Published on Let's Talk Development

Do households use improved cook stoves? What are the benefits? An Ethiopian case study

This page in:

About 40% of the human population, or about 2.8 billion people, find commercial fuels like electricity and gas inaccessible, too expensive or too irregularly supplied to use for cooking and heating (Smith et al., 2013; IEA, 2012).  Instead, they rely on solid fuels like coal, fuelwood, dung and charcoal that are combusted inside their homes (Jeuland and Pattanayak, 2012, Grieshop et al., 2011, Smith et al., 2013). Biomass fuels in particular are often self-collected and easy to use in inexpensive traditional stoves. This leads to severe public health problems, especially for women and children exposed to indoor smoke, and also can lead to forest degradation. Without major policy and/or technology changes, the global number of people depending on such fuels is projected to remain very large at least through 2030 (IEA, 2006, IEA and World Bank 2015).  
Improved biomass cookstoves that use less fuel and burn fuel more fully often are recommended as relatively affordable ways to deal with these concerns.

There is continued debate as to whether improved biomass cookstoves are “clean enough” to deliver the desired public health benefits (e.g. see Jeuland et al., 2015). A more fundamental issue, however, is that improved biomass cookstoves can only improve the wellbeing of households and the environment if they actually reduce fuelwood consumption and indoor smoke, and if people are willing to substitute them for more traditional cooking methods. If, for example, the stoves are not used because users find them inconvenient or strange, the technical performance characteristics of the stoves are irrelevant. 


This project focuses on measuring improved stove use in the field, and the fuelwood savings derived from that use, based on an experiment conducted in Ethiopia – a country with a strong commitment to expand use of improved stoves. The experiment is conducted using the so-called Mirt stove, a locally made concrete stove which can be used to bake injera, the main staple bread consumed in much of Ethiopia.  Reducing fuelwood to cook injera is important, because it represents the end-use for approximately half of all primary energy consumed in the country (Bizzarri, 2010). The Mirt stove has been found to use 50% less wood in laboratory tests (GIZ, 2011), and 40% to 50% based on user surveys (Megen Power, 2008; Dresen et al., 2014). It costs about $12, has a lifetime of 5 – 10 years, and has been promoted in Ethiopia since 1998 by the German technical assistance agency GIZ. 

Data and Methods
The project uses a randomized field experiment related to distribution of the Mirt stove. A random sample of 360 households received the stove as part of the experiment. A corresponding randomly selected comparison group of 144 households did not receive the stove. The households are in Amhara, Oromia, and Southern Nations, Nationalities and Peoples (SNNP) Regional States.  Survey data on socioeconomic characteristics were collected from all these households. 

We measured stove use in the field by the 360 households over a period of one year, using electronic stove use monitors affixed to the stoves. These monitors measure the surface temperature of the stove every 10 minutes for 60 days. The frequency and duration of stove use can be estimated from these temperature records.  Data were collected four times from the monitors over the course of the year.

Households receiving the stove were randomly allocated to different subgroups. One subgroup received the stove for free; another was offered the stove for a deeply subsidized price, about 13% of full cost; and a third was rewarded for use of the stove (as indicated by the electronic monitors). In addition, some randomly selected households received their stoves as a group, with training provided and subsequent opportunities to “network” about its use.

Potential fuelwood savings were obtained from three rounds of controlled cooking tests (CCTs) that compared the fuel use of the Mirt stove with a traditional three-stone tripod stove. Participants in the CCTs included 108 randomly selected households who received Mirt cookstoves in the experiment.

Key Findings
Over the course of the experiment, no households offered the stove either refused it or abandoned it subsequent to receiving it. Households also made significant use of the stove. The average rate of use was 2.4 times per week in the final round of monitoring, which is very consistent with traditional patterns of injera baking in areas without refrigeration. We also find that there is a general increase in frequency of use over time, though there is a slight falling-off in the last period of stove use measurement (Table 1). Households living in SNNP Regional State on average use the stoves about one-third less often than households in Amhara and Oromiya, illustrating the importance of accounting for local context.

Those that received the stove for free (i.e. full subsidy) made the most use of it. This contrasts with concerns that have been expressed in some experimental studies as to whether freely provided goods are really valued by recipients. Somewhat surprisingly, the rate of use was lower (though still significant) for those that paid the subsidized price. The rate of use was also high within the subgroup that was offered incentives to use the stove in the first period, but the bump from providing the reward for use declines over time so that ultimately, this approach was no better at promoting stove use than free distribution. Financial incentives tied to stove use therefore do not seem to be needed. In addition, the results suggest that networking among users can increase stove use frequency.  

As shown in Table 2, based on the CCTs the Mirt stove does save substantial fuel, though the savings are well below the 50% savings claimed on the basis of laboratory tests.  Average fuel savings increase over time as cooks gain experience with the stove.

As shown in Table 3, the Mirt stoves initially required substantially more cooking time, on average. However, the second and third CCTs showed that as cooks gained experience with the Mirt stove and improved their skills, cooking times fell. In the third CCT, the Mirt stove required marginally less time on average, compared to the three-stove tripod.


In a satisfaction survey carried out within the project, virtually 100% of users indicated that the stove is “good” or “very good” (the highest level) on all aspects, including convenience, durability, cooking time, firewood use and smoke reduction. These findings should be viewed with some caution because they are based on reported rather than experimental data on stove performance (other than cooking time), and respondents may want to give “correct” answers to project staff members. Moreover, we do not have experimental data on willingness to adopt the Mirt stove if offered at full price. Nevertheless, the correspondence between reported and experimental results indicates the Mirt stove may be a very successful intervention.

Many researchers have found that suitability is the most important reason cooks use rather than abandon improved biomass cooking technologies (e.g. see Mobarak et al., 2012). This experiment indicates that households offered the Mirt stove readily adopted it and made regular use of it. Doing so significantly reduced fuelwood use – thus also reducing time spent gathering fuelwood. In addition, the Mirt stove took no more cooking time than the traditional three-stone tripod once users gained experience with it, and users reported high levels of satisfaction with all aspects of the stove. 

These results correspond with other recent experimental research conducted in Africa (e.g. Bensch and Peters, 2013 in Senegal), but contrast markedly with recent findings from Orissa, India reported in Hanna et al. (2016).  Those authors find that even after an alternative stove was provided, almost all meals were still cooked on traditional stoves and there was little evidence of fuelwood savings. Our methods are very different from Hanna et al. (2016), which relies on self-reported data rather than CCTs and real-time in situ stove temperature monitors. The Mirt stove is also specialized for injera cooking, which likely makes for fewer issues of acceptability than the general-use stove promoted in Orissa.

So far the project has not collected information on possible respiratory benefits from the Mirt stove. It is possible that after as much as three years of regular use, Mirt stove users and others who are regularly in cooking areas during injera baking with the Mirt stove may enjoy better indoor air quality and have better lung function than corresponding households who have not used the Mirt stove. The project currently is measuring indoor air quality (in terms of PM2.5 concentrations), and the lung function of cooks and children (aged 7 and older) in households in the experiment with and without a history of Mirt stove use, to address these issues.


Michael Toman

Research Manager, Sustainability and Infrastructure, Development Research Group

Join the Conversation

The content of this field is kept private and will not be shown publicly
Remaining characters: 1000