Public opinion plays an important role in environmental politics of democratic countries. Legislators, fearing an electoral backlash, often hesitate to support bills that impose direct and visible costs on consumers. Environmental economists have dreams of carbon taxes to mitigate climate change, but elected officials worry about losing the next election if they vote for such policies.

Understanding the relationship between public opinion and legislators’ choices requires good data on public opinion and votes taken in the legislature. We have compiled and made freely available a comprehensive dataset of (i) important environmental roll-call votes in the U.S. Congress, (ii) state-level environmental public opinion, (iii) and campaign contributions by the oil/gas industry.

Consider an example. In a forthcoming RPR article, we used the data to examine patterns of partisan polarization in the U.S. Congress. Using a regression discontinuity analysis, we found that the effect of electing a Democrat in close elections over a Republican on the probability of a pro-environment vote is over 40 percentage points, a massive difference. We also found that the gap is the widest when fossil fuel interests strongly support the Republican candidate and public opinion itself is polarized over the environment.

The dataset, which we use for an analysis of partisan polarization over the environment in the U.S. Congress, enables scholars of environmental politics to examine the role of public opinion, interest groups, and electoral competition in environmental policy. For example, the dataset allows anyone to identify a given legislator’s environmental roll-call votes and link them to conservative, liberal, and moderate public opinion is his/her state. The researcher can find both the legislator’s margin of victory in different elections and campaign contributions from the powerful fossil fuel industry.

With this information, scholars of environmental politics can now conduct rigorous tests of hypotheses from sophisticated theories of environmental politics. With over three hundred thousand individual roll-call votes across a spectrum of environmental issues by thousands of legislators between 1971-2013, our dataset offers lots of statistical power and thus allows researchers to test contingent hypotheses on the interactions between different factors.

The dataset is freely available for non-commercial use, provided you cite the following:

Kim, Sung Eun and Johannes Urpelainen. 2017. Roll Call Votes on Environmental Issues by the U.S. Congress, 1971-2013. Harvard Dataverse, V1.

im, Sung Eun and Johannes Urpelainen. 2017. “The Polarization of American Environmental Policy: A Regression Discontinuity Analysis of Senate and House Votes, 1971-2013” Forthcoming. Review of Policy Research. DOI: 10.1111/ropr.12238

The codebook contains information on our sources, so that you can also locate and cite the original dataset for any particular variable. We are much obliged to the scholars and practitioners who have made their data available for anyone to use. That’s the future of social science.

We hope that the dataset encourages graduate students and junior faculty to do theoretically ambitious and empirically rigorous work on environmental politics. The environment is an inherently political subject fraught with conflict, so we badly need cutting-edge political science to inform environmental policy design. At the same time, political science as a field can reap enormous benefits from the rigorous analysis of high-quality data on environmental issues.


Note: This is a joint post with Eugenie Dugoua, a Ph.D. student in Sustainable Development in the School of International and Public Affairs at Columbia University.

There is no doubt that India has made rapid progress in rural electrification in recent years. Comparing the 2001 and 2011 Censuses of India, the household electrification rate in rural areas increased from 43.5% to 55.3% despite rapid population growth at the same time. Since then, progress in rural electrification has continued thanks to the government’s flagship electrification programs. By now, almost all villages in India are electrified.

And yet, hundreds of millions of households remain without electricity. Understanding the factors that contribute to low electrification rates in some areas is thus a necessary input to any policy framework that aims to universalize energy access within India. Why do some communities have stubbornly low electrification rates, and why do some households within electrified communities remain without domestic electricity?

In a recent paper (Energy Policy, forthcoming), we examined patterns of rural electrification using data from the 714 villages covered in the freely available ACCESS dataset. With the ACCESS survey conducted in 2014-2015, we matched all 714 villages to the 2011 Census data to assess changes in household electrification rates. For all villages in the dataset, we thus had information on changes in the household electrification rate between 2011 and 2014-2015.

Our analysis shows that geographic inequalities in household electrification have decreased significantly. While distance to the nearest town and the geographic area of the village are robust predictors of low electrification rates in the 2011 Census, they no longer predict such rates in the ACCESS survey. India’s rural electrification programs have been successful in extending the grid to remote households in sparsely populated areas.

Unfortunately, other inequities remain. Both household income and caste composition remain predictors of electrification. Villages populated by poor and/or low-caste households have lower electrification rates both in the 2011 Census and the more recent ACCESS survey. The same pattern holds if we look at differences between households within electrified villages.

To the extent that India’s rural electrification program seeks to universalize access to modern energy, these results suggest that finding solutions to social, as opposed to geographic, inequities should be the government’s next top priority. Policies that allow poor and socially marginalized households, instead of just villages, to benefit from household electrification are the next step on the path to offering every Indian affordable and abundant modern energy.

Note: This is a joint post with Sandra Baquié, a Ph.D. student in Sustainable Development in the School of International and Public Affairs at Columbia University.

One of today’s great energy crises is the continued use of traditional biomass for cooking by one-third of the world’s population. When people use firewood and other forms of biomass to cook inside their homes, they generate huge amounts of indoor air pollution, and the World Health Organization estimates that over four million deaths a year can be attributed to the damage done.

What can be done to solve the problem? Modern cooking fuels and technologies, such as efficient cookstoves, can substantially reduce indoor air pollution. The problem is that households – mostly rural – in different countries are not using these so-called “improved” technologies. If people consider modern technologies too expensive or inconvenient, they will not use them, regardless of what public health experts say.

In a recent paper coming out in Energy for Sustainable Development, we assess the relationship between modern cooking fuels and households’ subjective satisfaction with their cooking arrangement. The goal of this analysis is to test a simple but important hypothesis: households with access to modern cooking fuels are happier with their cooking than other households.

For the data analysis, we use the freely available ACCESS dataset of energy access among 8,568 rural households in six Indian states. We estimate statistical models to test whether access to modern fuels – most importantly, liquefied petroleum gas (LPG) – is associated with subjective satisfaction. In our dataset, 19% of households report using LPG for cooking.

The results are striking: LPG access is by far the most important predictor of subjective satisfaction. Households with LPG at home are much more satisfied with their cooking arrangement than other households, and the key reasons appear to be reduced smoke, faster cooking, and ease of cooking.

The very strong association that our data reveals suggests that modern cooking fuels can be very convenient and beneficial for rural households. Based on this reasoning, it seems that the cost of cooking with LPG might be a more significant obstacle to adoption than any perceived problems with the technology. There is, after all, nothing to prevent households from “stacking” biomass and LPG whenever necessary,

Our simple result is a first step, and future studies should rigorously investigate the causal impact of modern cooking fuels on subjective satisfaction, along with more objective indicators. Given the very strong association between subjective satisfaction and LPG access, there is every reason to be optimistic about the desirability of LPG as a clean cooking fuel in rural India.

The study of energy poverty is difficult because measuring the degree of access to modern energy is difficult. Wouldn’t it be awesome if there were detailed, representative, and freely available data on energy poverty?

Even in countries like India, where data on rural households is generally available, there are very few detailed and representative data sets on energy poverty. Census and National Sample Survey data are representative of the population, but they contain few details on energy access. More detailed surveys typically focus on small areas or their sampling fails to be representative.

To relax the data constraints that have impeded social science research on energy poverty, I and my collaborators are delighted to introduce ACCESS, a freely available data set on energy poverty in rural India.

The ACCESS data set contains detailed information on energy for 8,565 households from 714 villages across six states of India (Bihar, Jharkhand, Madhya Pradesh, Uttar Pradesh, Odisha, and West Bengal). The sampling is conducted such that the data are representative of rural areas in the six states, and we have also validated the representativeness with comparisons to recent household data from the National Sample Survey of India. Village identifiers from the 2011 Census of India are included.

Besides socio-economic characteristics and comprehensive modules on access to electricity and cooking fuels, the data set also contains information about subjective satisfaction and policy preferences. The data set also contains details on the use of decentralized energy technologies, such as solar power and efficient cook stoves.

The data set is described in a recent Nature Energy study (gated content) and an earlier report on rural energy access published by the Council on Energy, Environment and Water.

Besides studies of energy access, the data can be used for power analysis and as a baseline for field experiments.

Anyone can use the data for any non-commercial purposes, provided you cite the originaL NE study:

Aklin, Michaël, Chao-yo Cheng, Johannes Urpelainen, Karthik Ganesan, and Abhishek Jain. 2016. “Factors Affecting Household Satisfaction with Electricity Supply in Rural India.” Nature Energy 1, Article number: 16170. DOI: 10.1038/nenergy.2016.170. (

Energy poverty is one of today’s great global challenges and a key component of sustainable development, so we need all hands on deck. Making ACCESS freely available to anyone is but one step in the right direction.

If anyone’s interested in discussing new ideas on what to do with the data, what kind of data we should collect next, and – most importantly – how we can make progress in ending energy poverty in a sustainable fashion, please get in touch.

Tuesday night was a huge disappointment for so many of us. I personally never gave Trump a chance, and as the bad news kept coming in, I realized that I had no Plan B. My future plans as an environmentalist and a social scientist were based on the assumption that the next government would rely on academic researchers to offer advice on good policies and strategies.

I was wrong. It’s pretty clear that we all will be playing defense instead.

Over the past 48 hours, many people have shared their thoughts and ideas on the challenges that lie ahead. I am not going to comment here on what to expect under President Trump, as that issue has been covered more than thoroughly (NYT; Vox; Vox-2; Revkin; Stavins; Hale) and now we just have to wait and see.

Instead, I am going to talk about what we all can do. These thoughts are preliminary but at least I can say that this has been my obsessive focus for the past 48 hours.

There is little reason to expect President Trump to support the Paris Agreement, but international action on climate mitigation and sustainable energy remain critical. Renewable energy, fuel subsidy reform, and preventing deforestation are just some examples of approaches that can reduce greenhouse gas emissions. These approaches also promise many tangible benefits, such as reduced air pollution and improving energy access, to major emitters from China to India and Brazil. They are not critically threatened by President Trump, so those of us working on them should move ahead, full steam.

At the same time, President Trump resurrects an old question: what should global climate cooperation look like without American leadership? This question was hot in the days of the Bush administration, but the world has changed and the United States is not as pivotal as it was in, say, 2005. The Kyoto Protocol has disappeared and we now have the Paris Agreement. Practical, policy-relevant analysis of strategies is very important, and I intend to start new work on this topic pretty much immediately.

In the United States, the federal government will probably not do much in the coming four years and Obama’s signature contributions, such as the Clean Power Plan, are in grave danger. However, total paralysis across all levels is not in the cards. Exploiting opportunities for climate policy at the state level is more important than in a long time, so it’s time to ramp up research, advocacy, and policy work in the states. Many of us worked on this topic in the Bush era, and it’s time to revisit these lessons and adjust them to new realities.

The third issue that cannot be avoided under President Trump is opposition to fossil fuel projects. Recently, these projects had lost their momentum because of low energy prices and new environmental regulations. Under President Trump, mobilization against coal-fired power plants and tar sands may again become very important. Supporting organizations that engage in this kind of work, such as the Sierra Club, whether financially or by volunteering or with research, is a priority for me.

Finally, there is the question of electoral politics. Climate and energy issues played virtually no role in the campaign, so it seems that focusing on educating voters or trying to encourage climate voting is a lost cause. The reality is that one of the two political parties is committed to sustainable energy and climate mitigation – and the other is not. The implication is pretty clear: without a Democratic Congress and President, the federal government is not going to be in the high-ambition coalition.



Rural electrification efforts across the world have mostly focused on increasing the number of grid connections available to household. National flagship programs from Ghana to Brazil and India set ambitious goals of reaching universal electricity access by connecting household farther and farther away from urban centers.

The problem with this approach is that it ignores the quality of electricity supply. A grid connection is not worth much if electricity is rarely available. In the end, connecting households to the grid may do little to improve their welfare unless the generation, transmission, and distribution companies can offer high-quality electricity when the households need it.

In a recent Nature Energy study, we looked at this issue by estimating the relationship between different components of the quality of electricity supply and households’ satisfaction with their domestic lighting and electricity services. Using data from over 8,500 household surveys in six states, we examined how factors such as hours of electricity available (duration), the frequency of outages, and voltage fluctuation shaped households’ subjective satisfaction.

The results were striking: increasing hours of electricity available to an electrified household by one standard deviation would increase subjective satisfaction as much as connecting a non-electrified household. Because the average availability of electricity in the sample remained below 13 hours, increasing the duration of electricity supply furnished large benefits to the subjects.

The intuition behind this result is pretty simple. Rural households almost always use electricity for lights and mobile charging; many also use appliances such as fans and televisions. All these technologies are time-sensitive: the electricity has to be available at the time of use, unless the household is wealthy enough to invest in energy storage.

These results have some important implications for efforts to eradicate energy poverty. As household electrification rates across the world continue to increase, governments must begin to grapple with the much more difficult challenge of improving the quality of supply through power sector reforms and perhaps by encouraging high-quality off-grid electricity services.

Achieving these goals is much more challenging than connecting households, but the returns are large. Our study shows that regardless of whether household electrification produces direct economic gains, it is strongly associated with the quality of domestic life – and that’s gotta count for something.

In many countries, the low quality of electricity supply slows down economic growth. Without a continuous and reliable supply of electricity, firms face difficulties in improving their productivity and competitiveness. All key sectors of the economy – agriculture, industry, and services – benefit from electricity as an input.

In recent work published in the Journal of Policy Analysis and Management (gated; ungated pre-print), Joonseok Yang and I investigate the value of a specific power sector reform for encouraging private investment in power generation. We focus on the problem of electricity generation, asking a simple question: how can governments encourage private investors to increase generation capacity?

The results from the study suggest that allowing independent power production (IPP) is enough to produce a large increase in private investment in electricity generation. Both domestic and international investments increase, and the private increase easily compensates for any reduction in public investment in generation capacity.

These results are good news and important for policy because IPP reform is among the easiest available to governments. It does not require fundamental, politically controversial changes to the governance of the power sector. The government simply allows independent producers to generate and sell power to state-owned electric utilities based on purchase agreements.

The results are also somewhat surprising, given that governments’ inability to commit credibly to policies is a thread running through the literature on investment. But when we investigated cases of IPP reform, we found that governments had a good track record of honoring their commitments under purchase agreements.

As governments aspire to fuel their economies and reach universal electricity access, power sector reforms are bound to play a central role. While IPP reform does not solve difficult problems in the distribution sector, such as electricity theft or agricultural electricity subsidies, this policy does allow governments to rapidly expand their electricity generation capacity. In countries that face capacity constraints in the power sector, our evidence should encourage governments to pass policies that encourage independent power production.

“Negotiators from more than 170 countries on Saturday reached a legally binding accord to counter climate change by cutting the worldwide use of a powerful planet-warming chemical used in air-conditioners and refrigerators,” describes the New York Times the successful negotiations on the Kigali Amendment to the Montreal Protocol.

The Kigali Amendment is a big deal because it phases out hydrofluorocarbons (HFCs). The wealthy countries begin to phase out these powerful greenhouse gases already in 2019, most developing countries follow in 2024, and some major HFC producers such as India in 2028. Estimates suggest that the Kigali Amendment alone might reduce global warming by almost 0.5 degrees Celsius by 2100.

But the Kigali Amendment is also a small step. It does not get us any closer to solving the hideously complicated political problems related to reducing the use of fossil fuels. Because HFCs are only used in a specific, heavily concentrated sector and alternatives are readily available, the bitter conflicts related to carbon emissions do not surface.

In the Kigali Amendment, almost all countries in the world had very little to lose. Major industrialized countries benefit from a deal that phases out dirty chemicals and creates markets for cleaner substitutes – a pattern we have seen before in chemicals negotiations. Most developing countries do not produce these chemicals, so the number of countries that stand to lose from HFC phase-out is very, very small. In the negotiations, these countries – especially India – were given extra time to adjust, and a deal was closed.

The difference in the logic of HFC phase-out and carbon abatement can be seen by comparing the Kigali Amendment and the Paris Agreement. The Kigali Amendment says phase-out: the substances are to disappear from the face of the earth. The Paris Agreement says that countries can do whatever they want, but they need to submit documents for review.

This is not a bad thing. The Kigali Amendment is a good, aggressive solution to an easy but important problem. The Paris Agreement is a partial, complicated solution to a massive and massively important problem.

A few years back, I proposed that the future of climate policy is in big dreams and small wins. There is no silver bullet or master plan, given the complex global politics of climate change. But innovative, pragmatic negotiators can help governments solve a series of smaller problems that gradually reduce the rate of global warming. It probably will not be enough for 2 degrees Celsius, but every reduction of 0.01 degrees Celsius is worth celebrating.

Do democratic publics punish politicians for failing to provide public services? Sanctioning poor performance is a key means by which voters hold their government officials accountable. But when 700 million people lost power during the July 2012 blackout in India, public confidence in politicians increased, with troubling implications for dealing with abrupt failures of public service delivery in an age of climate change.

Democratic accountability relies heavily on voters’ ability to sanction and reward political leaders based on their performance. When voters assess their leaders, they can reward good outcomes such as high economic growth or high-quality public services.

Indeed, a great deal of evidence suggests that incumbents suffer at the polls for poor economic performance and failures in public service delivery. This logic of electoral backlash gives elected officials an incentive to work hard for their own political survival. (To be sure, voters also appear to punish incumbents even for events beyond their control, such as natural disasters and even shark attacks.)

However, evidence from India suggests that people may not always lose confidence in public officials in the wake of a massive failure in public service delivery. On July 30, after a circuit breaker tripped, seven northern Indian states plus Delhi completely lost power, followed by eleven more states in the northeast the following day. The July 30-31 outage was the largest blackout in human history, with 700 million people losing electricity.

Using data from the India Human Development Survey, which happened to be in the field during the July 2012 outage, we show in a working paper that the blackout caused an increase in people’s confidence in politicians. Moreover, the ruling Indian National Congress (INC) actually increased its vote share in the state elections that took place later that year.

We attribute these results to a “rally around the flag” effect: people increase support for their political leaders in times of crisis. Existing studies suggest that crises, particularly when they involve foreign enemies, tend to increase public support for politicians because of the fear and anxiety provoked by a traumatic event. This emotional reaction, in turn, leads people to seek a rapid, decisive solution to the crisis – people rally around the flag because they want their leader to solve the crisis. As a sudden, unexpected shock that threw much of the country into turmoil, the July 2012 blackout in India represented a period of national crisis from which voters sought relief from public officials and—at least in the short-term—held their anger for the failures of government energy policy at bay.

The implications of these findings are troubling. The less voters hold public officials accountable for policy failures, the less incentive politicians have to provide services effectively. If policymakers do not suffer during crises and disasters—even those caused by long-term failures of policy such as the outage—then preparedness efforts may hold less appeal than they should.

This is all the more alarming given that climate change is likely to make natural disasters more common in the coming years. Extreme weather events are going to strain public service delivery to India’s massive, growing, and poor population. If leaders in India expect a boost of confidence from abrupt failures in key services, their incentive to invest in disaster preparedness and climate adaptation is weakened. While the leaders gain, the people pay the price.

Note: This is a joint post with Brian Blankenship, a Ph.D. Candidate in Political Science at Columbia University.


Governments around the world spend vast amounts of money to subsidize the consumption of fossil fuels: oil, natural gas, and coal. In 2014, for example, consumption subsidies for fossil fuels amounted to $493 billion. These subsidies are both a waste of money and an environmental disaster, as they encourage people to burn fuels that release pollution into the atmosphere.

If fuel subsidies are really bad idea, why don’t governments simply eliminate them and use the money more productively? Political economy studies of fuel subsidies suggest three reasons: popular backlash, opposition by vested interests, and low institutional capacity for better social policies, such as cash transfers to the poor.

In a recent study, Joel E. Smith and I ask whether international organizations (IOs) could help governments remove fuel subsidies. We begin with the idea that IOs are mostly too weak to force governments do anything. If a sovereign government feels the need to hand out fuel subsidies, the typical IO cannot stop the waste.

But IOs can play a different role. If governments interested in reducing their fuel subsidies cooperate under an IO, they can make public commitments to eliminating their subsidies.

The IO can support these public commitments through mechanisms like regular peer reviews by the member states. Such peer reviews both facilitate learning and raise the reputational costs of reneging on one’s commitments to reduce fuel subsidies.

So while an IO cannot force governments to act, it can support their own efforts by making their promises more credible. The appropriate role for IOs in fuel subsidy reforms is to support governments interested in dismantling their bad policies but face political difficulty in doing so. With the help of the IO, these governments can improve their odds of weathering the political storm that fuel subsidy reductions produce.

For empirical evidence, we examine the role of two quite different IOs — Group of 20 (G20) and Asia-Pacific Economic Cooperation (APEC) — in fuel subsidy reforms. The key lesson from these studies is that the G20 has faced difficulties in committing to fuel subsidy removal largely because its membership is too diverse and fragmented to agree on policy priorities and commit to them.

Last month, I wrote in The Hill about the G20’s failure to make a clear public commitment on fuel subsidies at this year’s summit. Here, again, we saw the difficulties that the G20 faces: it is very, very hard to get 20 major powers to agree on what to do about fuel subsidies.

As countries develop their national plans to mitigate climate change under the bottom-up Paris agreement, eliminating fuel subsidies is increasingly urgent. Different IOs should be prepared to support interested governments in developing and implementing reputational commitments for fuel subsidy reform — without forgetting that fuel subsidy reforms are national and cannot be forced on reluctant leaders.