About

Subhrendu K. Pattanayak is the Oak Foundation Distinguished Professor of Environmental and Energy Policy at the Sanford School of Public Policy. He also holds faculty positions in the Department of Economics and the Division of Environmental Social Systems at Duke University. Additionally, he is an affiliate of the Duke Center for International Development and a Faculty Research Scholar of DuPRI's Population Research Center. His work involves exploring innovative solutions to environmental and energy challenges, with a focus on improving lives in developing regions. For example, he has contributed to discussions on how upgrading stoves in rural Kenya can simultaneously enhance living conditions and combat climate change, demonstrating his commitment to research that addresses global environmental issues and sustainable development.

Research topics

• Economics
• Engineering
• Economic growth
• Mechanical engineering
• Agricultural economics
• Econometrics
• Natural resource economics
• Business
• Environmental economics

Selected publications

• A stress test for unobserved confounders

  Nature Ecology & Evolution · 2026-05-07

  articleSenior author
  PublisherDOI

• Analysis of Irrigation Technologies and Their Impacts on Labor Allocation, Women’s Time Use, and Women’s Empowerment: Evidence from Ethiopia

  AgEcon Search (University of Minnesota, USA) · 2026-01-01

  otherOpen access

  A growing body of literature suggests that irrigation can enhance women’s empowerment, yet empirical evidence remains mixed and often treats irrigation as a homogeneous intervention. This study examines the impacts of distinct small-scale irrigation technologies on family labor demand, women’s time allocation, and women’s empowerment outcomes in rural Ethiopia. Drawing on primary panel data from 1,806 households and 6,774 plots across four regions—supplemented by interviews with spouses to capture intra-household dynamics—we estimate treatment effects using a doubly robust Inverse Probability Weighted Regression Adjustment (IPWRA) framework to address potential selection bias. Our results indicate that irrigated plots require substantially more family labor-days than rainfed systems, driven primarily by intensified cultivation rather than irrigation activities per se. Diesel motor pumps emerge as the most labor-intensive technology, followed by gravity-fed systems. Irrigation adoption also increases women’s unpaid workloads, particularly under capital-intensive technologies. Empowerment effects are heterogeneous. While irrigation expands women’s participation in groups, it is associated with reductions in women’s decision-making authority, asset control, and overall empowerment under diesel and gravity systems. Nonetheless, manual irrigation systems are linked to positive effects on women’s decision-making and empowerment outcomes. These findings demonstrate that the gendered impacts of irrigation are highly technology-specific. Irrigation expansion does not necessarily result in women’s empowerment; rather, outcomes depend on how different technologies interact with existing household power structures. Gender-responsive irrigation policies are therefore essential to promote more equitable development outcomes.

  PublisherDOI

• The Power to Aspire Aspirations and Women's Empowerment in Rural Myanmar

  SSRN Electronic Journal · 2026-01-01

  preprintOpen accessSenior author
  PublisherDOI

• Does the payment vehicle matter for valuing improved electricity reliability? A discrete choice experiment in Ethiopia

  Utilities Policy · 2025-02-01

  articleOpen accessSenior author

  Frequent and prolonged power outages severely impede business operations in many developing countries. Given resource constraints, estimating the value of improved electricity reliability in such contexts is crucial for justifying related investments. This analysis uses a split-sample design to examine whether business enterprises in Addis Ababa, Ethiopia, have different valuations for improved power supply reliability under two payment vehicles (electricity bill increases and tax revenue allocation). Results show that these businesses are willing to pay (WTP) an average of US$33 per year for a 1-h monthly reduction in outages and US$24 per year for one less outage per month. These amounts represent approximately 11% and 8% of the typical annual electricity bill of 10,615 Birr (US$295), respectively, highlighting that businesses place substantial value on electricity reliability. We find no significant differences in preferences or WTP estimates between the bill and tax payment vehicle sub-samples, suggesting that tax payment vehicles are as credible as bill increases in stated preference studies and that multiple mechanisms for financing power reliability investments may be feasible in practice. Highlights • Used a split sample design to elicit WTP for improved electricity reliability in Ethiopia. • Business enterprises are WTP more for a better quality of electricity supply. • No significant difference between bill and tax payment vehicles.

  PublisherDOI

• Building the evidence base for REDD+: Study design and methods for evaluating the impacts of conservation interventions on local well-being

  UNC Libraries · 2025-04-10

  articleOpen access
  PublisherDOI

• Electrifying Empowerment: Women Role Models and Solar Electrification in Rural Myanmar

  SSRN Electronic Journal · 2025-01-01

  preprintOpen access
  PublisherDOI

• Cost-benefit analysis of springs revival in the Indian Himalayan Region

  2025-08-27

  reportOpen accessSenior author

  Natural springs are the lifelines of the Hindu Kush Himalaya (HKH), sustaining over 200 million people, agriculture, livestock, and wildlife, while buffering fragile ecosystems against climate shocks. Yet, nearly half of these springs have dried or become seasonal due to geology, land-use change, and climate change. A group of economists from South Asian Network for Development and Environmental Economics (SANDEE) conducted a cost-benefit analysis of springs revival activities in the Indian Himalaya. Their findings suggests that the benefits of spring revival far exceed the costs—even before ecological gains are considered. Therefore, investing in restoring these springs is not only urgent but also offers exceptional returns, making spring revival one of the most powerful strategies to safeguard water security, livelihoods, and ecosystems in the HKH region.

  PublisherOA PDFDOI

• Grassroots Guardians: How Collective Institutions Protect Public Goods

  SSRN Electronic Journal · 2025-01-01

  preprintOpen accessSenior author
  PublisherDOI

• We Need the Largest Experiments on Earth to Achieve Our Climate Targets

  Global Change Biology · 2025-10-01

  articleOpen access

  Natural Climate Solutions (NCS) are the climate mitigation sector with the largest potential, yet the least investment. A major barrier to scaling NCS is lack of evidence-based scientific learning at the scale of actual interventions. To overcome this gap, we propose the world’s largest field experiments, supported by a global data-sharing platform and guided by a clear framework of ethical principles and guidelines for delivering robust NCS evidence. This involves consistently reporting outcomes in units of MgCO2e $-1 yr-1 while describing direct and enabling actor groups and their financial, policy, and information interactions. Decarbonizing energy and industry is essential but alone will not achieve global climate targets. Natural Climate Solutions (NCS), such as climate-smart forestry and agriculture, avoided deforestation, and reforestation, are described as the sector with the largest climate mitigation potential while contributing to a multitude of sustainable development goals (SDGs). Yet NCS (also reported as “AFOLU”) remains the least-financed sector (IPCC 2022). Here we posit this is linked to a lack of advanced scientific learning from actual NCS interventions clarifying how and where their theoretical potential can be scaled ethically to deliver the biggest bang for the buck while managing risks. NCS projects are bespoke, and almost none are designed for evidence-based learning. Scientific learning is critical to delivering effective climate solutions, as it is for other global challenges like pandemic vaccines, famine, and poverty. In 1948, the publication of the first randomized controlled trial (RCT) ushered in antibiotics and the era of modern medicine (Crofton and Mitchison 1948). Scientific learning delivered COVID-19 vaccines in record time, advanced the agricultural Green Revolution, and guided poverty reduction strategies around the world. This enabled the doubling of human life expectancy over the past century, tripling of crop yields since 1960, and the four-fold decline in extreme poverty since 1980. Experimental methods have long been used in the development and manufacturing of technologies, including solar panels and electric vehicles. While experiments have proven that plants sequester carbon, we lack evidence on the biophysical and socioeconomic outcomes of alternative large-scale interventions to reduce and reverse greenhouse gas emissions through NCS. That is, we need experimental approaches, and more broadly “prospective” impact evaluations, to transform NCS as they have for medicine and other fields like development economics (Duflo et al. 2007). Each of the breakthroughs above is linked to evidence-based scientific learning, coupled with a pressing need, explicitly measurable outcomes, rapid innovation, and both public and private sector investments. With NCS, the latter conditions are present, but the use of evidence-based scientific learning at the scale of interventions is lacking—for both climate and other environmental program implementation. We are aware of only a few NCS interventions that used experimental or quasi-experimental designs to determine CO2e outcome per dollar of intervention (e.g., Jack and Jayachandran 2018). Hence, we lack sufficient scientific evidence to identify, de-risk, and cost-constrain NCS interventions to deliver the speed and global scale required. If NCS are to feasibly deliver a major contribution to climate stabilization, the central question that needs innovation and evidence to answer is: “How much climate benefit do NCS interventions deliver per dollar spent, and why?” The measurable and globally comparable outcome for answering this question is cost per tonne of additional CO2e mitigation per year (MgCO2e $−1 year−1). MgCO2e refers to metric tonnes of carbon dioxide equivalents, a unit that standardizes the global warming potential of different greenhouse gases (and ideally also local biophysical effects like albedo). This index integrates two values shared by all societies (climate stability and economic wellbeing) using measurable units that circulate globally. By normalizing the climate benefit relative to the economic cost, this index allows comparisons across scales, geographies, and interventions. This index annualizes CO2e fluxes and associated costs and requires specifying both the discount rates and time periods involved. This index must differentiate “additional” fluxes from existing ecosystem fluxes associated with business-as-usual land use. Establishing NCS “additionality” has been challenging; however, ongoing improvements to carbon verification standards, such as those incorporating dynamic baselines (e.g., VM0047 and VM0045), are addressing this challenge. Even as they improve, carbon verification standards are not designed or intended to report costs or answer questions about differential intervention effectiveness, so they support but do not answer our question above. Innovation and evidence for delivering MgCO2e $−1 year−1 is needed across spatial scales at which interventions occur, from trees to nations, and across fields including ecology, silviculture, agronomy, economics, political science, psychology, sociology, and engineering. Engineering may seem antithetical to natural solutions; however, many NCS are enabled by technology. These include (i) technologies that physically protect and restore ecosystem fluxes (e.g., precision agriculture to reduce fertilizer application, low impact forestry machinery), (ii) technologies that monitor, report, and verify outcomes, and (iii) technologies for distributing incentives to stakeholders (e.g., microfinancing cell phone software linked to monitoring systems). We need two transformations to answer the question above: (1) a breadth of consistent data collection and disclosure, especially on costs, and (2) a depth of evidence to determine why and where some interventions perform better than others, based on causal impact evaluations. A breadth of evidence on MgCO2e $−1 year−1 outcomes depends on both transparency in reporting and consistency in describing NCS interventions. Existing biophysical descriptions of ~20 types of NCS (Griscom et al. 2017) are necessary but insufficient for describing interventions. NCS interventions involve biophysical, financial, political, and technological interactions among many actors. While many parameters are important to describe an NCS intervention, at minimum four types of parameters should be consistently described (Figure 1): (1) “direct actors” who directly steward lands and waters, (2) additional biophysical changes, measured in CO2e fluxes, due to changes in land and water stewardship, (3) “enabling actors” necessary for this change to be sustained, and (4) financial and other necessary interactions between and among actors for the additional CO2e fluxes to occur, distinguishing transfers from true costs. This breadth of evidence is essential but insufficient. We also need a depth of evidence from NCS interventions implemented as parallel treatments, alongside controls, to understand why differential MgCO2e $−1 year−1 outcomes emerge from different NCS interventions. In fields ranging from medicine to economics, RCTs are considered by many as the gold standard of “prospective” impact evaluation (Gertler et al. 2016). However, idealized randomized designs are not easy at large scales, and we thus need careful prioritization of the most critical questions that can be answered with ethically designed experiments. For example, one could design interventions with different contract lengths to assess trade-offs between length of interventions and uptake by smallholders. The question, the context, and the explicit theory of change should determine the study design, rather than constraining questions to fit an idealized randomized design. For some scales and some types of interventions, a variety of quasi-experimental methods are better suited to help identify solutions for climate, people, and biodiversity that work as well as possible at scale (Ferraro and Pattanayak 2006). To achieve this breadth and depth of NCS evidence at a speed and scale to match the urgency of climate change, we propose three parallel initiatives: (1) Principles and Guidelines for ethically delivering the breadth and depth of NCS evidence discussed here, (2) an NCS Data Platform for sharing MgCO2e $−1 year−1 and other outcomes crowdsourced by private and public project and program managers, and (3) NCS Acceleration Labs for launching NCS interventions as large-scale experiments. The first two will build on principles and guidelines and data sharing platforms already developed in other fields (e.g., World Medical Association 2024) and would begin by convening a broader group of scientists, practitioners, and other stakeholders including Indigenous Peoples and Local Community representatives. Aligned with the principles and guidelines and learning from evidence synthesis enabled by the data platform, the NCS Acceleration Labs would launch the largest set of experiments, and quasi-experiments, on Earth to identify and refine low-cost and low-risk NCS interventions best positioned for gigatonne scaling. While the Large Hadron Collider may be the largest engineering experiment with a footprint of about 6000 ha, the largest experiment we know of based on area and mass is ecological, occurring across 110,000 ha of the Olympic Experimental State Forest in Washington, USA (Washington Department of Natural Resources 2025). NCS projects and programs are already underway at similar and larger scales; however, they remain a series of bespoke interventions, not yet structured as landscape experiments to accelerate learning. Existing NCS projects and programs are significant yet remain tiny in the context of the > 300 billion dollar NCS investment gap (IPCC 2022) and the > 10 GtCO2e annual mitigation potential of NCS (Griscom et al. 2017). Globally coordinated NCS interventions structured as ethical experiments, at scales considerably larger than 110,000 ha, are urgently needed to deliver evidence about the relative performance—in MgCO2e $−1 year−1 and metrics of human well-being and biodiversity—across a range of NCS intervention designs and geographies. By anticipating project failure, as both unavoidable and an opportunity to learn, NCS Labs would demonstrate the virtuous cycle of rapid scientific learning that has been demonstrated in other fields. As such, and at a comparable cost to the $4 billion Large Hadron Collider, NCS Acceleration Labs would implement the largest—in hectares and mass—experiments on Earth to stabilize our biosphere. At the foundation of these advances is a shift in mindset. The chess grandmaster, Gary Kasparov, attributed his success to a “system mindset” where each game is played to better understand how to win, rather than just to win (Bahcall 2019). We are playing games with our biosphere and losing. Achieving a system mindset for NCS will require innovative experiments, financing, radical transparency, respect for indigenous and local communities, and widespread collaboration to rapidly deliver gigatonne-scaled, ecologically responsible, socially just, and cost-feasible Natural Climate Solutions. Griscom wrote the original draft following conversations with multiple co-authors. All authors reviewed and edited it to reach the final manuscript. We thank Anand Roopsind, Carlos Muñoz Brenes, and Micheal Wolosin for contributions to concepts presented here, and Alyssa Crozier for figure design work. The authors declare no conflicts of interest. Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.

  PublisherOA PDFDOI

• Critical mineral mining in the energy transition: A systematic review of environmental, social, and governance risks and opportunities

  Energy Research & Social Science · 2024-07-21 · 74 citations

  review
  PublisherDOI

Frequent coauthors

• Marc Jeuland

  Duke University

  88 shared

• Erin O. Sills

  72 shared

• Jui‐Chen Yang

  Clinical Research Institute

  44 shared

• Christine Poulos

  RTI Health Solutions

  25 shared

• Paul J. Ferraro

  Johns Hopkins University

  21 shared

• Sumeet Patil

  20 shared

• Ralf Seppelt

  Martin Luther University Halle-Wittenberg

  20 shared

• Aletta Bonn

  Friedrich Schiller University Jena

  20 shared

Education

• Ph.D., Public Policy and Management

  Carnegie Mellon University

  1999

• M.S., Public Policy and Management

  Carnegie Mellon University

  1995

• B.A., Political Science

  University of Calcutta

  1991