforestation. We also find that these impacts diminished in recent years, indicating that other factors, such as land appropriation motives, have become relatively more important. Our findings suggest that agricultural intensification could help decrease land pressure, and highlight the growing deforestation impacts of an ongoing expansion of infrastructure and legalization of land claims in the Brazilian Amazon.

Ulrich Wohak (p)

Lukas Vashold

This paper presents novel evidence from a dynamic discrete choice model on entry and exit
rates for Payment for Ecosystem Services (PES) programs. (PES) have been put forward as a
promising and economically efficient instrument to reign in deforestation in tropical rainforests,
complementing traditional command-and-control measures. The rising number of PES programs
indicate that they have become an integral part of the policy to tackle (inefficient) deforestation.
In this project, I present evidence from a reduced-form model on the determinants of participation
in PES schemes. Second, I propose a model for the dynamic choice process of farmers for (i)
entry and exit to PES programs and (ii) land-use choice that allows for heterogeneity in risk
attitudes. These choices have direct consequences for deforestation. I estimate the models’
structural parameters using data from a prominent PES program in Brazil and find that the
median farmer is risk-averse. I characterize the entire distribution of risk preferences of eligible
farmers and find that this distribution has strong implications for the payment design for PES
programs. In doing so, I also recover land-use elasticities and show that both land-use choices
as well as entry and exit decisions to PES programs depend on expected opportunity costs
as well as the risks associated with each particular choice. By the time the conference takes
place, I hope to be able to present results form several counterfactual exercises. For example,
the estimated model allows me to characterize optimal payments schemes to achieve critical
participation rates for a given ecosystem such that ecosystem service provision is guaranteed.
I also plan to quantify the costs of information asymmetry between farmers and PES program
designers and subsequently present evidence on the costs of this inefficiency.

Mathias Weidinger (p), Stafford Nichols, Stephan Dietrich

Ulrich Wohak

A large literature documents the positive relationship between forest ecosystem services and socioeconomic well-being in forest-proximate communities, in particular in lower income contexts. How deforestation affects these synergies, however, is less well understood. While aggregate forest loss metrics over any given administrative subdivision or area are widely used in social science research, measuring exposure to deforestation from an individual's or household's perspective is not commonplace. We combine geo-referenced data from a nationally representative survey in 34 sub-Saharan African countries with spatially explicit land use and land cover data at 10 by 10-meter resolution. For every survey location in our sample, we compose locally centred metrics of forest cover change. For robustness, we vary the detection sensitivity, circle radius, and recall period length of our exposure metrics to test how each of these parameters influence the final product, and we compare our local exposure metrics to traditional deforestation metrics. We conclude by envisioning applications for which localised exposure metrics might prove relevant in future research.