The limited impacts of an improved cooking stove programme in India suggest that testing interventions in real-world conditions is important
When you think about dangerous air pollution, the first image that comes to mind may be the smog-laden Delhi or the Beijing skyline. But indoor air pollution caused by people cooking with solid fuels inside their homes is an equally important challenge, responsible for about 4.3 million deaths worldwide each year according to the World Health Organization (WHO) — more than any other environmental cause. This is equivalent to the annual number of malaria and tuberculosis-related deaths combined.
Over 3 billion people around the world rely on solid fuels like wood, dung, and coal for cooking and heating their homes. The vast majority are low-income households in developing countries who historically have had few other affordable and reliable energy options. In India, about two thirds of the population uses these fuels for cooking (Census of India 2011). In addition to its health consequences for women and young children, cooking with biomass fuels contributes to climate change through the release of carbon dioxide (CO2) and black carbon.
Improved cooking stoves are increasingly seen as an important technology to improve respiratory health and combat climate change in developing countries. There are many different models and they are designed to reduce smoke exposure by directing smoke away from users, using less fuel, and/or putting off fewer harmful emissions. There are examples of improved cooking stove programmes dating back to the 1980s, including in India, but in recent years there has been a big push to increase their distribution and sale worldwide. For example, the Global Alliance for Clean Cookstoves calls for 100 million homes to adopt clean and efficient stoves and fuels by 2020, and estimates that 28 million have between 2010 and 2014.
Several studies and pilots in laboratory settings have shown that improved cooking stoves could have big effects on smoke exposure and emissions. Yet, like many preventive health technologies, whether they are as effective in real-world settings as they are in the lab depends on whether people want them in the first place, and whether people use them often and long enough for them to improve health. A recent randomised evaluation (Mobarak et al. 2012) in Bangladesh suggests low willingness to pay for an improved cooking stove among women, and especially among men.
Even in markets where there is demand for improved cooking stoves, for the technology to work, people must also be willing to pay the small everyday costs of not using their traditional cooking technology and using and maintaining the improved stove instead. What is the long-run impact of an improved cooking stove outside a controlled laboratory setting where its effectiveness depends on human behaviour over time?
Evaluating long-run impacts of an improved cooking stove programme in Odisha
To help shed light on this question, we partnered with the NGO Gram Vikas to conduct a randomised evaluation of the long-run health and environmental impacts of an improved cooking stove programme for 2,600 households in Odisha (Hanna et al. 2016). A public lottery determined when households received an offer for a stove. Some were randomly assigned to receive an offer in the first year, some in the second, and by the third year all households in the villages were offered a stove. We followed up with households for four years after the initial offer to measure the longer-run impact of a stove programme under realistic usage conditions.
The Gram Vikas stove had a chimney that directed smoke away from the user. It had been proven to reduce indoor air pollution and energy consumption in the lab, tests which Gram Vikas successfully replicated. It was constructed with locally sourced materials at a cost of about US$12.50. Gram Vikas subsidised the material costs so households were only responsible for providing mud for the stove base, labour, and a payment of about US$0.75 for the mason who assisted in building and maintaining the stoves.
Compared to today’s latest models, such as the Envirofit stove and recent rocket stoves, this stove is relatively old-fashioned. Yet stoves like the one we tested are similar to the vast majority of improved cooking stoves that have been distributed to date. In fact, the World Bank reports that stove programmes have typically distributed improved stoves in this category and over 166 million of them were in use as of 2011.
While many households in our study accepted the stoves after the offer, they did not use them very regularly — and use declined further over time. Many households continued to use their old stoves alongside the new. At peak use, households in the treatment group (those that were offered stoves) only cooked 3.5 more meals a week on the improved stove relative to comparison households (those who were not offered stoves). By year three, this had fallen to 1.8 meals.
Importantly, the stoves did not substantially reduce exposure to harmful pollutants. While smoke inhalation decreased among primary cooks in households in the first year (measured as carbon monoxide concentration in exhaled breath), this effect was statistically indistinguishable from zero by the second year and in all subsequent years. If use had been ‘perfect’, we may have got closer to the reductions in exposure seen in the lab.
Perhaps due to the low level of use and limited effects on smoke exposure, we saw no evidence that the stoves improved health. Nor did we see any effect on fuel costs or time spent cooking relative to the comparison group. If anything, households in the treatment group spent substantially more time repairing these new stoves relative to the traditional stoves, reducing time that could be spent on other activities.
Laboratory testing versus real-word conditions
Our study shows that a relatively inexpensive stove, used under real-world conditions, had limited long-run impacts — despite the fact that millions of these kinds of stoves have been distributed and that there was a lot of faith that they would help solve the challenging indoor air pollution problem. From a policy standpoint, this experiment suggests that we cannot rely solely on laboratory tests to measure the impacts of a new technology.
Rather, we need to go beyond laboratory testing and also test out new technologies in realistic settings where people’s use is not monitored by the research team — to understand how human behaviour may affect the way that individuals benefit from the technology. And, we need to run the experiments long enough to understand how use changes over time, not just measuring what is happening at the start when people may be particularly excited about a new technology.
Indoor air pollution is a serious problem. Improved cooking stoves may be one possible solution. But, as new technologies are being developed and tested in the laboratory, it is imperative that a couple of representative examples be tested in real-world settings before they are rolled out on a massive scale. Otherwise, in the future, we may be in the same place we are now — a field of broken stoves, millions of rupees spent, without solving the very serious problem at hand.
Editors' note: This column first appeared on Ideas for India.
References
Mobarak, A M, P Dwivedi, R Bailis, L Hildemann and G Miller (2012), “Low Demand for Nontraditional Cookstove Technologies”, Proceedings of the National Academy of Sciences 109(27): 10815-20 (summary at https://www.povertyactionlab.org/evaluation/demand-nontraditional-cookstoves-bangladesh)
Hanna, R, E Duflo, and M Greenstone (2016), “Up in Smoke: The Influence of Household Behavior on the Long-Run Impact of Improved Cooking Stoves”, American Economic Journal: Economic Policy 8(1): 80–114 (summary at https://www.povertyactionlab.org/evaluation/cooking-stoves-indoor-air-pollution-and-respiratory-health-india).
World Bank (2011), Household Cookstoves, Environment, Health, and Climate Change: A New Look at an Old Problem, Washington, DC: International Bank for Reconstruction and Development, p. 4.