From the Lab to the Field and Back

Transcription

1 October 2015 From the Lab to the Field and Back Synthesis Report Social Context, Gender, and User Needs in the Design and Promotion of Clean Stoves in Indonesia Photo: Voravate Tuntivate East Asia and Pacific A product of the EAP Gender and Energy Facility and the Clean Stove Initiative

2 Copyright 2015 The International Bank for Reconstruction and Development / The World Bank Group 1818 H Street, NW, Washington, DC USA All rights reserved First printing: 2015 Manufactured in the United States of America The findings, interpretations, and conclusions expressed in this report are entirely those of the authors and should not be attributed in any manner to the World Bank, or its affiliated organizations, or to members of its board of executive directors or the countries they represent. The World Bank does not guarantee the accuracy of the data included in this publication and accepts no responsibility whatsoever for any consequence of their use. The boundaries, colors, denominations, and other information shown on any map in this volume do not imply on the part of the World Bank Group any judgment on the legal status of any territory or the endorsement or acceptance of such boundaries.

3 i Contents Acknowledgments...iii Introduction...1 Why Promote Clean Stoves?... 1 Integrating Social and Gender Aspects in the Indonesia Clean Stove Initiative... 2 Methodology... 4 Qualitative and Quantitative Methods... 4 Development of a Social Protocol for Stove Testing... 6 Findings... 7 Phase 1: Sociocultural Context... 7 Phase 2: Assessment of Stoves and the Development of a Social Protocol for Stove Testing Conclusions and Recommendations...31 Recommendations on Stove Design...33 Recommendations on Stove Testing Recommendations on the Promotion of Clean Stoves Recommendations on Holistic Approaches to Reduce Exposure to Indoor Air Pollution Annexes 1 Additional Tables Overview: The Indonesia Clean Stove Initiative References... 47

4 ii From the Lab to the Field and Back: Synthesis Report Boxes 1 Cooking in Indonesia today Three examples of successful stoves Figures 1 Steps in the gender- and context-sensitive approach to promoting the use of clean stoves Household fuel use by levels of Income Declared frequency of each cooking task to prepare daily meals Social protocol framework Tables 1 Fuel market segments in Yogyakarta and Sumba Fuel expenditure among peri-urban households in Bantul and Sleman districts Average monthly fuel expenditure per income among peri-urban sample in Yogyakarta Cooking sequence and heat range by type of food Comparison of estimated average time needed to cook a complete meal by type of stove The top 10 features sought by cooks at sampled sites, in order of importance Focus group discussion of advantages of various stove types in Yogyakarta, December Different aims for clean cookstoves Summary table of recommendations Average fuel preparation time, by stove, May Average ignition time, by stove, May 2015 and December 2014 (in minutes) Average cooking time, by stove, May 2015 and December 2014 (in minutes) Fuel savings per stove compared with baseline, May 2015 and December Testers average rating of stoves on key criteria, May Testers average rating of stoves on key criteria, December

5 iii Acknowledgments This report summarizes survey findings gathered under a World Bank program to integrate social and gender dimensions into clean stove design and promotion. The program is a collaboration between the Social Development team in the Social, Urban, Rural and Resilience Global Practice (GPSURR) and the Energy team in the Energy and Extractives Global Practice (GDEER) in the East Asia and Pacific region. The overall work is led by Helene Carlsson Rex, and the program team members are Laurent Durix (co-task team leader), Yabei Zhang, Veronica Mendizabal Joffre, Voravate Tuntivate, Olivia Tanujaya, Crispin Pemberton-Pigott, and Cecil Cook. The World Bank team is grateful for support and contributions from the Indonesian nongovernmental organization Yayasan Dian Desa, led by Christina Aristanti, and to YDD s team, particularly Prianti Utami. This work is undertaken to support the Indonesia Clean Stove Initiative (Indonesia CSI), launched in 2012 as part of the broader East Asia and Pacific Clean Stove Initiative, which also covers China, the Lao People s Democratic Republic, and Mongolia. The report was written by Veronica Mendizabal Joffre. It was edited by Steven Kennedy. Laura Johnson designed the publication. The team thanks peer reviewers Gillian Brown, Jacqueline Devine, Sarah Keener and Vanessa Lopez, and who provided useful comments, as well as Ndiame Diop, Koffi Ekouevi, and Helle Buchhave, who provided guidance at the concept stage. The team is also appreciative of the valuable guidance provided by World Bank management throughout the process, particularly Charles Feinstein and Jan Weetjens. Finally, the team gratefully acknowledges the funding provided by the World Bank s Asia Sustainable and Alternative Energy Program (ASTAE). The outputs of the work program described in this report may be downloaded from For more information, please contact Helene Carlsson Rex, senior social development specialist at the World Bank (hcarlsson@worldbank.org), or Laurent Durix (ldurix@worldbank. org).

6 iv From the Lab to the Field and Back: Synthesis Report Photo: Voravate Tuntivate

7 1 Introduction Why Promote Clean Stoves? Approximately half of the world s population continues to cook and heat their homes using biomass fuel such as wood, dung, agricultural waste, and charcoal burned in rudimentary artisanal stoves or three-stone fires. Biomass burned in this manner releases considerable amounts of CO2 and short-term pollutants, contributing to climate change and accelerating the melting of the planet s ice and snow cover. The inefficiency of such stoves also affects local forest resources and associated ecosystems. Furthermore, the direct exposure of cooks and their families to cooking smoke and black carbon particulates increases their risk of cardiopulmonary disease. The World Health Organization has estimated that exposure to indoor air pollution causes more than four million premature deaths each year, mainly among women and children. In addition, cooking over open fires or in crude stoves, preceded by gathering the necessary fuel, is labor intensive and time consuming. As cooking and fuel gathering are generally the responsibility of women, the time and effort spend on these tasks may reinforce gender inequality and cycles of poverty. Improved cookstoves that burn biomass more efficiently have been identified as options with high potential to reduce negative health, environmental, and social impacts. However, their uptake has remained low despite development and dissemination efforts dating as far back as the 1960s. Some of the reasons for the limited success of clean stove programs include their initial small scale, unverified claims of the extent of improvement in efficiency and emissions, lack of agreement on testing protocols to assess such claims, the generalized lack of public awareness about the negative health effects of smoke exposure, and insufficient understanding of the factors affecting stove demand and use. In recent years renewed interest in the potential of clean fuels and stoves to enhance health and environmental outcomes has resulted in technical progress in the areas of emissions and efficiency, testing, business models, and financing instruments. A new generation of players has come to the fore, pushing technical improvements and competing for a bottom-of-thepyramid market of 2.5 billion people. The future looks promising, and the large-scale supply of clean stoves seems within reach. Nonetheless, substantial gaps remain in our understanding of the socioeconomic and cultural factors behind the adoption and sustained use of cookstove technologies. One of the key challenges ahead will be to decipher the black box of consumer acceptability and cultural norms affecting stove adoption and use. A failure to invest in understanding demand could limit the important gains made so far.

8 2 From the Lab to the Field and Back: Synthesis Report Integrating Social and Gender Aspects in the Indonesia Clean Stove Initiative The Indonesia Clean Stove Initiative (Indonesia CSI), a partnership between the government of Indonesia and the World Bank and part of the broader East Asia and Pacific Clean Stove Initiative, aims at accelerating access to clean cooking and heating solutions for millions of households that continue to rely on traditional biomass stoves and open fires to prepare their meals and warm their houses (box 1). Unlike initiatives where public agencies or nongovernmental organizations bear the costs of production and distribution, the Indonesia CSI is piloting a results-based financing approach under which the private sector is involved from the outset. Private sector suppliers assume investment and performance risks but receive subsidies based on their achievement of set results. Under this scheme, the creativity of the private sector can be unleashed to better target local consumer constituencies while the public sector focuses on policies, testing, and standards. It has been estimated that by 2020 the program could reach a market of about 10 million households out of the 25 million that still use biomass in Indonesia. The other key characteristic of the Indonesia CSI is the close integration of social and gender considerations into the testing of clean stoves and into efforts to promote their use. Box 1. Cooking in Indonesia today Today, approximately 40 percent of Indonesia s households about 24.5 million located mainly in rural areas still rely on traditional biomass as their primary cooking fuel. A successful conversion program ( ) made significant progress in incentivizing households to switch from kerosene to liquefied propane gas (LPG), but the program had limited impacts on biomass users. Rural and poorer households located far from the distribution network continue to depend on firewood collected from the local environment to meet most of their cooking needs, and many peri-urban households routinely use LPG and biomass contiguously. Indoor air pollution resulting from the inefficient use of traditional biomass is linked to an estimated 165,000 premature deaths annually in Indonesia, primarily among women and their young children. Despite past efforts, a market for clean biomass cookstoves has yet to develop, and progress remains limited to small one-off government distributions and sporadic NGOled projects. As part of the effort to expand access to modern energy, proactive policy is needed to help this 40 percent of the country s households gain access to clean cooking technologies.

9 Introduction 3 Figure 1. Steps in the gender- and context-sensitive approach to promoting the use of clean stoves The Challenge: Increase uptake and use of clean stoves The Solution: Experimental approaches that integrate social and gender aspects into testing and promotion of clean stoves 1. Qualitative sociocultural exploration 2. Surveys to validate qualitative findings 3. Integration of sociocultural factors into technical tests 4. Protocols for social testing of clean cookstoves 5. Insights into marketing and promotion stategies The Indonesia CSI launched its first pilot in Yogyakarta in 2014 and is working to establish a certification system and to develop the capacity of key market players. As the program deploys a market mechanism, the challenge is to work with stove designers, manufacturers, and distributors to develop products that meet the expectations of consumers at affordable prices. This in turn entails an intimate understanding of the needs and wants of potential buyers, and of the social aspects influencing purchase and adoption (figure 1). To achieve this objective the World Bank assembled a team of social scientists and stove engineers and asked them to elucidate the sociocultural and economic factors likely to affect adoption of new stoves in selected pilot areas. Chief among the teams taks were to gather market intelligence on consumer needs and preferences and to identify messages that could raise consumer awareness of the benefits of clean cookstoves. Although the field work was carried out in Indonesia, it is hoped that the recommendations coming out of the work will be broadly applicable. The work encompassed qualitative research in Sumba and Central Java, including in-depth interviews, focus group discussions, and case studies. In addition, a survey of cooking fuel consumption, biomass cookstove preferences, and cooking habits in peri-urban households outside Yogyakarta City, in Central Java, was undertaken. With the information gained during this initial sociocultural exploration, a social protocol for clean stove testing was developed, tested, and validated in various locations in Central Java. This report presents the findings of that work.

10 4 Methodology With the renewed dynamism in the cookstove sector, acceptability trials of finished products are becoming more popular. Less emphasis, however, has been given to understanding local practices and cultural factors at the stages of stove development and prototyping, although work in this area is also gaining momentum (see, for example, Simon and others 2014; Miller and Mobarak 2013; and Urmee and Gyamfi 2014). The approach used in Indonesia has been based on an understanding of the adoption of innovation through the interplay of technical and social factors. Consequently, the work was designed to understand the context of stove use, with attention to (i) existing cooking fuels and devices, (ii) how cooks use and integrate these into everyday life and routine actions, and (iii) the process by which other household members and society as a whole mediate this use. Further, the study considered that rapid changes in urbanization and economic development are affecting how people approach day-to-day tasks, including cooking. Qualitative and Quantitative Methods In its first phase, the design included a review of fuel-consumption patterns and market segmentation, as well as an analysis of cooking practices, user preferences, and social and gender relations in the context of stove and fuel use. The data-collection methods included a survey of 1,434 respondents in the peri-urban districts of Sleman and Bantul in Yogyakarta, and focus group discussions, interviews, and ethnographic research with 205 people and 15 households in urban, peri-urban, and rural communities in Yogyakarta and Sumba. Focus group discussions were organized with separate groups of men and women to obtain disaggregated views about stove preferences. The design was exploratory and did not make an a priori selection of people using a specific stove or fuel, but rather sought to discover what stoves and fuels were prevalent in the areas visited. A conscious decision was made not to meet with villages that had been the site of previous efforts to promote clean stoves, as the main aim was to obtain a snapshot of the context of stove use in real life or as is. A multidisciplinary team comprising a sociologist and gender specialist, an anthropologist, a statistician, and a stove engineer conducted the work. Team members provided complementary perspectives and concepts spanning social and technical aspects, and used a

11 Methodology 5 variety of tools to interpret the complexity of stove adoption. Qualitative and quantitative methods were used for data gathering among them focus groups and in-depth interviews, a survey, participant observation, and case studies. The methods built on one another and made it possible to obtain extensive as well as in-depth information and to triangulate findings. The work was carried out in close collaboration with Yayasan Dian Desa (YDD) in Yogyakarta and Yayasam Alem Lestari (YAL) in Sumba, local nongovernmental organizations working on the development and promotion of clean cookstoves. The partnership allowed for a deeper understanding of the context and the history of the clean stove movement in the country, and teams were able to exchange knowledge and methodologies. Data was collected in the two sites selected for the implementation of the first two pilots of the Indonesia CSI: Yogyakarta Special Region in Central Java and Sumba Island in Nusa Tenggara province. The sites represent two extremes of the rich geographical and cultural diversity of Indonesia. Yogyakarta typifies Java Island: a large population; a dynamic center of culture, production, and services; good access to markets; and higher incomes. Sumba characterizes the peripheral islands, with low population density, distinct non-javanese cultures and traditions, at the margins of trade, with few opportunities for paid employment, and subsistence economies. The energy policies applied to date in the two sites differ. Between 2007 and 2009 Yogyakarta was targeted by the national kerosene-to-lpg conversion program, which aimed to convert more than 50 million Indonesian households and micro-businesses to LPG as their primary cooking fuel. Households received a basic conversion kit that included a single-burner LPG stove and a refillable 3 kg gas cylinder, plus training on proper use of the stove. The process was accompanied by kerosene withdrawal, the establishment of LPG gas terminals and filling stations, and engagement with local governments, sales agents, and retailers. Though the conversion program was national in scope it did not cover Nusa Tenggara, Maluku, or Irian Jaya provinces. For that reason, in Sumba kerosene continues to play a role alongside firewood. In 2012 Sumba was selected to showcase conversion to renewable energy. As part of that program, the organization Hivos had installed 140 home biogas digesters by 2014.

12 6 From the Lab to the Field and Back: Synthesis Report Development of a Social Protocol for Stove Testing The second phase of the work included the development of an experimental method to assess the performance of clean stoves in their context of use. The six most promising stoves that had undergone the emissions and efficiency tests of the Indonesia CSI (a waterheating test developed by the Indonesia CSI and the Indonesian National Standard Test) were selected for assessment under field conditions. The methodology was developed and tested in December Its results were validated in a second round of testing in May The testing and validation were conducted in collaboration with the YDD, which is also in charge of conducting the CSI technical tests. Seven YYD stove operators were trained in the use of the social method. The sample selected for the first test included six rural household cooks in the southern district of Bantul. The sample was extended during the validation to six more local testers in the densely populated district of Sleman. The two testing sessions suggested that the tested stoves met some of the users expectations, but some problems with usability or compatibility with context were also observed. Key issues included the requirement of some clean stoves for small cuts of wood. Because fuel preparation at the point of use is a woman s task, a requirement for smaller cuts would be likely to increase the burden on women. The test also suggested that some of the clean stoves might require much longer ignition and operation times than the baselines available in the region, which could lead to low levels of interest in transitioning from existing options to clean stoves. The method also identified some problems between stove design and local practices. Furthermore, although some stoves were intuitive enough for the user to learn after a few trials, others appeared likely to require detailed training to be operated correctly, adding to the overall costs of promotion. Photo: Veronica Mendizabal

13 7 Findings Phase 1: Sociocultural Context Patterns of fuel consumption The review of patterns of cooking fuel consumption confirmed that the Indonesian LPG conversion program successfully transformed the fuel market toward LPG in targeted areas kerosene is now rare. The largest market for LPG is in urban areas, where adoption is practically universal. A large percentage of the peri-urban sample was also using LPG either on its own (27 percent), in combination with firewood (48 percent), and/or alongside electric rice cookers (53 percent). LPG is also slowly penetrating the rural market, although its uptake has been affected not only by competition from free firewood, but also by weak rural distribution systems, the initial costs of acquiring the stove and gas cylinder outside of the already completed conversion program, safety concerns, and food taste preferences, specially among older populations. Despite the success of the conversion program, reliance on firewood remains high among peri-urban and rural communities who use it as their primary fuel even when also using LPG. In the sample, 25 percent of the peri-urban households used firewood as their only fuel, in addition to the 48 percent that used it in combination with LPG. Figure 2, which reports household fuel use by income level, illustrates the continuing use of firewood. Figure 2. Household fuel use by income level Percent of households Quintile 1 <$ Quintile 2 $75>x>$ Quintile 3 $112>x>$ Quintile 4 $150>x$ Quintile 5 >$240 Use firewood only Use both firewood and LPG Use LPG only Levels of monthly income Source: Peri-urban survey in Sleman and Bantul districts, Yogyakarta Special Region, December 2013.

14 8 From the Lab to the Field and Back: Synthesis Report In areas not targeted for LPG conversion, such as Sumba, the better-off urban segment that has cash income and is connected to the power grid mixes fuels (kerosene, firewood, and rice cookers). Lower-income urban and peri-urban households, however, use firewood combined with small amounts of kerosene. Rural households depend entirely on firewood. In both urban and rural settings, the market for charcoal remains small and is concentrated among small-scale urban food vendors. Among urbanized households, firewood is considered the cheaper option to LPG, a belief contradicted by the study s finding that purchased firewood used in inefficient stoves can be more expensive than LPG. Market segments Fuel-consumption patterns divided the market into four distinct groups: Users of LPG exclusively (kerosene in Sumba) Users of firewood exclusively Fuel mixers Users of electric rice cookers. Table 1 and figure 2 describe the socioeconomic and demographic characteristics of each group. Besides income, age, and education, household size was also found to affect the pattern of fuel use. In Yogyakarta, smaller households (average 3.7 persons) were found to prefer a single fuel, either LPG or firewood, while larger households (average 4.2 persons) used a mixed fuel strategy, combining LPG and firewood. Table 1. Fuel market segments in Yogyakarta and Sumba Only LPG (or kerosene in Sumba) Higher average incomes, better educated, and younger than other segments. Located mainly in urban and peri-urban sites. Less common in rural areas. Fuel mixers Middle-range incomes spanning age brackets and with a relatively high level of education. The share of LPG (kerosene in Sumba) is linked to income. The less income, the lower the share of LPG. Concentrated in peri-urban areas and slowly expanding into rural communities. Only firewood Lower average incomes compared with the other groups. Cooks on average less educated and older. Located mainly in peri-urban and rural areas. Rice-cooker users Highest average income compared with all other segments. Located in urban and peri-urban. Slowly penetrating rural communities connected to the grid. Source: Findings of the Indonesia CSI Qualitative fieldwork and Survey conducted in Yogyakarta and Sumba during

15 Findings 9 In Sumba, larger households (>7) prefer firewood for all their cooking. The findings point to two different segments as potential markets for clean stoves. One is lower-income consumers who use firewood for all of their cooking; the other, middle-income consumers who combine wood with LPG (or kerosene where LPG is not available). These two markets have different characteristics, including purchasing power and varying levels of exposure to cleaner fuels, with fuel mixers more likely to be interested in cleaner and more efficient biomass options. A strategy in which products and marketing are differentiated to match each segment is recommended. Household expenditure on fuel Despite distinctive patterns of fuel use, overall household expenditure on fuel in absolute terms was found to vary only slightly among segments (table 2). But in relative terms, firewood users, who on average earn less than other consumer segments, spend a larger share of their income (4.2 percent) on fuel compared with users of a fuel mix or LPG only (2.3 percent and 2 percent, respectively). The lower income quintile was found to spend as much as 7 percent of their income on fuel. This finding suggests that LPG subsidies may be benefiting better-off households rather than poorer ones, which are more likely to use nonsubsidized firewood. However, most firewood is still collected for free. Of the survey sample, 77 percent gathered free firewood, 12 percent bought and collected, and 10 percent bought (table 3). Collectors had, on average, lower incomes than buyers, but those collecting and buying had the lowest average income, possibly owing to more limited access to land and/or firewood resources. This indicates a potential economic incentive among poor households to use more efficient stoves. Table 2. Fuel expenditure among peri-urban households in Bantul and Sleman districts Fuel user segment Average income in millions IDR (US$) Average monthly HH expenditure on fuel in IDR (US$) Monthly HH expenditure on fuel in % Firewood only 1,454 (121) 61,086 (5.0) 4.2 LPG-firewood combination 2,534 (211) 58,239 (4.8) 2.3 LPG only 3,094 (258) 62,391 (5.1) 2.0 Source: Peri-urban survey in Sleman and Bantul districts, Yogyakarta Special Region, December 2013.

16 10 From the Lab to the Field and Back: Synthesis Report Table 3. Average monthly fuel expenditure per income among peri-urban sample in Yogyakarta Percentage of sample Average income in million IDR (US$) Average monthly expenditure on firewood in IDR (US$) Percentage of income spent on firewood Collect 77 2,161 (180) 0 0 Collect and buy 12 1,809 (151) 60,000 (6) 3.4 Buy 10 2,740 (228) 100,000 (10) 3.6 Source: Peri-urban survey in Sleman and Bantul districts, Yogyakarta Special Region, December Types of firewood used In Indonesia, hard wood is preferred because it burns longer and produces charcoal. However, the essential characteristic for users is not the quality of the wood per se, but its humidity content. Wet or damp wood is harder to light, produces more smoke than dry wood, and increases overall cooking times. In Yogyakarta, as on most Indonesian islands, the rainy season lasts for approximately seven months. Households prepare for the long rains and collect large stocks of firewood in advance. Nevertheless, stocks rarely last the whole season, and cooks are forced to use wet wood. This preoccupation with the humidity content of the wood is not concentrated solely in areas with the most rainfall. In Sumba, which has a shorter rainy season (three months), wet wood was an important issue for firewood users. In addition, while purchased wood is generally the hard type, collected biomass comprises all types of burnable material, such as agricultural waste, bamboo, dry leaves and twigs, old furniture, and anything that can be ignited, including plastic, which is commonly used to start the fire. In practice, therefore, stoves are burning a wide range of organic matter, of which firewood is only a fraction. The implication is that clean stoves should be able to operate with wood of varying humidity and that potential consumers may expect them to burn other commonly used combustible materials. Firewood collection and chopping The collection of firewood is widely viewed as a burden that could be reduced through the use of more efficient stoves. Fieldwork indicates variations in how users perceive this task. In Yogyakarta, wood collection is not considered difficult, as the task is shared among men and women and integrated into other daily activities. In Sumba, school-age girls spend an average of six hours each week on this task. The few girls interviewed for the study indicated that collection is a burden, but this group has no real voice within household decision-making.

17 Photo: Veronica Mendizabal Findings 11

18 12 From the Lab to the Field and Back: Synthesis Report Chopping, by contrast, is generally considered a burden. Men s involvement is limited to chopping large logs for storage, and women do a second-level fine chopping at the time of use, if needed. In rural areas people are not doing much chopping, as traditionally firewood is used in lengths of 1 2 meters to feed three-stone fires and large mud stoves. The situation is different for peri-urban households using smaller portable stoves (known as Keren stoves). In this case, women devote more time and physical effort to chopping the wood into smaller pieces. The use of pellets or small cuts of wood has been linked to the potential of job/microenterprise creation. Given that in Indonesia the greater use of firewood is concentrated among lower-income households, such prospects may need to be carefully assessed, as it is highly likely that households would prefer to chop the wood themselves rather than purchase it, and women would end up assuming this task. Cleaner stoves requiring small cuts (such as rocket/gasifiers) may, at least initially, be better suited to sites where fine chopping is more common, such as peri-urban locations. Biomass stove prevalence The study found that Indonesian households have a large assortment of biomass and manufactured stove options. The diversity was particularly high in rural villages, where the range included biomass stoves made from various materials such as mud and brick, cement and brick, clay, or carved stone in configurations with single, double, or triple burners. Within this diversity, three biomass-burning devices stood out as the most prevalent. In Yogyakarta s rural communities, the most prevalent form is the self-built, large, stationary mud stove. Among peri-urban households in Yogyakarta, the small, portable single-burner Keren clay stove is the most common option. In Sumba the three-stone fire prevails. The stoves respond to specific contexts. The fixed mud stove, usually built and repaired by husbands, fits the large traditional Javanese kitchen. Owing to its size, it can accommodate two or three pots simultaneously. Its box design generates strong heat, and it is used for a range of cooking tasks, among them daily cooking, large festive meals, water boiling, and palm sugar production. Mud stoves also allow for the production of ash for use as fertilizer and dishwasher detergent; smoke is used to dry and conserve seeds and harvested foods, as well as to dry humid firewood. The introduction of new stoves, therefore, particularly in rural areas, will have to account for any trade-offs in benefits currently obtained from traditional stoves. Although some large mud stoves are found in peri-urban neighborhoods, the stove of choice there is the portable Keren. This stove is well suited to more densely populated areas as it fits the smaller peri-urban kitchens and is easily stored after use. Manufactured locally, Kerens are widely available in corner stores and markets. Although they are fragile, they are

19 Findings 13 also cheap to replace. In Sumba, regardless of location, the three-stone fire is the common practice, although some urban and peri-urban households combine it with a kerosene stove. The stove inventory found that all households that used biomass were using at least two stoves of the same or different type. In Yogyakarta, the average number of biomass stoves per household was 2.4. In Sumba, the use of a single three-stone fire was uncommon. They are usually paired, or else a five-stone arrangement is used. Both solutions allow two pots to be supported simultaneously. Special ritual kitchens, where large amounts of food are cooked, include configurations of seven-stones for three simultaneous pots. The common use of multiple stoves is valuable information for designers, but also for producers trying to penetrate the Indonesian market, which appears larger than the number of households using biomass. Stove costs and willingness to pay Available stoves can be classified as low and high end. Low-end stoves such as the Keren, the fixed mud-and-brick stove, and the 3- and 5-stone (all of which represent a cash cost of $0.50 $2) are among the most popular options. Nonetheless, the study found willingness to pay for high-end stoves such as LPG stoves, the improved Jolentho stove with chimney, and the locally produced Doughnut stone stove ($15 $30). In the case of LPG stoves, most of the 75 percent of the peri-urban sample using LPG had purchased the stove without subsidies that is, outside of the conversion program. Willingness to pay for the Jolentho as a productive asset was observed among palm sugar producers, which organized informal credit schemes to enable the purchases. Finally, even poor households are willing to pay for the durable and powerful carved stone stove, as it is traditionally offered as a wedding gift. Local cooking practices Indonesian cuisine is complex and varies from island and island. Nonetheless, the following widespread routine practices are observed: The composition of a complete meal includes three elements: (i) a staple (rice, corn); (ii) vegetables; and (iii) a source of protein (fish, meat, tofu, or tempeh). Households cook these three elements only once a day and reheat leftovers. In urbanized areas, where women have entered the paid labor market, the main cooking session is after work. In rural areas women usually prepare the full meal early in the morning before agricultural tasks. Indonesians employ an array of cooking modalities including boiling, steaming, simmering, deep-frying, stir-frying, and grilling (figure 3).

20 14 From the Lab to the Field and Back: Synthesis Report Figure 3. Declared frequency of each cooking task to prepare daily meals Percent Cooking Reheating Breakfast Cooking Reheating Lunch Cooking Reheating Dinner Boil water Cook rice Deep fry Make soup Stir fry Steam/boil Grill Source: Peri-urban survey in Sleman and Bantul districts, Yogyakarta Special Region, December Everyday Indonesian dishes require wide heat ranges, from high to extremely low. These need to be achieved rapidly in order for the dishes to cook properly. Rice requires one of the widest heat ranges from high to minimal. When cooking on a three-stone fire this radical temperature change is achieved by burning firewood at high flame to boil the staple in water and pulling the firewood out to allow the pot to steam on hot charcoals and the heated stones. This is one of the reasons why wood types that produce charcoal are preferred for cooking rice in Sumba. With this method, once the pot is left to steam, the cook need not pay further attention until the rice is done, very much as is done with a rice-cooker. In Java two different methods for cooking rice on biomass were observed: (i) boiling the rice until the water has almost evaporated and then transferring it to a steamer pot; (ii) boiling the rice until the water has almost evaporated and then reducing the heat to low to create and maintain steam. The second method is less popular as it requires more attention to keep the flame low enough to steam the rice without burning it. Table 4 presents the cooking Table 4. Cooking sequence and heat range by type of food Type of food Dish Sequence Heat range Rice Steamed rice Boil, steam High flame, burning coals Fish Soup Boil, simmer High flame, low flame Vegetable Stir-fry Stir-fry High flame Source: Focus group discussions and participant observation during cooking episodes in Yogyakarta and Sumba, December 2012 and December 2013.

21 Findings 15 Table 5. Comparison of estimated average time needed to cook a complete meal by type of stove Type of stove Time needed in minutes Difference from 3-stone (minutes) 3- or 5-stone 87 0 Traditional biomass stove 79 8 LPG Source: Participant observation, Sumba December 2013; Peri-urban survey in Sleman and Bantul districts, Yogyakarta Special Region, December Note: Time required for cooking three dishes, not accounting for fuel preparation time. sequence and heat ranges required for some of the most common dishes as an example of the variations in heat output required by a stove. The information gathered constitutes detailed feedback to stove developers on how people are using their current stoves and what outcomes they expect from their efforts. New stoves aspiring to enter a specific market will have to be able to achieve results that are at least as good as those obtained from their traditional counterparts, not only in terms of efficiency and emissions, but, most important for users, in terms of the resulting food. As such, new stoves must enable local dishes to be prepared properly. Time spent cooking Findings indicate important variations in the time required for cooking when using different stoves. A complete meal including three dishes cooked on a three-stone fire can take as long as 87 minutes. Cooks using other traditional biomass stoves or a combination of biomass and LPG spend on average minutes cooking a similar meal, while the average times for reheating lunch and dinner were 6 and 30 minutes, respectively. Users of LPG stoves spend an average of 65 minutes for the main meal, and 5 and 21 minutes for lunch and dinner, respectively. This data indicates that cooks would save approximately 22 minutes on a main meal when using an LPG compared with a three-stone fire, without accounting for the additional time required for fuel preparation. 1 Water boiling Water boiling is a daily activity conducted by almost all households surveyed. Of these, 60 percent boil water twice a day. Biomass stoves are by far the most popular for this task. The frequency and amounts of water boiled suggest a potential market for dedicated 1. This data is approximate as other factors including family size, type of biomass stove used, and humidity content in the biomass were not considered.

22 16 From the Lab to the Field and Back: Synthesis Report water-boiling devices. Nonetheless, it is recommended that additional information be collected among potential targets to assess viability. Currently, powerful fixed stoves that are also used for other tasks (including palm sugar production) constitute a strong competition for dedicated water boiling devices in rural areas. However, such devices could appeal to peri-urban households that currently use Keren stoves, mainly because the Keren stoves are not stable enough to permit boiling large amounts of water at once. The street biomass economy Supplementing home cooking with purchased meals is a widespread practice among urbanized households in Indonesia. Most of those in the sample who purchased food belong to the group that combines fuels (14 percent), whereas those who buy the least food outside the home tend to use only biomass (5 percent). The foods purchased from food vendors are usually complex traditional/regional dishes that are time-consuming to prepare at home. Thus, an important segment of the population that uses an LPG mix is outsourcing some of the dishes that require intensive use of the stove to street vendors who use either firewood or charcoal. What cooks want: desired stove features In focus group discussions, participants were asked to characterize the positive and negative features of stoves, identifying gaps in each. The qualitative findings were then validated through a survey questionnaire administered in the Bantu and Sleman districts. Responses point to four stove characteristics as the most important at the sites sampled: (i) speed and power; (ii) rapid ignition; (iii) durability; and (iv) fuel efficiency. Respondents also found important the stove s convenience and its capabilities for solving certain contextual problems. Under convenience people indicated ease of use, ease of heat adjustment, ease of cleaning, and the potential of the stove to produce less or no smoke and to operate with any type or size of wood (table 6). Price did not rank among the top areas of interest for respondents, although it was still important for 60 percent of the sample. Aesthetics and a modern look were important for almost 40 percent. Regarding differences in preference between male and female respondents in focus group discussions, men expressed more interest in what the stove might look like and in the ideal of a modern stove made of metal. In rural areas, men were also interested in maintaining the additional by-products of traditional stoves such as ash to use as fertilizer and smoke to dry crops and seeds. In addition, in rural areas, where men are still traditionally the builders of the stove, they also appreciated hand-made stoves. Women, on the other hand, were far more interested in the functional aspects of the stove, including speed and power, durability, efficiency, and convenience. Interviewed retailers (all of whom

23 Findings 17 Table 6. The top 10 features sought by cooks at sampled sites, in order of importance Feature Percentage of peri-urban sample that considered the feature very important or important Focus group discussion (peri-urban-rural) Powerful/fast 99 Ranked first in preference among focus groups. Rapid/easy ignition 99 The largest amount of smoke is produced at ignition. Durable 99 Less of a concern in rural than in peri-urban areas where the Keren clay stove is prevalent but lasts only a year on average. Fuel efficient 97 Despite the availability of free/cheap firewood, households are interested in using it more efficiently. Convenient/easy to operate 90 People want stoves that are easy to operate, ignite, clean, and regulate. Less/no smoke 90 Cooks consider smoke uncomfortable but do not perceive it as a major threat and are not aware of the long-term effect of sustained exposure. Uses any type and size of firewood Can operate with damp wood 90 Stoves that require good-quality wood to operate are considered less attractive, since hard wood has to be purchased rather than collected. 59 Humidity directly affects combustion, ignition times, and levels of smoke; stoves that can burn wood that is somewhat damp would be useful during the long rainy season. Portability 58 Fixed stoves are nontransferable and cannot be resold or given away as household assets, reducing their value once purchased or built. Multiple burners 49 Important for preparing dishes simultaneously, saving time, and using firewood more efficiently. Source: Focus group discussions in Yogyakarta and Sumba December 2012 and 2013; Peri-urban survey in Sleman and Bantul districts, Yogyakarta Special Region, December were men) were interested in the modern, metallic, light look and believed that simpler stoves would be unappealing to consumers. The safety aspects mentioned during group discussions referred to the potential risk to children, the cook, and the household as a whole and were mentioned mainly by women, although some men also deemed this aspect important. Traditional stoves are considered generally safe; by contrast, some people are afraid of metallic surfaces that could burn children or cooks. Accidental LPG explosions during the early days of the substitution program

24 18 From the Lab to the Field and Back: Synthesis Report Table 7. Focus group discussion of advantages of various stove types in Yogyakarta, December 2012 Stove Fuel Traditional mobile clay 1-pot (Keren) Wood Improved mobile clay 2-pot (Keren Wood Gandeng) Traditional mobile stone 1-pot (Doughnut) Apollo Wood chips/ sawdust Sawdust rice husk Clean/smoke-free Fast/hot Durable Easy use/ignition Multi-purpose Safe Efficient Movable Traditional Cement fixed 2-pot self-made Wood 3 3 Traditional Clay fixed 2-pot self-made Wood Traditional Stone 2-pot Wood Improved Jolentho fixed 2-3-pot clay, cement, brick/ liners and chimney Wood Anglo Traditional clay 1-pot Charcoal Anglo supra Improved clay, cement, zinc 1-pot Charcoal Kerosene Kerosene 3 3 LPG 1 or 2 burners LPG Electric rice cooker Electric Source: Authors. Beautiful Economic instilled fear among users that appears to have subsided in urban and peri-urban areas, although some people in rural communities are still afraid of LPG stoves. Interestingly, health aspects were not particularly salient for men or women. While women cooks consider smoke uncomfortable, they do not perceive it as a major threat and are not aware of the long-term health effect of sustained exposure. This suggest that the health argument may not be the most effective way to convince users to use clean stoves. User assessment of stoves currently present in Yogyakarta During fieldwork, focus group respondents assessed their current stoves using the criteria they had already identified as important. None of the stoves met all of the criteria, but

25 Findings 19 those that came close included the LPG, the Doughnut, and the Jolentho (table 7). These are also among the most expensive in the Indonesian market. The stoves that ranked the lowest included the most prevalent mud and Keren stoves. For the CSI program in Indonesia, this means that there is a latent market for stoves that offer a closer fit with consumer preferences. The next question is how users address the deficiencies of the stoves currently available. One of the answers is through stacking. The lesson from stove stacking The combined use of two or more cooking devices, a practice known as stacking, allows users to benefit from the positive characteristics of their stove portfolio while minimizing the disadvantages. Stove stacking is widely practiced at the sites visited, improving efficiency, responding to changing taste preferences, obtaining niche benefits, and addressing fuel scarcity. In order to achieve fuel efficiency, biomass stoves are used for dishes requiring longer cooking times and using larger amounts of fuel. The LPG is used for tasks such as rapidly reheating already prepared food or boiling small amounts of water for immediate use. Time efficiency is also achieved by the simultaneous use of stoves allowing various foods to be prepared in parallel. This type of stacking is also practiced using the same or different fuels. Stacking also accommodates changing taste preferences. For example, some respondents using LPG for most of their cooking would still use a biomass stove to cook rice, which is considered more fragrant when cooked over a wood fire. Others would use firewood for most meals but boil drinking water on the LPG stove to avoid the smoky taste and smell of wood. In addition, some stoves serve specific purposes. It is unlikely, for example, that palm sugar would be cooked on an LPG stove owing to fuel requirements and costs. In Sumba, kerosene permits baking with an add-on oven. This task is impossible on the current biomass stoves available. In the sampled sites, all the fuels become scarce at times. When stocks of firewood run out during the long rainy season, leaving only damp wood, households intensify the use of their LPG stove if they have one. At times, shortages of cash to purchase fuel force LPG or kerosene users back to their biomass stoves. Frequent power blackouts make the use of the rice-cooker uncertain. All of the stacking practices observed suggest that respondents do not perceive the various stove technologies as mutually exclusive, but rather integrate stoves and fuels in a coherent system that maximizes the benefits and minimizes the constraints of each type of stove. This in turn indicates that whereas clean stove programs may aim at the adoption of improved stoves over traditional ones, the user s strategy is to adopt devices that complement rather than replace current cooking systems. The uptake of clean stoves will thus depend in part

26 20 From the Lab to the Field and Back: Synthesis Report Box 2. Three examples of successful stoves To better understand trade-offs, some examples of relatively successful stoves were analyzed. The Keren stove. Despite its fragility, the Keren has been able to penetrate the peri-urban market by offering an alternative to the large, fixed mud stoves common in rural areas. This is the cheapest stove produced by local artisans. Its advantages include its small size, light weight, and portability, which allows it to be used in smaller kitchens in densely populated peri-urban areas and to be moved as needed, making it attractive for people renting living space rather than owning a house. It is offered in three sizes to suit amounts cooked and household size. The Keren is also flexible, burning twigs and leaves as well as firewood. Its distribution network is wide; it is available at most corner stores. The driving force for the widespread adoption of the Keren has been its capacity to successfully address the contextual and lifestyle changes of people moving from rural to peri-urban communities. The HOCK kerosene stove. The HOCK is another example of the importance of quality and customer support. Various brands of kerosene stove are sold in Sumba, but the Indonesian HOCK is the most popular despite also being the most expensive. Customers indicate that original spare parts for the HOCK are widely available and that local technicians are skilled in fixing the stove. More importantly, the HOCK is ISO9001 certified, and the company offers customer support in locations across the country and by phone, text (SMS), , Facebook, and Twitter. Indonesia has almost 85 percent mobile phone penetration (212 million people), and 28 percent have Internet access (71 million). The Jolentho improved stove. Developed by Yayasan Dian Desa, the improved Jolentho is an example of adoption of an expensive clean stove ($25 30) for the artisanal production of palm sugar in Central Java. Despite the cost, interviewees who had adopted the stove touted its multiple benefits: efficiency, with self-reported fuel reduction of percent; no soot; and power, which yields important time savings over traditional fixed stoves. The Jolentho can also operate with damp wood when already hot. The characteristics that have driven sales of the Jolentho match the user preferences identified by this study (power, speed, use of damp wood). The adoption of the Jolentho in Central Java has also provided key information about the acceptance of innovation in the area. Because adopting any innovation entails a degree of risk, the process of approaching an innovation is one of risk reduction. Interviewees adopted similar strategies to reduce uncertainty. They tried to increase their level of information on the Jolentho before making a decision by (i) requesting information from people who had had direct experience with the stove, such as friends or promoters; (ii) traveling to observe the Jolentho at work and to try it out; (iii) inquiring of members of farmers groups (peers) or people they trusted (friends); and (iv), in all cases, referring to more than one source of information. The Jolentho case suggests that in Indonesia, people are generally not averse to innovation and are open to new technologies.

27 Findings 21 Table 8. Different aims for clean cookstoves Promoters of improved stoves Common goods and delayed individual benefits Improved health from lower emissions Energy efficiency Improved resource use Lower workload for women Stove users Personal, direct, and immediate benefits Speed: stove cooks fast High heat Fast and easy ignition Fuel efficiency (lower costs and less dependence on scarce dry wood) on their ability to solve problems perceived by users and to show clear benefits that can be integrated into the existing stove-fuel system. Stoves that fail to provide added value that exceeds the trade-offs of having to adapt to a new device have the lowest chance of success. To provide value, improved stoves must produce the same or better results in terms of taste and suitability for cooking common meals, while also providing additional benefits such as improved ignition, speed, durability, and fuel efficiency. Some successful stoves are profiled in box 2. What has to be improved to justify purchasing a new stove? The previous section presented findings on consumer preferences and highlighted the motivations of potential consumers in purchasing a new stove. The findings indicate a disconnection between consumer preferences and many aspects of the clean stove movement as it is presently configured. In paying for research and development work on clean stoves, the development community is interested in common goods and delayed individual benefits such as improved health outcomes from reduced exposure to smoke, energy efficiency, lower environmental impact, and resource efficiency (table 8). Stove users, by contrast, are interested in direct, immediate, and concrete benefits, including stoves that cook fast and are powerful, that light quickly and easily, and that are fuel-efficient. The disconnection between the motivations of users and the promoters of technology is not a rare phenomenon. The Green Revolution, for example, was motivated by yield increase, while farmers emphasized taste and the ability to prepare familiar food from their produce. In the case of stoves, these various sets of objectives and motivations need to be integrated to ensure that improved stoves address global as well as individual concerns. The experience so far indicates that clean stove programs that do not address consumer demands may face adoption problems even if the stoves are technically superior to traditional stoves. The challenge will be to strike a balance between supply-side objectives and demand-side needs and preferences.

28 22 From the Lab to the Field and Back: Synthesis Report Who does what? The household division of labor In Indonesia the household division of labor reflects the view that women are responsible for children and family. Rapid economic changes in the country over the last decade have not fundamentally altered that view or the associated gender roles. Despite women s inclusion in the labor market, the view of women as nurturer of the family permeates all aspects of women s lives. Women who work outside the home have to a large extent continued to bear primary responsibility for household chores, including cooking and related activities (96 percent of the women in the sample were the primary cooks in their households). The more urbanized a household, the less men tend to be responsible for fuel gathering and chopping, stove building, and cooking tasks, indicating that changes in stove type can affect the distribution of labor within the household. In rural areas, although cooking is mainly the woman s role, men take part in tasks surrounding the cooking process. Women perform tasks that are deemed light, such as fine chopping of firewood, cooking, clearing ashes, cleaning the kitchen, and maintaining the stove, while men assume tasks believed to be more physically demanding, such as chopping larger logs for storage and building and repairing stoves. In cases where the stove has a chimney, men clean and repair it, as this is considered a dangerous task. In urbanized areas, where households have started to use manufactured stoves such as the portable clay Keren or an LPG stove and to purchase rather than collect firewood, men have become less involved, as building and repairing are no longer necessary. Design features can have an effect on gender-equality outcomes The diameter of chopped wood varies according to the stove used. Given women s responsibility for firewood chopping, stoves that require small pieces of firewood present the risk of adding to women s tasks and increasing the time they must devote to operating the stove. In Sumba, for example, most cooks who use firewood burn it in open fires, and branches are kept long as a way to control the heat. Gasifiers and other stoves requiring small cuts of wood would increase the time needed for chopping, with the burden falling on the girls responsible for chopping. This consideration extends beyond biomass stoves. In households that adopt biogas, girls are likely to have to spend more time cleaning the stalls of the small animals that produce the waste for the biogas digester, as girls are traditionally in charge of caring for small animals, including pigs and chickens. This can lead to a situation in which the time savings obtained by not having to collect firewood are offset by the additional time required to care for animals and clean their stalls. This points to the need for new stoves to be assessed in light of their direct and indirect effects on the current roles and potential new roles of women and girls.

29 Findings 23 Decision-making over household resources Husbands are not directly affected by the negative effects of traditional stoves, but they have critical decision-making power over household spending. In Sumba as well as Java, women are traditionally the managers of household finances. As a general practice, husbands give most of their cash income to their wives, who then are responsible for seeing that all household needs are met. However, management of household resources does not automatically translate into decision-making over allocation. Focus groups and interviews revealed that women have spending discretion only to a certain threshold (between $5 and $10, in our sample). From the current selection of stoves, then, women can decide to purchase a Keren, an Anglo, or an Anglo Supra. (The last two burn charcoal.) For more expensive stoves including the Jolentho, a kerosene or LPG stove, and rice-cookers the final decision lies with the husband. Female respondents indicated that this is not always an easy or successful negotiation. These findings strongly suggest that marketing and outreach strategies will need to target men, since their buy-in will be critical to increasing uptake. Furthermore, actions to increase uptake will need to go hand in hand with enhancing the status of women and promoting gender equality. Photo: Veronica Mendizabal

30 24 From the Lab to the Field and Back: Synthesis Report Holistic responses to address smoke exposure Alongside fuels and stoves, the kitchen is another element to consider when examining cooking practices. Kitchens reflect their context. At high altitudes, the traditional Javanese kitchen has little or no ventilation, reflecting the need to conserve the heat released by stoves. Kitchens in the warmer lowlands have permeable roofs and walls made of porously woven palm leaves and bamboo that allow smoke to escape. Some households have small openings in the high roof directly above their fixed biomass stoves. In urbanized areas, the kitchen has undergone transformations, becoming smaller. The survey found that better-off peri-urban households have, on average, more ventilation outlets in their kitchens (windows, chimneys, vents, small openings in the roof) than poorer households. This finding is troublesome, as poorer households are the ones most likely to use biomass. If their kitchens are less ventilated, these households are at higher risk of emission exposure than the better-off who tend to use a variety of fuels, including cleaner ones. An interesting link was observed between a choice of clean fuel (LPG, kerosene, rice-cooker, Jolentho) and kitchen improvements. Most fuel mixers have two kitchens, a traditional one, detached from the main house, with earth floors and woven palm walls, where the biomass stove is used, and a second space, usually attached to the main house, with brick walls and a cement floor, where clean stoves are stored and used. Other valuable appliances may also be kept in this area, including the water filter or the refrigerator. This spatial segregation of cooking devices casts light on the perceived hierarchy of stoves, suggesting that clean stoves lead to new use patterns and to kitchen improvements. The survey in peri-urban neighborhoods in Yogyakarta asked respondents about their priorities when changing their cooking environment. About 76 percent answered that their first priority would be to have a cleaner kitchen without soot. Other desired improvements included brighter and airier kitchens (57 percent), concrete floors (39 percent), and more space (27 percent). When asked specifically about smoke and soot, more than 50 percent of the sample indicated that they would want less soot on ceilings and walls and less smoke while cooking; 53 percent would want less smoke when lighting the stove; and 48 percent would want less soot on pots and pans. However, when asked about changes in terms of stoves and fuels, people did not focus on changing either, but rather indicated that the first priority would be to reduce firewood collection times (30 percent). Approximately 40 percent of the sample considered buying a new stove unnecessary (71 percent in the case of an LPG stove). A narrow emphasis on clean stoves limits the potential for wider transformations encompassing the three aspects in the system: cooking fuel, cooking device, and cooking environment. Understanding this dynamic (and for the additional purpose of increasing kitchen ventilation to reduce exposure to indoor air pollution), YDD has piloted a loan scheme to

31 Findings 25 allow households to upgrade their kitchens after purchasing an improved biomass stove. The scheme promotes ventilation and is relatively successful, but outreach has been limited to a few villages. The importance of engaging community groups Community groups were found to be strong and to play key roles in social cohesion, promotion of innovation, and informal saving and credit schemes. These groups are widespread across Indonesia but are particularly well established in Java, where they have high membership rates, making them excellent entry points to the community. In Sumba, extended families play similar roles in community cohesion and in linking rural and urban populations. People in Java and Sumba try to avoid credit whenever possible. One of the main reasons is the fear of losing face if repayment is not possible. Nonetheless, informal credit and savings groups (Arisan) are popular. Members usually pay monthly fees that are withdrawn as savings after a certain time. Women and men organize separate groups, however respondents indicated that women need their husband s consent if requesting a large credit from the Arisan. Producer groups, including farmers groups, cattle growers, and palm sugar associations, are also important, particularly in rural areas. While the groups represent different production interests, they have the same underlying goal of serving their members with information on markets, innovations, credit, and financing. The promotion of consumer goods through organized community groups is common. Retailers target savings and credit groups on a regular basis, and development NGOs such as YDD have also engaged with the community through producers associations. Such groups play a key role in facilitating informal credit to members, and this is one of the ways in which the Jolentho stove has spread in Central Java. Community groups can play key roles in facilitating credit to purchase cleaner stoves, diffusion of information, awareness raising, on-site stove testing, provision of feedback, and monitoring stove uptake and use. However, although purchases of stoves can be facilitated by the access to credit, the family will consider the decision carefully; as with other important household expenditures, the final decision will generally remain with the husband. Phase 2: Assessment of Stoves and the Development of a Social Protocol for Stove Testing The second phase of the work included an assessment of stoves from the user s perspective and the development of an experimental method for evaluating the performance of clean stoves in their context of use. Five stoves that had passed the Indonesia CSI emissions and efficiency tests (a water heating test developed by the Indonesia CSI and the Indonesian National Standard Test) were selected for assessment under field conditions. The

32 26 From the Lab to the Field and Back: Synthesis Report Figure 4. Social protocol framework Social assessment Fit with context Functionality and usability Emotional response Local fuel conditions Cooking environment Stove systems Gender roles Allocation of household resources Affordability Credit Fuel preparation Placement Ignition Overall speed Operation Attention requirement Heat regulation Satisfaction with: Operation Ergonomics Functions Design features Outcome methodology was developed and tested in December 2014 and its results were later validated in a second round of testing in May Figure 4 presents the framework developed for this social protocol, along with the areas assessed. How did clean stoves perform in the field? The findings of the first and second application of the protocol are presented below. In the tables reporting the results, the stoves are referred to only by number. The results obtained by each stove have been shared with its manufacturer. The approach tries to replicate as much as possible a scenario of large-scale market-based promotion. For this reason, no training or demonstration was provided to testers before they used the stove for the first time, although all questions were answered throughout the assessment. This design, without previous training, made it possible to observe users first encounter with and decoding of the stove, specific areas of difficulty, and the level of success in users own problem-solving strategies, allowing us to better understand the issues users face in their initial handling of the product and clarifying the degree of training that might be required for each product. The design also recognizes that funding and organizing training or demonstration is not always possible in large-scale interventions, as is the case of the Indonesia Clean Stove Initiative.

33 Findings 27 Fuel moisture Findings suggested that clean stoves in Java would have to be operable with high levels of moisture in the fuel given the seven months of annual rains (October to May). In December, three months into the rainy season, the average humidity of the firewood used to operate the clean stoves and the baseline equipment was percent. In May, seven months into the rainy season, the moisture content averaged percent. 2 For comparison the wood in the lab was measured at 8 11 percent moisture. This indicates that stoves are tested in the lab under conditions very different from those in the field. Fuel preparation The additional time required to prepare wood for some clean stoves was identified as a potential obstacle for their sustained use. Stored wood in the field is generally long (1 2 meters) and of varying diameter (2 10 cm). The baseline stove can operate with such sizes. Owing to their frontloading design, some clean stoves can also operate with long wood, requiring only some splitting to reduce the diameter. In our testing, additional preparation time required for these stoves ranged from 2.5 to 5 minutes. Top-loading stoves need much shorter cuts of wood (approximately 3 12 cm). To prepare those cuts, stove testers needed up to 28 minutes. Table 10 (Annex 1) presents the average additional time required to prepare the fuel for each type of stove. In view of the gender division of labor in Java (men and women collect wood, men cut and store it, women cut at time of use), the additional burden of preparing fuel for use in a clean stove would likely fall on women. This would increase the time they need to spend on cooking tasks and decrease the time available to them for other activities (including incomegenerating activities). Ignition Ignition times for clean stoves were found to be higher than for the baseline. In some cases, much longer times were required. The baseline stove could be lit in minutes, whereas clean stoves needed up to 15 minutes in some cases. The average ignition times for all stoves is reported in table 11 (Annex 1). Cooking times The baseline needed 90 minutes for the five cooking tasks in both December and May, despite higher fuel moisture in May. Only one clean stove matched this time. All others required longer times, ranging between 120 and 170 minutes that is, percent longer than the baseline. See table 12 (Annex 1) for average times. 2. The fuel moisture was calculated using the SeTAR MC1.14:2013 protocol.

34 28 From the Lab to the Field and Back: Synthesis Report Re-ignition Two of the clean stoves tested had to be relit frequently. In one case the fire tended to go out following refueling; in another the fire would die owing to the presence of large amounts of charcoal in the chamber. In one case, the stove might increase levels of indoor air pollution and exposure during re-ignition, thereby offsetting the reduction in pollution achieved during operation and making it critical that people be able to Photo: Voravate Tuntivate operate the stove properly. Training as well as other strategies, such as post-purchase support networks and access to trouble-shooting, would be needed. (A few users might be able to operate the stove without training, but if a stove is too complex for most people to be able to decode it intuitively, it will not catch on.) Power Testers were in general satisfied with the power (heat output) of the clean stoves, although in a few cases the stoves could not generate enough power for frying. To increase the power of the stove, most testers overloaded the chamber, leading to lower performance and charcoal build-up. Safety Most clean stoves were considered safe, in fact much safer than the LPG stove, which is a great opportunity for marketing. A couple of stoves were considered dangerous owing to charcoal build-up and storage of glowing charcoal next to the stove, and because of a design element that is sharp and difficult to manipulate when hot. Design Testers liked the appearance of the clean stoves. However in one case, a design element was found to be inconsistent with local practices. This is an add-on skirt that has to be fitted to the pot to improve efficiency and reduce emissions. The problem is that Indonesian dishes are cooked in a myriad of pot types and sizes, and the skirt has to be fitted to these. It works

35 Findings 29 with some pot diameters but not others. In addition, it is difficult and dangerous to handle, and in most cases people do not adjust the skirt at all. In a couple of cases in May people decided to stop using it after the first cooking task. 3 Resulting food and fuel consumption Testers in general considered that the food was good, except the fried food prepared with the one stove that could not achieve high power ranges. The clean stoves used less wood in all cases. Table 13 (Annex 1) shows the results. User satisfaction ratings At the end of the test, each tester was asked to assess the stoves on six areas of importance for biomass stove users in the region (as identified during the previous social work and survey). The areas are ignition, power, speed, ease of use, cleanliness, and efficiency. Testers were asked to rate each stove from 1 to 5, five being the best rating (fully satisfied) and 1 the worst (completely unsatisfied). The results are presented in tables (Annex 1). Clean stoves rated high on efficiency, but testers were not satisfied with their overall performance on ignition, speed, and power. Two clean stoves rated below satisfactory on cleanliness. Price and credit Indicative prices were given to the testers based on distributors initial quotes. Prices ranged between IDR 250,000 and 900,000 ($21 75). In all cases the prices were deemed too high and unaffordable. In all cases testers would have had to discuss and agree with their husbands before purchasing the stove. Only in a couple of cases did testers say that they would consider credit. Credit would also have to be agreed to by their husbands. Introduction of pellets This new fuel intrigued the testers. Their first reaction was to calculate pellet vs. LPG consumption and cost. Their calculations suggest that pellets would be more expensive than subsidized LPG. 3. Based on this finding, the stove has to be retested for emissions and efficiency in the lab to see if it will continue to be included under the subsidy scheme of the Indonesia CSI.

36 30 From the Lab to the Field and Back: Synthesis Report Photo: Laurent Durix

37 31 Conclusions and Recommendations The study provided an overview of the landscape of fuel and stove use in two very different areas of Indonesia. Despite the difference between the two areas, some strong similarities were found. These similarities point to key opportunities and constraints for clean stove programs. Key opportunities include the generally open attitude to innovation and new technologies, strong communal organizations that support access to informal credit and information, and, with specific reference to stoves, the widespread practice of stove stacking, in which stoves are seen as complements rather than competing devices. 4 Constraints include the limited decision-making power of women over household spending decisions, which limits their capacity to decide on the purchase of clean stoves over a certain threshold. Men, who are decision-makers but are not directly affected by the negative effects of smoke exposure, may not see the added value of investing scarce household resources in devices that can be produced at home with local materials and at no cost. Another critical constraint is the misalignment between clean stove supply and demand. Promoters and funders of clean stove programs are interested in achieving long-term common goods and focus their efforts on reducing emissions, exposure to emissions, increasing burning efficiency, and improving health outcomes. Users, on the other hand, are interested in direct personal benefits resulting from the functional features of stoves (heat, speed, ignition, and fuel efficiency, among others). Finally, the findings of the CSI assessment of stoves indicates that there is room to further improve cookstoves, to invest in a next generation of stoves that are tested in the lab and in the field through successive iterations in which user feedback is gathered and systematically integrated. The assessment also suggests that marketing strategies can benefit from a systematic assessment of how people would actually use the products and from identification of the products and areas that would require the most support and investment. Furthermore, understanding problems of usability at the early stages through tests involving actual users can help in determining the types of investment that would be required in marketing, demonstration, training, and post-purchase support. Recommendations follow. 4. Real gains in terms of reduced exposure to emissions would be achieved with the exclusive use of stoves for a majority of cooking tasks. Nonetheless, stacking provides an entry point for users to begin taking advantage of clean stoves.

38 32 From the Lab to the Field and Back: Synthesis Report Table 9. Summary table of recommendations Recommendations on stove design Recommendations on stove testing Recommendations on the promotion of clean stoves Recommendations on holistic approaches to reduce exposure to indoor air pollution Integrate the interests of consumers in the design of improved stoves; move from supply- to demand-based stove technology Address the differentiated preferences of women and men Assess the level of improvements in areas of interest to users Ensure functionality Simplify to increase usability Use feedback mechanisms to improve design, distribution, and outreach Invest in understanding local cooking practices Determine and define cooking cycles and incorporate these in the lab testing protocol Conduct social adoption tests and feed the results back into testing protocols Share results with stove designers and promoters Address the differentiated needs of rural and urbanized populations; consider prioritizing areas that have been exposed to a diverse range of fuels and devices Ensure that the pricing structure responds to the purchasing power of the target groups and to the scope of women s power to make purchasing decisions Consider conducting a gender audit of selected stoves Engage community groups in promotion Conduct awareness-raising campaigns Involve men/husbands Target additional consumer groups beyond household consumers Use mutually enhancing solutions to increase the impact of work on clean stoves Explore the potential of supporting fuel diversification Monitor and evaluate impacts Continue to develop the CSI social testing protocol

39 Conclusions and Recommendations 33 Recommendations on Stove Design Integrate the interests of consumers in the design of improved stoves; move from supply- to demand-based stove technology Based on the field work, technical improvements in smoke reduction will not be enough to attract consumers unless other features that people seek in a stove are also present. In other words, uptake will depend on how well stoves respond to consumer preferences. The CSI should encourage the integration of key features that users consider important as entry points to increase the acceptability of clean stoves. In the specific cases studied, users found the following characteristics important: Powerful and easily regulated heat output Speed and ease of ignition Durability Efficiency Convenience and ease of operation Less smoke than traditional stoves Ability to accommodate many types and sizes of firewood Address the differentiated preferences of women and men Women as primary users are interested in the functional aspects of stoves, whereas men are interested in cost and aesthetics. These aspects should be taken into account in marketing and promotion strategies. Assess the level of improvements in areas of interest to users Current tests are largely focused on the efficiency and emissions of clean stoves. The results of testing are used to certify the stove and to inform consumers of improvements in these areas. To encourage uptake, however, increases in cooking speed, ignition times, durability, and other features of interest to consumers should be formally assessed, rated, and reported to users. The Indonesia CSI has developed a social testing protocol as a way to integrate the social dimensions of stoves and user preferences into testing. The findings of this initial work enforce the argument for conducting field tests under real conditions as part of the stove

40 34 From the Lab to the Field and Back: Synthesis Report certification process. Through this work the CSI can engage with and contribute to the working group on field-testing methods of ISO Technical Committee 285. (The working group was launched in late 2014.) To date, field tests have focused on user satisfaction; only a few testers have tried to develop more systematic tools to assess field performance (such as the kitchen performance test developed by Rob Bailis in 2007), and these have concentrated on validating emissions and efficiency results, rather than looking at performance on a set of criteria of interest to users. The Indonesia CSI is in a good position to move testing in the right direction. Ensure functionality Beyond desired design features, the stove must be able to perform its main task of cooking food that suits the tastes of users in their specific contexts. One of the areas in which the CSI has made progress is in the use of a water heating test to ascertain whether stoves proposed under the program allow local dishes to be properly prepared. This work has been further complemented by the social assessment discussed in this report, which assesses performance in areas such as ease of use, time savings over traditional stoves, fuel savings in financial terms, and a stove s capacity to operate with damp wood, among others. Simplify to increase usability The stove design should be intuitive, minimizing the level of complexity needed for operation. Because cooking is an everyday task carried out simultaneously with other tasks, it should be made as easy as possible and require little or no formal training. Training is always mentioned in discussions of how to improve the rate at which technology is adopted. However, if a program to promote the adoption of clean stoves is to succeed widely, it cannot depend on training beyond the rudimentary level. Instead, stoves must be designed to reduce the amount of training needed to operate them. The study in Indonesia has demonstrated that ease of operation is one of the characteristics most important to users, and the social assessment has showed that some stoves are getting it right, with intuitive designs, while others are still being developed without much consideration of how the user will interact with the product. Use feedback mechanisms to improve design, distribution, and outreach Feedback mechanisms: Ensuring that supply responds to consumer preferences requires a thorough understanding of what works and what does not in terms of technical design, development, distribution, and retailing. For this reason, interaction with users to obtain their feedback is recommended at two levels:

41 Conclusions and Recommendations 35 Technical feedback: on features, usability, maintenance, and repair in order to understand device constraints and thus improve design, maintain key features, and change others in future models Social monitoring: on usage, integration into cooking, and stove stacking practices in other words, how is the clean stove actually being used? While sales provide an indication of the degree of acceptance of new stoves, more detailed feedback on and analysis of the specific factors affecting consumption is useful to determine the extent to which price and other factors such as functionality, fit with cooking practices, beliefs, household relations, and so on are driving demand. This information can help fine-tune design, as well as pricing, distribution, retailing, promotion, and marketing. There are various ways to obtain the necessary feedback: Direct feedback: The sustained use of focus groups among representative market segments is recommended to gather data on technical and social aspects at various stages of the introduction of new stoves. Collaborating with local groups: This is achieved through the deployment of testing teams composed of local evaluators (cooks) to test the stoves in the laboratory and in the field for their performance when preparing local dishes and in delivering the features most valued by users. The findings from such testing can also be used in marketing. Pre- and postassessment: Users should participate in prototyping and testing. In both cases user trials help determine how the stove is used and how it fits within the cooking system. An initial step in this direction has been the development of the CSI social testing protocol. Its application should be considered at various stages from prototypes to final products. Customer service: A customer service system should be established to provide technical support on use, maintenance, and repair, and to obtain specific data on technical aspects through the systematic recording and analysis of complaints and requests for trouble-shooting. It is vital to ensure that customer services, including information and support, are established and operational by the time clean stoves are released into the market. Customer outreach: Whereas a customer service system may be more focused on problem solving, a customer outreach system can be an excellent tool to communicate proactively with users. Examples of such systems include the use of social media as forums to share information with consumers about stove use and features, points of sale, costs, and available credit schemes, and to allow consumers to share their successful recipes, tips, and so on. This is one of the strategies used successfully by the Indonesian stove producer HOCK.

42 36 From the Lab to the Field and Back: Synthesis Report Who should collect this data and how should this feedback be used? It is difficult to imagine that small individual producers would be inclined to invest in such a feedback system. In addition, it is only through aggregation that data will become relevant. Thus, it is recommended that this work be supported by the World Bank or other development partners, or by large actors in the clean stove field (such as the Clean Stove Alliance) in collaboration with local governments and NGOs. The available data should be analyzed periodically and fed to designers and market aggregators. This information will help not only to fine-tune the manufacturing and marketing of stoves, but also to improve the CSI, its incentives, and its targets. Through the work done by the CSI to develop both the technical and social tests, national capacities have already grown in Indonesia. Recommendations on Stove Testing Invest in understanding local cooking practices Testing that does not emulate the actual conditions of use (such as the need for cooks to attend to other household tasks while cooking) is not likely to provide an accurate indication of a stove s appeal. In other words, artificial testing conditions produce artificial results. Determine and define cooking cycles and incorporate these in the lab testing protocol Indonesian cooks have other duties both within and outside the household. Their cooking sessions reflect these realities, and so should stove tests. If three meals are prepared in the morning, with two of those meals stored for later reheating, the testing protocol must reflect that reality. Conduct social adoption tests and feed the results back into testing protocols Cultural norms affect stove adoption and use, but unless testing protocols make testers aware of their importance, tests may not measure or detect certain features of a given stove that ultimately will affect the stove s appeal to consumers. Testing that gauges the social or cultural acceptability of stoves is a key part of the vital task of understanding demand. Share results with stove designers and promoters This recommendation is self-evident. The designers and manufacturers of clean cookstoves, as well as the organizations that promote their use, have a clear incentive to consider the results of finely tuned field testing. The success of their efforts depends on it, particularly when the availability of subsidies or other financing is contingent upon results obtained.

43 Conclusions and Recommendations 37 Recommendations on the Promotion of Clean Stoves Address the differentiated needs of rural and urbanized populations; consider prioritizing areas that have been exposed to a diverse range of fuels and devices In the case of Indonesia, the areas of potential prioritization would be peri-urban sites that are in the midst of a fuel transition, where the stove selection has been reduced to small, fragile stoves, and where women have access to cash income. These are sites where small, one-burner rocket stoves could have the most appeal and traction, given the similarities in appearance and operation with the Keren stove. Ensure that the pricing structure responds to the purchasing power of the target groups (high-end and low-end stoves) and to the scope of women s power to make purchasing decisions As indicated for the surveyed areas, the allocation of household financial resources is generally not under the control of women. The threshold of women s control over household finances has been identified and should be considered when pricing stoves, taking into consideration that the less integrated women are into the labor market and cash economy, the lower their threshold of control. In addition, the survey found a willingness to pay for high-end stoves, even among the poorest households, when the stove was clearly perceived as a good investment because it was durable (the carved stone Doughnut), fast (LPG), or productive (Jolentho). Thus, there is a market for high-end stoves, even among poorer households, if they can show direct and immediate benefits. High-end clean stoves have to be marketable on the basis of their direct benefits to consumers. Consider conducting a gender audit of selected stoves Technologies are rarely gender neutral. Most often they affect women and men in different ways. Even if a technology appears to be appropriate for all (gender neutral), it should be assessed for specific potential impacts on women. In particular, to what extent do proposed stoves empower women by reducing the time they spend cooking and collecting fuel? To what extent do they reduce women s exposure to smoke? Because household roles are gendered it is possible that a stove that reduces smoke may also increase women s work in the kitchen for example, if it requires small cuts of wood or increased attention during operation.

44 38 From the Lab to the Field and Back: Synthesis Report Stoves should be assessed in terms of their effects on pertinent gender roles and responsibilities. To what extent do stoves affect gender equality in positive or negative ways by: Adding or reducing time spent gathering fuel or cooking? Adding or reducing the physical burdens related to cooking and gathering fuel? Adding or reducing exposure to smoke or soot? Increasing or limiting women s decision-making power within the household or community? Engage community groups in promotion Market aggregators will require support with promotion. A concerted, centrally conceived campaign (which may be managed in a decentralized fashion) is required to ensure users needs are met and to make the product known and wanted. Mechanisms may include a promotional stove kit for early adopters or the establishment of user testing groups, such as trial kitchens or Arisan meetings where women participate in stove demonstrations and have the opportunity to try them first-hand. It is important to bear in mind that in the areas visited for this study it is men who make the final decision; for that reason their engagement and participation in testing, trials, and promotion is vital. People are willing to try a new technology if they hear about it from peers and friends and can test the technology themselves and experience the benefits of its use. Conduct awareness-raising campaigns Currently, efforts to promote stoves using a preventive health message as the entry the point encounter varying levels of awareness. Over all, however, awareness of the dangers of cooking smoke is low among users and their families. In addition, environmental messages are not likely to resonate, given that firewood is still generally considered abundant. Messages that link locally expressed needs and wants with solutions provided by the stove, plus additional benefits to which users may relate, probably have the best chances of engaging potential buyers. Another aspect that has a good chance of attracting interest as an entry points to sensitize consumers is the savings obtained when using an efficient stove compared with a traditional one. Messages should highlight the improved stove s unique features: This stove can do something that the traditional stove does not do well, or does not do at all. Identify concrete benefits of the improved stove and gaps that it can fill in the cooking system.

45 Conclusions and Recommendations 39 Involve men/husbands Targeted messages. Possibilities to engage men and increase the chances of allocation of household resources for medium-range or high-end stoves can be opened up through messages likely to interest men, such as ignition and other technical features of stoves. Engage men in prototyping and testing stoves and in developing promotional messages. Explore the potential of men s interest in advanced options, such as biogas (including the construction of the digester). Small, simple stoves such as the Keren, charcoal, or three-stone fire did not seem to interest men at the sites studied. Consider the use of incentives for groups that do not benefit directly from the use of cleaner or more efficient stoves, such as husbands, but that have decision-making power. Interests identified among husbands include the multi-functionality of stoves and stoves that can do more than just cook. Recently, some companies producing improved cookstoves have started producing models that bundle technologies such as mobile phone or LED charging into stoves. On Indonesian islands that are still not connected to the grid, such integration of direct benefits into the stove offers a way to interest men in acquiring an improved stove. Be aware of disincentives. Stoves that could increase the labor input of men (for example, through additional firewood collection or chopping time) might turn out to discourage the adoption of stoves. Worse, the additional tasks may be transferred to women or children in the household. Target additional consumer groups beyond household consumers Home-based enterprises are an important target group, given their importance in urban and peri-urban sites, where complex and time-consuming cooking tasks are increasingly outsourced to them. Similarly, palm sugar and salt producers are some of the groups that have been identified as intensive users of biomass (and thus highly vulnerable to smoke exposure). Recommendations on Holistic Approaches to Reduce Exposure to Indoor Air Pollution Use mutually enhancing solutions to increase the impact of work on clean stoves The work on clean stoves should be complemented with additional work on improving kitchen ventilation. Although it is reasonable to expect that efforts to promote the uptake of improved stoves will increase their use in the short and medium term, it is equally reasonable to believe, based on the practices reported here, that people will continue to stack stoves and fuels rather than switching completely to a clean stove, at least initially. Thus

46 40 From the Lab to the Field and Back: Synthesis Report cooks will continue to be exposed to various levels of harmful emissions. Findings indicate that the poorest households, those that depend heavily on biomass, are also likely to have the lowest levels of ventilation in their kitchens. Although kitchen ventilation may be outside the scope of clean stove programs, outreach and awareness-raising activities could beneficially include information on the benefits of kitchen ventilation and information on how to increase it in the most effective manner. In addition, findings indicate that, if they can afford to do so, households will make kitchen improvements following the acquisition of a cleaner stove. Credit schemes for improving kitchens could be used as an incentive for early adopters, with due attention to the risks of distorting markets. Explore the potential of supporting fuel diversification Although the CSI s main focus is on improved biomass cookstoves, the impact of efforts to reduce exposure to indoor air pollution could be increased by promoting the use of efficient energy mixes that combine LPG, firewood, and electricity. What is currently lacking is a full understanding of what the most promising energy mixes might look like and what gains they could bring in resource use and costs in different contexts. In this connection, the potential benefits of the already popular practice of stacking fuels should be explored with a view to the possibility of subsidizing the adoption of a set of devices powered by a variety of fuels, including electricity (for rice-cookers). Subsidies would reduce the upfront costs for households ready to jump into more advanced and cleaner cooking solutions. Monitor and evaluate impacts This is an area where a great deal of work is still needed. Studies so far are inconclusive, and evidence on the types and scale of impacts is mixed. What happens in the black box of technology adoption? How does the technology affect household relations, gender equality, socioeconomic indicators, and, of course, how it does it affect heath outcomes? Since 2015, the Indonesia CSI has included the monitoring and evaluation of stove adoption and sustained used as part of its marketing strategy. The results of this work were built into the CSI social testing protocol to obtain a baseline of initial use and a certain number of buyers to be monitored for adoption, use, and impacts. Continue to develop the CSI social testing protocol The key work now is to define how and in what form the new social testing protocol should be used (i) to support the design and development of new products and (ii) as a necessary stage in stove testing. Close coordination with the ISO working group on field testing is recommended in this regard.

47 41 Annex 1 Additional Tables In the tables below, stoves are referred to by number only not by name. The results obtained by each stove have been shared with its manufacturer. The source for all tables is the CSI social assessment and validation, conducted in December 2014 and May The approach tries to replicate as much as possible a scenario of large-scale market-based promotion. For this reason, no training or demonstration was provided to testers before they used the stove for the first time, although all questions were answered throughout the assessment. This design, without previous training, made it possible to observe users first encounter with and decoding of the stove, specific areas of difficulty, and the level of success in users own problem-solving strategies. The design is based on the recognition that funding and organizing training or demonstration is not always possible in large-scale interventions, as is the case of the Indonesia Clean Stove Initiative. Table 10. Average fuel preparation time, by stove, May 2015 Stove Additional fuel preparation time (minutes) Baseline 0 Clean Stove 1 0 Clean Stove Clean Stove Clean Stove Clean Stove Clean Stove Note: Fuel prepared ranged from 1.5 to 2 kg for 1 cooking session (5 cooking tasks).

48 42 From the Lab to the Field and Back: Synthesis Report Table 11. Average ignition time, by stove, May 2015 and December 2014 (in minutes) Stove May 2015 December 2014 Baseline Clean Stove 3 a 5.8 Clean Stove Clean Stove Clean Stove Clean Stove b Clean Stove b a. This stove underwent some design variation between December and May. b. Kerosene was used in December to aid ignition of these stoves. Table 12. Average cooking time, by stove, May 2015 and December 2014 (in minutes) Stove May 2015 December 2014 Baseline Clean Stove Clean Stove Clean Stove Clean Stove Clean Stove Clean Stove 3 a a. This stove underwent some design variation between December and May.

49 Annex 1. Additional Tables 43 Table 13. Fuel savings per stove compared with baseline, May 2015 and December 2014 Stove Fuel used (grams) May 2015 December 2014 Percentage of baseline Fuel used (grams) Percentage of baseline Clean Stove 6 1, , Clean Stove 4 1, , Clean Stove 3 a 2, Clean Stove 2 1, , Clean Stove 5 1, , Clean Stove 1 1, , Baseline 2, , a. This stove underwent some design variation between December and May. = data not available. Table 14. Testers average rating of stoves on key criteria, May 2015 Ignition Power Speed Easiness Cleanliness Efficiency Average Baseline Clean Stove Clean Stove Clean Stove Clean Stove Clean Stove Clean Stove

50 44 From the Lab to the Field and Back: Synthesis Report Table 15. Testers average rating of stoves on key criteria, December 2014 Ignition Power Speed Easiness Cleanliness Efficiency Average Baseline Clean Stove Clean Stove Clean Stove Clean Stove Clean Stove Clean Stove 3 a a. This stove underwent some design variation between December and May. = data not available.

51 45 Annex 2 Overview: The Indonesia Clean Stove Initiative Objectives In 2012, the World Bank, in collaboration with the Directorate of Bioenergy within Indonesia s Ministry of Energy and Mineral Resources (MEMR), launched the Indonesia Clean Stove Initiative (CSI). The Indonesia CSI aims to scale up access to clean and efficient cooking solutions in Indonesia through capacity building, policy development, and support for selected government action plans. Approach The Indonesia CSI took a phased approach. Phase I ( ) centered on initial stocktaking, which was critical for developing the implementation strategy, designing subsequent program phases, and establishing policy dialogue with the country s institutional focal point. Based on the findings from Phase I, Phase II ( ) focused on four areas of activity: (i) establishing a system of standards, testing, and certification of stoves; (ii) strengthening institutions and building stakeholder capacity; (iii) designing and implementing the pilot program; and (iv) designing and preparing a master plan for the national program. Impacts The Indonesia CSI: Provided in-depth assessments of the existing stove market (demand and supply) and reviewed the sector s institutions, policies, and key programs, thus building a technical foundation for further program interventions. Raised the issue on the policy agenda by (i) identifying and strengthening institutional champions within the national government and among stakeholders, and (ii) conducting broad national consultations. Generated a number of knowledge products that have been used to engage and raise awareness of stakeholders and share knowledge and experiences to date.

52 46 From the Lab to the Field and Back: Synthesis Report Developed a conceptual roadmap to achieve universal access to clean cooking by 2030 and engaged both the government and partner organizations to work on the roadmap. Designed and launched the pilot of a new business model using a results-based financing approach that aims to incentivize the private sector and promote market sustainability. The early results show promise for potential scale-up or replication using this approach. Developed an innovative stove-testing method that incorporates local cooking practices and sociocultural preferences based on anthropological findings and comprehensive household surveys. The new stove-testing method has contributed to ongoing ISO discussions on defining and testing clean cookstoves. Built a platform for learning, information dissemination, collaboration, and capacity building within and among key partner organizations through (i) establishment of the Indonesia Stove Alliance, (ii) frequent communication and active sharing of resources and expertise among country teams within the World Bank s East Asia and Pacific region, and (iii) strengthened engagement with key regional and international players.

53 47 References Cook, C Report to World Bank Clean Stove Initiative/ Indonesia Social Assessment Team. June. EAP Gender & Energy Facility and CSI (Clean Stove Initiative). 2015a. Stoves, Fuels, and Cooking Practices on Sumba Island, Indonesia: Findings and Recommendations of Qualitative Field Research. GPSURR, World Bank, Washington, DC b. Stoves, Fuels, and Cooking Practices in Central Java, Indonesia: Findings and Recommendations of Qualitative Field Research. GPSURR, World Bank, Washington, DC c. Clean Biomass Cookstoves in Central Java, Indonesia: A Quantitative Market Analysis. GPSURR, World Bank, Washington, DC. Miller, G., and A. M. Mobarak Gender Differences in Preferences, Intra-Household Externalities, and Low Demand for Improved Cookstoves. NBER Working Paper 18964, National Bureau of Economic Research, Cambridge, MA. April. Simon, G., R. Bailis, J. Baumgartner, J. Hyman, and A. Laurent Current Debates and Future Research Needs in the Clean Cookstove Sector. Energy for Sustainable Development 20 (June): Urmee, T., and S. Gyamfi A Review of Improved Cookstove Technologies and Programs. Renewable and Sustainable Energy Reviews 33 (May):

54

55

56 This report is part of a program of work on social and gender aspects of the development and promotion of clean stoves. Grounded in extensive research in Indonesia, the series consists of practical documents that can be used to integrate social and gender dimensions into work on clean stoves in East Asia and the Pacific and beyond. The target audience is clients and development partners active in the development and promotion of clean stoves. The publications in the series are available at publication/social-gender-support-to-indonesia-csi. For additional information, please contact Lourdes Anducta (landucta@worldbank.org). Photo: Veronica Mendizabal