Conference Agenda

This paper studies the long-run environmental consequences of India’s Green Revolution by linking agricultural intensification to air pollution. I show that the adoption of high-yield variety (HYV) seeds in rice and wheat shifted production in North India from traditional monocropping toward an intensive rice–wheat system, increasing pressure on farmers to clear fields between crop cycles and raising incentives for crop residue (stubble) burning.

To quantify intensification, I construct a Rice–Wheat Intensity (RWI) measure that captures both the combined importance of rice and wheat in a district’s agricultural output and the balance between the two crops. Using a district-level panel from 1966–2011 in a multi-way fixed effects framework, I find that a 10 percentage point increase in rice–wheat HYV adoption increased RWI by 1.1 percentage points within districts. Household-level evidence shows that HYV-adopting households in exposed states are substantially more likely to follow an exclusive rice–wheat cropping pattern by about 13 percentage points and to engage in residue burning by about 12.5 percentage points.

I then link agricultural intensification to air pollution using a wind-based spatial spillover design. Exploiting variation in wind-driven exposure across origin–receiver district pairs from 2006 to 2019, I show that stubble burning and higher rice--wheat intensity in upwind districts significantly increase downwind PM2.5 concentrations, while intensification in non–rice--wheat crops reduces downwind pollution. I also find that pollution transported from other districts accounts for approximately 8.6% of total pollution in the Delhi NCR during the stubble-burning season. This paper provides a unique link between the Green Revolution and its effects on air pollution, and also addresses a long-standing question regarding the contribution of stubble burning to pollution in Delhi.

Global forest cover was in steady decline for millennia prior to modern industrialization. This paper studies how technological progress in agriculture shaped long-run forest cover in pre-industrial societies. I begin by developing an equilibrium bio-economic model to analyze how agricultural productivity influenced land allocation decisions and forest cover in the presence of limited trade and Malthusian population dynamics. I show that while early agricultural innovations may have spared forests, beyond a threshold further advances necessarily expand cultivation and accelerate deforestation. I test these predictions using high-resolution Holocene forest reconstructions from fossil pollen data. Exploiting variation in the timing of the First Agricultural Revolution, I find that early adoption led to a sustained 25% per-millennium decline in forest cover relative to later adopters. In a complementary case study focusing on the introduction of the three-field rotation system in medieval Europe, I show that the system caused an immediate 20% reduction in forest cover followed by a long-run decline of 50%. Together, these results indicate that agricultural innovation was an important driver of global deforestation in the pre-industrial era.

Time use is a critical yet undermeasured dimension of individual and household wellbeing, especially in settings where access to natural resources is highly time intensive. In this paper, we assess and compare the estimates of time use captured from (i) stylized time use questions, (ii) traditional time diaries, and (iii) hybrid time diaries for primary cooks in rural households across Kenya, Tanzania, Zambia, and Malawi. Using data from an experimental study, we analyze discrepancies in time-use estimates generated by each time-use elicitation method for different activities: fuel collection, cooking, household chores, unpaid work, and paid work. Results indicate that, compared to traditional time diaries, stylized questions, and hybrid time diaries yield higher estimates for certain activities. Importantly, the hybrid method generated the least bias. These findings highlight that choices in time-use measurement have implications for the impact assessment of development and environmental interventions that aim to reduce or reallocate time burdens. They underscore the need for careful selection of time-use instruments when designing, targeting, and evaluating such policy interventions.

Environmental sustainability efforts in many developing countries are constrained by limited fiscal capacity, making external environmental assistance an important instrument for advancing global commitments under SDGs 12, 13 and 17. However, evidence on whether such assistance improves environmental outcomes and how its effectiveness interacts with domestic fiscal conditions remains limited. This study examines the environmental impact of aid in Africa’s ten largest recipient countries and assesses whether its effectiveness depends on domestic fiscal capacity. Using balanced panel data from 2002 to 2022 and applying fixed effects (FE), instrumental variable fixed effects (IV-FE) and instrumental variable generalised methods of moments (IV-GMM) estimators, the analysis shows that environmental aid is, on average, negatively associated with environmental quality but significantly enhances domestic revenue mobilisation, indicating a fiscal crowd-in effect. The results further reveal that the environmental effectiveness of aid is conditioned by domestic fiscal capacity, with higher income levels not necessarily strengthening revenue mobilisation, and by trade openness, which generates environmental and fiscal pressures. These findings highlight the need to integrate environmental aid into national budgeting systems, strengthen domestic fiscal capacity and coordinate aid with trade and industrial policies to support cleaner production, stronger environmental regulation and green industrial upgrading.