Fueling Demand: The Effect of Rebates on Household Purchase of Improved Cookstoves in Rural India

Over one half of the world’s population uses solid and biomass fuels, such as wood and crop residues, for cooking and heating (Legros, et al., 2008). Inefficient combustion from solid fuel use leads to emission of smoke, particulate matter, and black carbon and is associated with increased health risks (Bonjour et. al., 2013), local and global environmental degradation (Grieshop et. al. 2011; Ramanathan & Carmichael, 2008), and barriers to household economic development (Wilkinson et. al. 2009).

Improved cookstoves (ICS) represent a compelling option for decreasing the health, environmental, and economic costs associated with solid fuel use (GACC, 2012). ICS draw on clean energy sources or improve combustion of solid fuels, which decreases exposure to HAP in households and may limit the negative health outcomes in women and children, as well as reducing emissions that contribute to climate change (Ramanathan & Carmichael, 2008). Improved fuel efficiency suggests that smaller quantities of wood are required for cooking and heating, thus reducing each household’s time spent collecting wood and decreasing local deforestation.

However, the challenge in realizing the potential gains from ICS lies in encouraging both initial investment and sustained use of ICS technology (Jeuland & Pattanayak, 2012, Ruiz-Mercado et al., 2011). Households in rural, low-resource settings, where solid fuel use is high, are often budget constrained and have exhibited low demand for ICS and other preventative health technology (Hanna et al., 2012, Levine & Cotterman, 2012, Lewis & Pattanayak, 2012, Dupas, 2011). Previous studies suggest that the low demand for ICS may be the result of a range of barriers, including inability to pay or low willingness to pay for ICS (Levine & Cotterman, 2012), as well as a lack of understanding of ICS benefits and use (Shell Foundation, 2013), low trust in new technologies (Miller & Mobarak, 2011), and poor cultural acceptability (Tronsoco et al., 2007).

Using two rounds of survey data from Duke University’s stove sales randomized control trial in rural Uttarakhand, India, I use a household adoption framework to model a household’s decision to purchase ICS (Pattanayak & Pfaff, 2009). I specifically examine the effect of a rebate offer in incentivizing ICS purchase and additionally consider the influence of local institutional, community, and household-level factors associated with a household’s stove purchase decision.

The study’s stove sales intervention targeted key barriers to ICS adoption by incorporating 1) information, education, and communication (IEC) activities related to stove benefits and use; 2) a choice of two improved stoves, including an electric G-Coil and natural draft Greenway biomass stove; 3) an installment plan option, wherein households spread out stove payments over three visits; and, 4) a randomly assigned rebate offer, which reduced the price of the stove by one of three-levels, and was contingent upon stove use.

Sales results indicate a high demand for ICS among households offered the stove sales intervention. In the entirety of the treatment group, 51% of households purchased a stove. Of the stove types offered, demand for the electric Gcoil stove was highest, encompassing 70% of the stoves sold. Of the group that purchased a stove, 20% purchased a biomass Greenway stove and 10% purchased one of each type of stove. Following the intervention, 65% of treatment households owned any kind of improved stove, compared with 31% owning an improved stove at baseline.

The randomized rebate offer shows a positive and highly significant effect on household ICS purchase. In all models, the percentage of households purchasing stoves increased as the rebate increased (and price paid decreased). At the highest rebate level, 72% of households purchased an ICS, with 54% and 27% of households purchasing at the middle and lowest rebates, respectively. Further, average marginal effects of the rebate offer on the type of stove purchased indicate that assignment to one of the two higher rebate levels causes a household to be more likely to purchase a Gcoil electric stove over their traditional stove.

A number of community and household characteristics are significantly correlated with stove purchase, giving insight into types of households that may be more likely to adopt ICS. Examination of the role of local NGOs in a community introduces a nearly 16% increase in stove purchase suggesting the importance of understanding local institutions in ICS service delivery. Additional analyses demonstrate the influence of a household’s use of savings and credit, finding that rebate’s effect on stove purchase is significantly higher among households that lack experience with savings.

This analysis finds that there is a high demand for improved stoves, especially with substantial ‘use-related’ rebates. Deliberate experimentation with various rebates provides further understanding of price elasticities, which may guide planning and marketing. However, further focus is needed in building a reliable supply of ICS, especially given the challenging environments that small market-based approaches to ICS distribution face in developing countries. When further challenged with low ICS demand and a market distorted by subsidies, local market-based supply chains may flounder. This study’s findings suggest that NGOs may serve as an important institutional complement to market-based supply that leverages local networks of trust and contextual knowledge.