Clean cooking in sub-Saharan Africa: modeling the cooking fuel mix to 2050

(1)

Master of Science Thesis

KTH School of Industrial Engineering and Management Energy Technology EGI_2017-0057-MSC

Division of Energy and Climate Studies SE-100 44 STOCKHOLM

Clean cooking in sub-Saharan Africa:

modeling the cooking fuel mix to 2050

Henri Casteleyn

(2)

(3)

KTH ROYAL INSTITUTE OF TECHNOLOGY

Clean cooking in sub-Saharan Africa:

modeling the cooking fuel mix to 2050

Submitted by Henri Casteleyn

Supervisor: Prof. Semida Silveira Examiner: Prof. Semida Silveira

A thesis submitted in fulfillment for the degree of Master of Science

in the

Division of Energy and Climate Studies Department of Energy Technology

School of Industrial Engineering and Management

June 2017

(4)

(5)

“There is no substitute for hard work.”

Thomas A. Edison

(6)

(7)

Abstract

As of 2014, 81% of sub-Saharan population or 792 million people rely on the traditional use of biomass to provide in their cooking needs. This situation causes harmful health, environmental, and development hazards with a substantial annual economic cost of USD58.2 billion. The concern about the issue of access to clean cooking facilities is growing as international organizations and national governments define steps to transform the existing situation.

Literature provides a good view on determinants for the cooking fuel choice in developing regions, but comprehensive outlooks for the future cooking fuel mix in sub-Saharan countries are limited. To this extent, the presented master's thesis aims to shed light on a history-inspired pathway for the evolution of the biomass dominated cooking fuel mix in sub-Saharan countries to 2050.

A quantitative model was developed to estimate the future uptake of various cooking technologies, from which the fuel mix can be derived using energy intensities. Projections were constructed for urban and rural areas in 45 countries. Economic development, population expansion, urbanization, and to a certain extent policies are the key drivers of the model.

Despite a moderate improvement in the share of population relying on traditional biomass, 808 million people in sub-Saharan Africa are expected to make use of traditional three-stone fires in 2050, an increase compared to 2014. Biomass remains the dominant cooking fuel as a result of limited switching and the low efficiency of employed stoves. Driven by higher incomes and a better developed infrastructure, urban areas experience a faster shift to modern fuels. Demand for LPG grows at an annual rate of 6% across sub-Saharan Africa, in sharp contrast with the phase out of kerosene and the limited uptake of electric cookstoves. The speed of evolutions is dissimilar across countries because of differences in economic growth and urbanization, and non-homogeneous starting points. The results demonstrate the vast size of the challenge to improve living conditions in sub-Saharan Africa and suggest that universal access by 2030, a target stated by several international organizations, is rather unrealistic.

Keywords: clean cooking, biomass, sub-Saharan Africa

Master of Science Thesis EGI_2017-0057-MSC

Clean cooking in sub-Saharan Africa: modeling the cooking fuel mix to 2050

Henri Casteleyn

Approved Examiner

Prof. Semida Silveira

Supervisor

Prof. Semida Silveira

Commissioner Contact person

(8)

Acknowledgements

The study was carried out during a 6-months internship at McKinsey Energy Insights in Poland.

I am grateful for the incredible learning opportunity and the daily coaching I received. In particular, the intellectual support of Bram Smeets, Matt Frank, Arjan Keizer, and Magdalena Wlodarczak proved of great value. Without their commitment, this master’s thesis would not stand where it is today. Furthermore, the broader Energy Insights team ensured that I had an unforgettable experience during this period.

I would like to thank my supervisor and examiner Prof. Semida Silveira, head of the Unit of Energy and Climate Studies at KTH, for the guidance and interesting discussions during the course of the thesis work.

Finally, I would like to thank my family for their continuous support and encouragement.

This thesis was typeset inL^ATEXusing a template provided bySunil Patel, who himself modified a template provided by Steven Gunn. The template was published under CC BY-NC-SA 3.0 and has been abridged and altered by the author.

(9)

List of Figures

3.1 Link between modeled cooking fuels and technologies . . . 20

3.2 High-level calculation logic for estimating future cooking fuel mix . . . 21

3.3 Methodology to distinguish between the use of improved wood stoves, traditional wood stoves, improved charcoal stoves, and traditional charcoal stoves . . . 22

3.4 IEA projections for cooking technology penetrations to 2040. . . 23

3.5 Calculation logic to project future urban fuel penetrations . . . 25

3.6 Example of general methodology to project fuel penetrations . . . 28

4.1 Regression analysis: household size versus GDP per capita . . . 34

4.2 Technical analysis for urban wood penetration. . . 36

4.3 Technical analysis for rural wood penetration . . . 38

4.4 Technical analysis for urban charcoal penetration . . . 40

4.5 Technical analysis for rural charcoal penetration . . . 42

4.6 Technical analysis for urban LPG/natural gas/biogas penetration . . . 44

4.7 Technical analysis for rural LPG/natural gas/biogas penetration . . . 46

4.8 Technical analysis for urban kerosene penetration . . . 48

4.9 Technical analysis for rural kerosene penetration . . . 50

4.10 Link between urban electricity access and penetration of electric cooking . . . 51

4.11 Technical analysis for urban electricity penetration . . . 52

5.1 Comparison of model results for fuel penetration in urban areas against IEA projections . . . 56

5.2 Comparison of model results for fuel penetration in rural areas against IEA projections . . . 57

5.3 Cooking technology mix in sub-Saharan Africa to 2050 . . . 58

5.4 Number of people in sub-Saharan Africa cooking with biomass by technology, to 2050 . . . 59

5.5 Number of people cooking with biomass in 2050, by region and technology . . . . 59

5.6 Cooking fuel mix in sub-Saharan Africa to 2050 . . . 60

5.7 Comparison of technology penetrations in urban and rural areas, by region, in 2016 and 2050. . . 61

5.8 Number of people cooking with LPG/natural gas/biogas in 2016 and 2050, by region and residence . . . 62

5.9 LPG/natural gas/biogas cooking energy demand in sub-Saharan Africa to 2050, by region . . . 62

5.10 Impact of variation in economic growth on change in cooking energy intensity . . 63

B.1 Projected urban fuel penetration for East Africa . . . 83

B.2 Projected urban fuel penetration for West Africa . . . 83

(14)

List of Figures

B.3 Projected urban fuel penetration for Central Africa . . . 83

B.4 Projected urban fuel penetration for Southern Africa . . . 83

B.5 Projected rural fuel penetration for East Africa . . . 84

B.6 Projected rural fuel penetration for West Africa. . . 84

B.7 Projected rural fuel penetration for Central Africa . . . 84

B.8 Projected rural fuel penetration for Southern Africa . . . 84

C.1 Projected urban technology penetration for East Africa . . . 87

C.2 Projected urban technology penetration for West Africa . . . 87

C.3 Projected urban technology penetration for Central Africa . . . 87

C.4 Projected urban technology penetration for Southern Africa . . . 87

C.5 Projected rural technology penetration for East Africa . . . 88

C.6 Projected rural technology penetration for West Africa . . . 88

C.7 Projected rural technology penetration for Central Africa . . . 88

C.8 Projected rural technology penetration for Southern Africa . . . 88

D.1 Projected urban cooking fuel mix for East Africa . . . 91

D.2 Projected urban cooking fuel mix for West Africa . . . 91

D.3 Projected urban cooking fuel mix for Central Africa . . . 91

D.4 Projected urban cooking fuel mix for Southern Africa . . . 91

D.5 Projected rural cooking fuel mix for East Africa. . . 92

D.6 Projected rural cooking fuel mix for West Africa . . . 92

D.7 Projected rural cooking fuel mix for Central Africa . . . 92

D.8 Projected rural cooking fuel mix for Southern Africa . . . 92

(15)

List of Tables

2.1 2030 scale-up targets for clean cookstoves in Ethiopia . . . 16

4.1 List of cooking technologies and their range of energy intensity, expressed in toe/year per household. . . 33

4.2 Definition of model rules for the wood penetration in cities . . . 37

4.3 Definition of model rules for the wood penetration in rural areas . . . 39

4.4 Definition of model rules for the charcoal penetration in urban areas . . . 41

4.5 Definition of model rules for the charcoal penetration in rural areas . . . 42

4.6 Definition of model rules for the LPG/natural gas/biogas penetration in urban areas . . . 45

4.7 Definition of model rules for the LPG/natural gas/biogas penetration in rural areas 45 4.8 Definition of model rules for the kerosene penetration in urban areas . . . 48

4.9 Definition of model rules for the kerosene penetration in rural areas. . . 49

4.10 Definition of model rules for the electricity penetration in urban and rural areas. 52 4.11 Summary of the exceptions to the general model rules . . . 53

A.1 List of modeled countries and their respective region in sub-Saharan Africa . . . 82

B.1 Relative fuel penetration in urban and rural areas in East Africa, as percent of urban/rural population cooking with a particular fuel . . . 84

B.2 Relative fuel penetration in urban and rural areas in West Africa, as percent of urban/rural population cooking with a particular fuel . . . 85

B.3 Relative fuel penetration in urban and rural areas in Central Africa, as percent of urban/rural population cooking with a particular fuel . . . 85

B.4 Relative fuel penetration in urban and rural areas in Southern Africa, as percent of urban/rural population cooking with a particular fuel . . . 85

C.1 Relative technology penetration in urban and rural areas in East Africa, as percent of urban/rural population cooking with a particular technology . . . 88

C.2 Relative technology penetration in urban and rural areas in West Africa, as percent of urban/rural population cooking with a particular technology . . . 89

C.3 Relative technology penetration in urban and rural areas in Central Africa, as percent of urban/rural population cooking with a particular technology . . . 89

C.4 Relative technology penetration in urban and rural areas in Southern Africa, as percent of urban/rural population cooking with a particular technology . . . 89

D.1 Relative fuel mix in urban and rural areas in East Africa, as percent of urban/rural cooking energy demand supplied by a particular fuel . . . 92

D.2 Relative fuel mix in urban and rural areas in West Africa, as percent of urban/rural cooking energy demand supplied by a particular fuel . . . 93

(16)

List of Tables

D.3 Relative fuel mix in urban and rural areas in Central Africa, as percent of urban/rural cooking energy demand supplied by a particular fuel . . . 93 D.4 Relative fuel mix in urban and rural areas in Southern Africa, as percent of

urban/rural cooking energy demand supplied by a particular fuel . . . 93

(17)

Abbreviations

IEA International Energy Agency GDP Gross Domestic Product

UN United Nations

GACC Global Alliance for Clean Cookstoves SDG Sustainable Development Goal SE4ALL Sustainable Energy for ALL

GHG GreenHouse Gas

BAU Business As Usual

GTF Global Tracking Framework CRGE Climate Resilient Green Economy WEO World Energy Outlook

MGI McKinsey Global Institute toe Tonnes of Oil Equivalent

(18)

(19)

Physical Constants

Million tonnes of oil equivalent¹ Mtoe = 41.868 · 10¹⁵ Joule

1As different types of crude oil have different calorific values, the exact value of one Mtoe is defined only by convention. Here, the value defined by the IEA is used.

(20)

(21)

Symbols

◦C Temperature Degrees Celsius

(22)

(23)

To my parents.

(24)

(25)

Chapter 1

Introduction

1.1 Background

Access to modern forms of energy is fundamental to meet basic human needs. A lack of access to affordable and reliable energy services results in poor healthcare and education, and a lagging economic growth (International Energy Agency, 2014). Moreover, increased energy access can trigger reduced poverty, improved health, increased productivity, a competitive economic system, and an overall accelerated economic growth (International Energy Agency,2011).

Therefore, it is of uttermost importance to be aware of the energy access situation and its future in developing countries.

Energy access is a broad concept and literature does not bring forward a single definition. The International Energy Agency (further abbreviated as IEA) summarizes and identifies four key attributes that are commonly found across definitions (International Energy Agency,2016b):

Access to a minimum level of electricity

Access to cooking fuels and technologies that are safe and sustainable, and pose only minimal harmful health effects

Access to modern energy for productive activities

Access to modern energy for public services

In addition to these important aspects, one needs to be aware that other issues are connected to energy access as well, including the safety, affordability, reliability, and adequacy of the energy system. Due to data constraints however, most literature pays attention to the topics of access to electricity and modern ways of cooking (International Energy Agency,2016b).

When taking a closer look at sub-Saharan Africa, the global epicenter of energy poverty, one observes that the current state of affairs related to these two aspects of energy access is far from satisfactory. Firstly, 620 million people in the region do not have access to electricity (Interna- tional Energy Agency, 2014). Although a very stunning situation, this master’s thesis focuses on the second aspect of energy poverty, namely access to clean cooking facilities. In sub-Saharan Africa, 792 million people use traditional biomass for cooking (International Energy Agency,

(26)

Chapter 1. Introduction

2016a). Traditional biomass cooking refers to the use of highly inefficient three-stone fires fueled with fuelwood, charcoal, dung, agricultural waste, and other waste. With population growth outgrowing the number of people gaining access, the absolute number of people without access to clean cooking technologies increased during the period 2012-2014. Sub-Saharan Africa repre- sented 33% of the global population without access to clean fuels and technologies for cooking in 2014 (Sustainable Energy for All, 2017a). Even though the number of people using solid biomass for cooking is larger in developing Asia, the share of the population in sub-Saharan Africa is bigger, namely 80% in sub-Saharan Africa, compared to 50% in developing Asia (In- ternational Energy Agency, 2015). In 23 countries, more than 90% of population cooks with biomass, driven by extreme poverty (Lambe et al., 2015). Moreover, the pace of improvement has been extremely slow since 1990 (Sustainable Energy for All,2017a). The described situation is problematic since the high penetration of traditional biomass cooking brings about a variety of social, environmental, and economic hazards.

Traditional biomass cooking causes severe indoor air pollution, the effects of which are broadly described in literature. The incomplete combustion of biomass is accompanied by the release of particulate matter, carbon monoxide, and other products. The inhalation of these harmful pol- lutants causes many diseases, among others respiratory infections (e.g. acute lower respiratory infections, chronic obstructive pulmonary disease), cancer, vascular diseases, cataract, and low birth weights. Estimates show that 600,000 people die each year as a result of the exposure to biomass smoke in sub-Saharan Africa, with girls having a higher chance to die from hazardous fumes than from malaria or malnutrition (Lambe et al.,2015; The World Bank, 2014;United Nations Environment Programme,2017).

The vast use of biomass, both fuelwood and charcoal, for cooking has a negative effect on the environment, causing forest degradation and biodiversity loss (Rivard and Reay,2012;Beyene and Koch,2013). With a depletion rate that is larger than the rate of growth, East Africa can be considered the center of unsustainable biomass use. In addition, the burning of fuelwood is expected to make up 5.6% of Africa’s projected business as usual (BAU) greenhouse gas emissions (The World Bank,2014).

As biomass is often sourced by gathering the cooking fuel, productive time is lost. The bur- den is typically placed on women and children, who are responsible for this task. On average, women spend 2.1 hours per day gathering wood for cooking, with a maximum of 5 hours in Sierra Leone (Lambe et al.,2015). After fuel is gathered, women spend another couple of hours cooking with the traditional three-stone fires. These practices prevent women and children to generate income for the household, pursue an education, or fulfill other household tasks. In addition to the waste of productive time, women are vulnerable to physical and sexual violence outside their communities (United Nations Environment Programme,2017).

The above mentioned negative effects on society cause a substantial opportunity cost of USD58.2 billion per year, which represents 4.4% of gross domestic product, further abbreviated as GDP (high estimate values) (Lambe et al., 2015). This high value suggests that cost-savings could be achieved when citizens switch to modern cooking technologies, a statement supported in the Climate Resilient Green Economy plan of Ethiopia (Federal Democratic Republic of Ethiopia, 2011).

(27)

Section 1.1 Background

The gravity of this topic is demonstrated by the attention paid by influential international bodies, among others the IEA, the United Nations (UN), the G7, and the G20. Moreover, multiple national governments started issuing policies on clean cooking facilities, even though the focus of governments in the past decades was on electricity access, in particular grid extensions.

The concern about the issue of energy access is clearly growing and steps are taken to transform the existing situation. In the following paragraphs, some examples of meaningful international and national actions are presented.

The UN has been actively working on the issue of energy access for years. In 2010, the Global Alliance for Clean Cookstoves (GACC) was established, with an aim to provide 100 million households with access to clean and efficient stoves by 2020 (Global Alliance for Clean Cook- stoves, 2017). The UN also declared 2012 the International Year of Sustainable Energy for All (International Energy Agency,2011). In 2015, the UN adopted a goal on energy in the Sustain- able Development Goals (SDGs), namely to ensure access to affordable, reliable, sustainable and modern energy for all. A cornerstone of this goal is to provide access to modern energy in developing regions, including sub-Saharan Africa (United Nations, 2017). Another noteworthy action of the UN is the Sustainable Energy for All (SE4ALL) initiative. This initiative targets universal access to modern energy services, improved efficiency, and an increased use of renew- ables (Sustainable Energy for All,2017b). Very recently, the World Bank and the IEA released the Global Tracking Framework, the key take-away being that the progress on the objectives for 2030 is not fast enough (Sustainable Energy for All, 2017a). Furthermore, the G7 committed to accelerate access to renewable energy in Africa and other regions, and the G20 launched the Energy Access Action Plan for sub-Saharan Africa and developed a plan for Asia and the Pacific as well (International Energy Agency,2015,2016a). In 2015, the African Development Bank launched A New Deal on Energy for Africa, aiming to achieve universal access to energy in Africa by 2025 (International Energy Agency,2016a).

Other noteworthy initiatives that touch upon access to clean cooking are GIZs Energizing De- velopment Program, the West African Clean Cooking Alliance by the Economic Community of West Africa States, Energy+ hosted by Norway, United Kingdom and others, and the EnDev program by Germany, Norway and others (International Energy Agency,2014).

As mentioned, also national governments are committing to the case of clean cooking. Some countries with recent policies on clean cooking are (International Energy Agency,2014,2013):

Indonesia: successfully implemented a kerosene-to-LPG program in 2007 and continues to make efforts to promote clean cooking

Nigeria: set a target to provide 20 million households with access to clean cooking facilities by 2020

Senegal: implemented a national LPG program

Ghana: implemented a national LPG program

Cote d’Ivoire: implemented a national LPG program

Kenya: plans to eliminate kerosene by 2022

(28)

Ethiopia: aimed to disseminate 9 million improved cookstoves by 2015 and develops Biomass Energy Strategy Plans in collaboration with the European Union Energy Ini- tiative Partnership Dialogue Facility

Rwanda: plans to reduce the share of bioenergy in primary energy demand to 50% by 2020

Mozambique: develops Biomass Energy Strategy Plans in collaboration with the European Union Energy Initiative Partnership Dialogue Facility

Liberia: develops Biomass Energy Strategy Plans in collaboration with the European Union Energy Initiative Partnership Dialogue Facility

Sierra Leone: develops Biomass Energy Strategy Plans in collaboration with the European Union Energy Initiative Partnership Dialogue Facility

Also, multiple countries, such as Ethiopia, Nigeria, Kenya, Central African Republic, Burkina Faso etc. state in their Intended Nationally Determined Contributions that the promotion of clean and improved cookstoves is an action area to increase access to energy among citizens and cut greenhouse gas (GHG) emissions (United Nations Environment Programme,2017;Federal Democratic Republic of Ethiopia,2015;Ministry of Environment and Natural Resources Kenya, 2015;Government of Nigeria,2015).

A last example of the growing attention is the emerging involvement of multinationals, illus- trated by partnerships between Philips and the Industrial Development Corporation of South Africa, and between the firms General Electric, Burn Manufacturing, and the US Overseas Pri- vate Investment Corporation (International Energy Agency,2013).

The previous paragraphs clearly described the issues related to a lack of access to clean cooking facilities and the growing attention paid by national and international organizations. Relevant literature has mapped the negative effects of traditional biomass cooking and does touch upon drivers for cooking fuel choice in developing regions. However, research papers do not provide a comprehensive, quantitative, and transparent outlook for the future cooking fuel mix in sub- Saharan countries, a research gap to be filled in order to develop a view on the future situation, and identify opportunities for increased efforts.

1.2 Purpose of the master’s thesis

Starting from the observation that cooking energy consumption averages approximately 80% of residential energy use in sub-Saharan Africa, a good understanding of this service is essential to assess the dynamics in the overall energy system (International Energy Agency, 2014). More- over, modeling the cooking energy consumption is desirable as the model insights can be used to guide policy decisions. Governments and international aid organizations can benefit from having a BAU perspective on the future cooking fuel mix to identify priorities, which is what this report aims to deliver.

This master’s thesis report aims to fill a gap in the existing research literature by providing a driver-based quantitative outlook for the cooking fuel mix in urban, rural, and national areas

(29)

Section 1.3 Limitations

for 45 countries in sub-Saharan Africa, the global epicenter of energy poverty. It is crucial to understand that differences in used cooking fuel among cities and rural villages are considerable. To the author’s best knowledge, transparent estimates of the future fuel mix on an urban, rural, and/or national level are not available in scientific literature for this geographical scope.

By investigating past evolutions and historical datasets, the author hopes to develop a BAU scenario to 2050.

To this extent, the following research question is defined:

What is a history-inspired pathway for the evolution of the biomass dominated cooking fuel mix in sub-Saharan countries to 2050?

The research question is further broken down into three objectives:

1. Develop a view on the drivers behind the fuel switch in residential cooking in sub-Saharan Africa

2. Create a quantitative model that captures the changes in the use of traditional biomass for cooking, based on the uptake of a number of cooking technologies, in order to establish a perspective on the future cooking fuel mix on urban and rural level in 45 sub-Saharan countries

3. Explore the connection between the clean cooking evolution and change in cooking energy demand intensity

In order to be able to formulate an answer to the above stated research question and achieve the objectives, a quantitative model was developed. The model estimates the future uptake of several cooking technologies in 45 sub-Saharan countries (see appendixAfor a list of countries).

Projections for urban and rural areas are formulated separately to account for significantly different dynamics in the two residences. Examples of the different circumstances are the difference in level of income and availability of substitute fuels for biomass. The model rules are based on regression analyses (penetration of cooking fuel against income), insights from literature, and expert interviews. Once projections for the diffusion of technologies are obtained, the fuel mix can be calculated, given assumptions for the energy intensity of each cooking technology.

1.3 Limitations

However, rather than pretending to predict the future with certainty, it is acknowledged that the produced model is per definition an approximation of the reality. Therefore, it is useful to understand the limitations associated with its use.

This master’s thesis project aims to provide a driver-based outlook for the cooking fuel mix in sub-Saharan Africa to 2050. Making long-term projections is challenging because a long time horizon increases the uncertainty of projections. As one tries to capture fuel dynamics, it is possible to miss the impact of certain disruptions, such as the uptake of solar cooking or an accelerated development of the power system.

(30)

Another inherent limitation of this project is the strong connection between the projections and historical evolutions, as observed in the available data. The outlook for the future uptake of technologies is largely based on the speed of trends in the past. As a result, the obtained model results display a BAU case, without fully accounting for drastic changes in policies and other developments. When observing the model results, it is crucial to be aware that the future situation might very well be totally different from what is presented.

As described in chapter2, literature shows that cooking stove/fuel choice is influenced by many determinants, broadly categorized as socioeconomic variables, cultural or behavioral habits, product-specific attributes, and external factors. The developed model focuses on economic development, population growth, and urbanization as drivers. The rationale behind this approach is given by severe data constraints and the belief that the above mentioned macroeconomic variables are the key drivers of past and future evolutions in cooking energy consumption.

The model is based on the assumption that households use only one fuel for cooking. How- ever, in reality, households often use multiple fuels as the transition to clean cooking fuels is accompanied by the simultaneous use of biomass and a more modern cooking fuel. This phenomenon is known as fuel stacking and provides households with flexibility in case of disturbed fuel supply or high prices. Due to lack of data and complexity of the model, this phenomenon is not captured by the presented model.

The thesis work has been characterized by severe data limitations. Residential energy consumption data is often hard to find because of privacy issues that limit the availability of data, and the costs of detailed energy metering. This is especially the case for sub-Saharan Africa, which is known for fairly poor data quality. As a result, the model often employs strong assumptions and pragmatic actions. The data that was used for representing the cooking fuel penetration is based on surveys. Inherently, surveys are not always a good representation of reality because of mistakes during data collection, respondents that are not representative for the broader population, or the poor design of the survey. Several different surveys were used to obtain data for as much countries as possible. The scarcity of data called for a pragmatic approach, resulting in different countries having a different base year for the projections.

1.4 Organization of report

Chapter 2 further elaborates on the research gap by describing relevant work that has been produced by others. Both academic literature and publications from international organizations are covered to give the reader a good idea of the overall clean cooking environment. In chapter 3, the methodology is discussed in a brief and concise way. Chapter4covers in detail the model rules that determine the results of the modeling exercise. Some specific aspects of the results are described and presented graphically in chapter 5. This chapter also contains a benchmarking of the results. Due to the vast size of the model, results are mostly presented for urban and rural areas on a regional level. The report ends with a discussion, and a description of potential future work in chapter 6.

(31)

Chapter 2

Frame of reference

This chapter aims to provide the reader with a synopsis of literature on clean cooking in developing regions. In the first two sections, academic research papers are covered. In the third section, some attention is reserved for the efforts of national and international organizations to alter the current dominance of traditional biomass cooking.

2.1 Factors influencing cooking energy choice

A number of studies have been conducted to understand the factors that affect cooking stove choices and fuel consumption patterns. Even though determining or quantifying determinants of fuel choice is not the aim of this thesis, this research is discussed as it can provide a view on the drivers behind increased access to clean cooking facilities, which is one of the objectives of this report. The starting point of these studies is mostly a conducted survey, to which statistical manipulations are applied. Understanding the determinants of the uptake of improved or clean cookstoves is crucial for the design of effective policies that aim to expand access to clean cooking facilities in a cost-efficient way. One can group the identified variables in four groups, namely socioeconomic variables, cultural or behavioral habits, product-specific attributes, and external factors.

2.1.1 Socioeconomic variables

This group of determinants characterizes the households or individuals deciding on the stove type that will be used for cooking activities. Examples of factors include income, education, and the size of households. This group of factors has received most attention in literature.

The influence of income on the cooking fuel choice is a widely discussed topic. Takama et al.

(2012) provides prove for the rationale that households switch to more advanced and cleaner cooking technologies when their wealth level increases. This concept is known as the energy ladder. Other studies confirm the influence of income on fuel choices, provide empirical evidence, and state that rising income is a key driver to climb the energy ladder (Barnes et al., 2010;

(32)

Chapter 2. Frame of reference

Yu-Ting Lee, 2013; Ouedraogo, 2006; Rehfuess et al., 2010; Beyene and Koch, 2013; Karimu, 2015;Rahut et al.,2016).

A different perspective on the fuel switching behavior of households is presented by the fuel stacking model, which states that households make use of a portfolio of techniques and fuels for cooking (van der Kroon et al.,2013;Masera et al., 2000). The use of multiple fuels allows households to be flexible and react to supply shortages, price fluctuations, or irregular income flows. Income is still accepted to be a key determinant of fuel choice, but the impact of other influences may not be overseen. Therefore, the authors describe the fuel choice in a household decision environment which includes variables such as the age of occupants and their education level.

Several papers agree that education is a significant factor in the fuel choice decision process, making improved education an important lever to foster a transition towards modern cooking facilities (Heltberg,2005;Rao and Reddy,2007;Yu-Ting Lee,2013;Ouedraogo,2006;Rehfuess et al., 2010; Rahut et al., 2016; Nlom and Karimov, 2015). These sources also recognize the influence of human capital.

A number of other factors that impact the decision of cooking fuel were identified in literature. Examples include occupation of household members, household size, age of household members, availability of public infrastructure, and housing standards (Rao and Reddy, 2007;

Yu-Ting Lee,2013;Ouedraogo,2006;Nlom and Karimov,2015).

Note that it becomes clear why the cooking situation in cities is distinct from the one in rural villages. As factors such as level of income, education, occupation of household members, availability of infrastructure, and housing standards differ among the two residences, the penetration of modern fuels versus traditional biomass is different. This observation supports the decision to model the fuel mix in both residences separately.

2.1.2 Cultural and behavioral habits

Less academic work has focused on cultural or behavioral habits that affect the choice for a particular cooking fuel. However, household characteristics such as the frequency of cooking certain meals (e.g. rice), the position of women, and the religion of household members have an impact on the used cooking technologies (van der Kroon et al.,2013;Ouedraogo,2006;Masera et al.,2000;Heltberg,2005;Urmee and Gyamfi,2014;Ruijven et al.,2011).

2.1.3 Product-specific attributes

As opposed to previously discussed considerations, product-specific attributes cover a rather different set of variables with an influence on the cooking fuel choice decision, namely the characteristics of the cooking stoves and the associated fuels. This research focuses more on the elements to which consumers pay attention and the behavior of household consumers in switching away from traditional biomass cooking.

(33)

Section 2.1 Factors influencing cooking energy choice

One study which reflects on stove attributes is Takama et al. (2012). The authors researched the importance of several product-specific characteristics in influencing the stove choice, aiming to understand how households make the trade-off between different attributes. To do so, a choice experiment was implemented, which is a behavior modeling technique aiming to assess the strength of various influences. Through a survey of 200 households in July 2008 in Addis Ababa, Ethiopia, the preferences of consumers were identified as respondents chose the preferred option from a series of products. Afterwards, the value of the attributes was estimated to convey the relative strength of product features in influencing households’ choices. The report covers three cooking fuels, namely wood, kerosene, and ethanol, and four stove attributes; stove price, monthly usage costs, indoor smoke emission level, and safety. Moreover, by comparing responses from three different wealth groups, the impact of wealth could be evaluated. The investigated product-specific factors significantly affect stove and fuel choices and differences across wealth groups are considerable.

As expected, the research paper finds that stove price and usage cost reduce the utility of a cookstove, except for high-wealth groups, suggesting that other factors such as convenience play a significant role for those households. The results show that the low-wealth group is more sensitive to the stove price than the middle-wealth group, illustrating that as the wealth level of a household increases, the stove price becomes less important in the decision process. As with stove price, the low-wealth group is most sensitive to usage cost, which is affected by the efficiency of a stove and the fuel price. In comparison to low-wealth and middle-wealth groups, the high-wealth group shows the strongest aversion for smoke, illustrating that rich households pay attention to user convenience. Finally, the safety characteristics of a stove prove to be a determinant of fuel choice across wealth groups, with higher wealth groups willing to pay more for increased safety than lower wealth groups. The results of this research point to the energy ladder model, in which households shift to more advanced cooking options when income increases. Wealthy consumers prefer clean fuels and stoves and are willing to pay for it.

Another important influence is the price level of fuels. Higher fuel prices lead to a reduced consumption of that particular fuel and switching to substitutes, as described by Yu-Ting Lee (2013);van der Kroon et al.(2013);Nlom and Karimov(2015).

2.1.4 External factors

The last broad category of determinants can be called external factors, as these are not determined by household characteristics or directly influenced by cookstove suppliers. Often cited determinants include fuel availability, reliability of supply, policies or clean cookstove initiatives (if well targeted and executed), and the market structure for clean cooking technologies (The World Bank,2013;Bacon et al.,2010;Urmee and Gyamfi,2014;Mainali et al.,2012).

(34)

2.2 Modeling cooking fuel mix

Several studies have aimed to quantify the process of cooking fuel choice. This master’s thesis aims to produce a perspective on the future cooking fuel mix in sub-Saharan Africa. In order to do so, the decision was made to build a bottom-up model that quantifies important trends as a function of input parameters. In this section, other relevant academic work related to the research objective is discussed.

2.2.1 Two distinct types of models

Swan and Ugursal(2009) provides a comprehensive overview of the different modeling techniques to determine end-use energy consumption in the residential sector. The authors distinguish between top-down and bottom-up models, each characterized by advantages and drawbacks. The paper provides a review of each technique and pays attention to strengths, weaknesses, and purposes. The two most critical differences between the two archetypes of models are the depth of required input data and the range of modeled scenarios. Top-down approaches are easier to develop because the data requirements are less stringent. On the other hand, these models do not allow to account for technological breakthroughs or shocks. Bottom-up models, being inherently more detailed, allow for a deeper investigation of consumer behavior and the impact of new technologies. The authors exemplify their statements by briefly describing models that are present in literature.

Even though the model which was developed to support this master’s thesis is largely based on macroeconomic variables, one can categorize the model as bottom-up. The reasons for this are the detailed consideration of the different cooking technologies and their characteristics from an end-user perspective, the assessment of adoptions for individual fuels, and the distinction between urban and rural areas. Moreover, rather than regressing the national cooking energy demand, combining projected adoption rates for several technologies and technology ratings allows to disaggregate cooking energy consumption among fuels and technologies, thereby also paying attention to consumer behavior.

2.2.2 Quantifying cooking fuel choice

Discrete choice modeling is a commonly used technique to analyze fuel choices (Takama et al., 2012, 2011). A stated preference survey is conducted to understand consumer preferences as they choose the preferred option from a number of cooking technologies. In a next step, consumer choices are correlated with characteristics of correspondents to quantify the influence of attributes and assess the trade-offs in choice of cooking stoves and fuels. Takama et al.

(2011) employs the estimated attribute coefficients to simulate the impact of a change in stove attributes on the stove market share. This allows to understand the potential effect of policy measures. However, this approach cannot be used to develop a plausible perspective on the future (Mainali et al.,2012;Takama et al.,2011).

Another approach to model the cooking fuel choice is discussed by Rehfuess et al. (2010).

(35)

Section 2.2 Modeling cooking fuel mix

The authors aim to quantify the influence of household-level factors and area-level characteristics, including differences between communities, by employing Bayesian hierarchical and spatial modeling with a geographical focus on Benin, Kenya, and Ethiopia. As found by the authors, the heterogeneity between areas may point to the considerable influence of contextual effects and supply-side limitations. As one will understand later, this master’s thesis accounts for these contextual factors by applying the slope of the calculated regression curves to the starting point of each country, rather than applying the equation of the regression curve to all countries. As this work of Rehfuess et al. (2010) concentrates efforts on determining determinants of solid fuel use in a set of selected countries, it does not provide insights for the future evolution of the cooking fuel mix.

2.2.3 Projecting cooking fuel mix

As the transition from traditional to modern fuels is a specific dynamic in developing regions, the literature study focuses on quantitative projections for developing regions. Several academic papers contain projections for the composition of cooking energy demand by fuel in developing regions. Often, relevant reports cover the total residential energy sector, of which cooking constitutes an important share, focus on one particular country, and opt for a shorter projection period compared to what was done for this report. Moreover, performed analyses are often characterized by the application of a cost-optimization model, thereby assuming rational agents with perfect foresight. A point of similarity lies in the identification of model drivers, often being economic growth, urbanization, and population expansion. The author is not aware of a comprehensive research paper on the future cooking fuel mix to 2050 for the whole geographical region of sub-Saharan Africa. Furthermore, projections for specific sub-Saharan countries proved almost impossible to find, except for South Africa. Therefore, one can state that this lack of knowledge presents a clear research gap, which this master’s thesis hopes to address.

This report aims to deliver a BAU perspective on the future cooking fuel mix in urban and rural areas of 45 sub-Saharan countries to 2050. In the following paragraphs, relevant literature is commented to give the reader an idea of what work has been performed.

Howells et al. (2005) aim to compute the optimal energy system in a rural village in South Africa, namely the village of Nkweletsheni. For this purpose, the authors developed an extension of the MARKAL energy modeling tool, the TIMES model. TIMES is a linear optimization model which identifies the least-cost supply options, from a lifetime perspective. While taking into account details on technology cost and performance, the model assumes perfect foresight by energy consumers, an assumption which can be questioned.

Rather than focusing on cooking fuel mix only, the model calculates the future energy consumption for six energy services; cooking, space heating, water heating, lighting, refrigeration, other (radio, TV, etc.). A set of appliances is assumed to satisfy the demand for the six energy services. The demand for considered fuels (biomass, coal, electricity, LPG, paraffin, candle way) is estimated from the quantities of final energy required by the set of appliances. Five scenarios were developed, trying to capture different levels of access to electricity. Two differentiating aspects of the approach are the disaggregation of energy consumption according to the time of day through load curves, and the consideration of appliances that can supply more than one

(36)

energy service (e.g. a wood brazier serves as a source for cooking, space heating, and water heating). The described approach allowed to obtain a holistic view on energy consumption in rural villages, but lacks a clear focus on cooking. The work is also characterized by a narrow geographical scope and limited predictive time frame (2003-2018).

The base case scenario projects biomass remaining a dominant cooking fuel to 2018. LPG fulfills the role of secondary or back-up fuel when wood is not available, supplying only a very small part of the cooking energy demand. According to the authors, LPG is not economically competitive when wood is present.

Ruijven et al.(2011) presents a bottom-up model for residential energy use in developing countries, making a distinction between five end-uses, namely cooking, water heating, space heating, lighting, and appliances. The model, which was designed to be incorporated in the global energy system model TIMER, determines energy use and fuel choice.

The authors applied the developed model to India because of good data availability. The identified drivers of residential energy use are population growth, urbanization, household ex- penditure, household size, and temperature changes. The analysis employs correlations from econometric studies and regressions. In a first step, the level of useful energy demand for the various end-uses is determined. The second step constitutes of assigning fuels to meet this demand, based on economic considerations and fuel availability. To do so for cooking fuel choice, the authors used a capital vintage model, which determines the market shares of stoves from assumptions on investment and depreciation rates. The annual perceived cost of technologies is crucial for this purpose, but the model also accounts for up-front investment challenges and household preferences not captured by economics. As a result of this modeling approach, the switch from one fuel to another follows the concept of the energy ladder, which states that households opt for more efficient and modern fuels as they get richer.

The developed model is used to investigate the impact of the level of rural electrification and income distribution. The authors conclude that an equal income distribution and adequate rural electrification play an enhancing role in the diffusion of commercial fuels. However, projections show that rural areas will continue to rely mainly on the use of traditional fuels to 2050 and that biomass will make up between 34% and 45% of total national final residential energy use.

Vassilis et al. (2012) applies an extension of the model, now called REMG, to analyze future developments of residential energy use in India, China, South East Asia, South Africa, and Brazil. The projection timescope is reduced from 2050 to 2030. The authors state that over time, non-cooking end-uses, including space heating, cooling, and appliances, become more important in the total residential energy demand. At the same time, the shift towards modern fuels only develops slowly in urban and rural areas, but more explicitly in rural areas where biomass still shows a strong dominance in 2030 across all five regions. However, the per capita cooking energy demand decreases to 2030 as a result of moderate fuel switching and increases in efficiency for any given fuel. Even though this report does not account for improvements in efficiency for any given technology, the switch to modern and improved cookstoves is also expected to display a reduction in cooking energy intensity.

(37)

Section 2.2 Modeling cooking fuel mix

Another paper which shows efforts to estimate the future cooking fuel and stove choices in developing countries is presented by Mainali et al. (2012). Focusing on China, the authors project the cooking fuel mix to 2030, thereby making use of the MESSAGE-Access modeling tool. Similar to the previously mentioned MARKAL model, MESSAGE is an optimization model which aims to identify least-cost energy supply options under predefined constraints. The model is mostly used to construct long-term projections (up to 2100). The MESSAGE-Access model is an extension of the MESSAGE model and covers the residential sector with a focus on cooking and electrification related demand.

The analysis of cooking fuels was performed separately for urban and rural areas, based on the fact that energy consumption patterns differ, as mentioned before. To account for the distribution of wealth across the population, households were sub-divided into five income quintile groups. Different quintiles are characterized by different discount rates, technology preferences, inconvenience costs, and budgets constraints.

The following variables are treated as inputs: per capita final energy consumption in rural and urban quintiles in the base year (2005), population growth, income growth and distribution, and cost of stove technologies. The prices of fossil fuel energy sources are determined endogenously within the model. The model assumes that per capita solid fuel (biomass and coal) use cannot increase over time. Also, per capita modern fuel consumption is not allowed to decrease from one year to another.

The total cost of cooking with a particular stove is calculated as the sum of fuel cost, annual- ized cooking technology cost, and inconvenience cost, which aims to capture the non-monetary disutility of using traditional solid fuels versus modern cooking technologies. Thus, the outcome of the optimization process represents the least-cost solution from a lifetime perspective. It is noted that a similar, detailed determination of the various costs of cookstoves is not required for this report because the model is not based on a cost-optimization algorithm.

In addition to a base case perspective, the authors developed a number of scenarios, reflecting different levels of credit access to cover the upfront stove costs and support mechanisms to reduce fuel cost. Biomass is expected to remain a dominant fuel to 2030 in rural areas in the BAU scenario. The authors predict that 24% of rural population and 17% of urban population will rely on solid fuels (biomass and coal) in 2030 under BAU scenario. However, a clear shift towards LPG in both urban and rural areas can be observed.

A similar approach was followed by Ekholm et al. (2010). The objective of the study was to analyze the effectiveness of policies that aim to improve the penetration of modern cooking facilities in India.

The authors developed a choice model based on economic considerations and practical determinants such as consumer preferences, while maximizing end-user utility. As in the previously discussed model, heterogeneity of households was taken into account by considering fuel choices separately for populations living in rural and urban areas and by differentiating between income groups. The choice model was then implemented within the MESSAGE framework to forecast the diffusion of different fuels to 2020.

(38)

The baseline model results display the domination of biomass in rural areas. According to these projections, the number of people using traditional biomass would increase with almost 100 million in the period 2000-2020. It is remarkable that the authors project a decrease in the use of LPG, stating that existing LPG stoves are replaced with biomass stoves after their lifetime.

The baseline scenario is supplemented with additional scenarios, with the purpose to assess the impact of policies, including fuel subsidies and improved financing opportunities. This analysis pointed out that high investment costs are a major barrier for the adoption of modern fuels, especially LPG. Therefore, the authors conclude that fuel subsidies should be combined with improved access to financing to cover the appliance investment.

2.3 National and international organizations

Several national and international organizations performed studies on cooking patterns in sub- Saharan Africa. However, most of this work does not include a quantitative view on the future.

The produced publications often focus on past evolutions, success stories of selected countries, and the design of strategies.

The UN is one of the major institutions that does efforts to improve the uptake of clean cooking technologies. The SDGs, introduced in 2012 as replacement of the Millennium Development Goals, aim to meet global environmental, political, and economic challenges. The 7^th SDG is focused on affordable and clean energy. Apart from targeting an increased share of renewable energy in the global energy mix, a doubling of the improvement rate in energy efficiency, en- hanced international cooperation, and an expansion of energy infrastructure, the objective is to ensure universal access to affordable, reliable, and modern energy services (United Nations, 2017). In the reports that assist this ambitious goal, insights are provided in how households and communities are supported to complete the switch away from traditional biomass cooking.

Examples of activities are developing sustainable charcoal value chains and realizing improved cookstove projects. Figures on the use of funds and the number of projects are also published.

The project portfolio covers almost 100 projects on energy access in 70 countries and the UN is involved in USD410 million of grant financing and a further USD1.5 billion of co-financing.

Even though the largest portion of supported projects is related to renewable energy, 19 projects in 2015 were addressing the uptake of energy efficient cookstoves (United Nations Development Programme, 2016; United Nations Development Programme Global Environmental Finance Unit, 2015). 12 of these projects were located in Africa, namely in Burkina Faso, Ethiopia, Guinea, Guinea Bissau, Kenya, Liberia, Mali, Niger, Tanzania, and Uganda (United Nations Development Programme Global Environmental Finance Unit,2015). However, the published reports do not include projections for the uptake of cooking technologies or cooking fuel mix.

The UN Foundation hosts a partnership called The Global Alliance for Clean Cookstoves (GACC) (Global Alliance for Clean Cookstoves,2017). This initiative aims to create a strong market for clean and improved cooking solutions and set a target of 100 million households adopting clean and efficient cookstoves by 2020. The GACC concentrates its projects in eight

(39)

Section 2.3 National and international organizations

countries, four of which in sub-Saharan Africa, namely Ghana, Kenya, Nigeria, and Uganda.

GACC reporting mainly covers insights on stove/fuel production and distribution, progress, and market assessments. The 2016 Progress report is a good illustration of a publication which tracks the progress towards cleaner cookstoves and fuels (Global Alliance for Clean Cookstoves,2016).

This documents includes a high-level historical estimate and forecast to 2020 of the cumulative global stove distribution, making a distinction between clean or improved cookstoves and traditional biomass stoves. According to their analysis, 250 million improved or clean cookstoves will be distributed in the period 2010-2020. This perspective does not differentiate between regions or fuels, and thus contains no information on fuel mix in sub-Saharan Africa, let alone on country-level. Another key component of the report is a more detailed investigation of the latest progress and improvements in the eight focus countries. Other relevant publications mainly aim to develop a more comprehensive market understanding and therefore assess the current market structure, consumer preferences, and active stakeholders. Examples of work in this area include a market assessment for Nigeria, an LPG market assessment for Kenya which provides a BAU scenario for the LPG penetration to 2020, and a consumer segmentation analysis for Ghana (Accenture, 2011; Dalberg, 2013; AddedValue, 2014). Again, no comprehensive, quantitative outlook for the future situation is provided in these documents as the focus is on enhancing market intelligence and characterizing current cooking habits. A last type of document that is developed under guidance of the GACC is the country action plan, which summarizes the priority interventions and opportunities on a country level for selected countries and sometimes includes targets for clean cookstove penetrations. Examples include the Kenya Country Action Plan and the Ghana Country Action Plan (Ghana Energy Commission, 2013; Global Alliance for Clean Cookstoves,2013).

Another global platform on energy access in Sustainable Energy for All (SE4ALL) (Sustain- able Energy for All,2017b). This organization aims to achieve universal access to energy while limiting the impact on the environment, keeping global warming well below 2°C. In this regard, SE4ALL defined three objectives for 2030, namely universal access to modern energy services, doubling the share of renewable energy in the global energy mix, and doubling the global rate of improvement in energy efficiency. To achieve these targets, high impact opportunities were identified. One of the 11 Action Areas is the universal adoption of clean cooking solutions, which calls for a close collaboration with the GACC, the Global LPG Partnership, and the World LP Gas Association. Moreover, national governments were encouraged to complete gap analyses to determine key challenges and opportunities in achieving the three SE4ALL objectives. Seven countries, all of them in Africa, already completed the next step in the process, which is to draft an Action Agenda. These national Action Agendas elaborate on a long-term vision and should serve as a starting point for donor aid. On a regular base, the SE4ALL organization assesses past developments at the national, regional, and international level to clarify whether the move- ment is on track to reach the objectives and identify where more action is required. The tool that is used for this purpose is called the Global Tracking Framework (GTF) and an update is produced bi-annually. Each edition delivers insights on trends in energy access, renewable energy, and energy efficiency. The GTF 2017 represents the latest thinking and claims that 850 million people in sub-Saharan Africa lacked access to clean cooking in 2014, representing 74% of the population (Sustainable Energy for All,2017a). To obtain universal access to clean cooking facilities by 2030, the SE4ALL states that progress in the remaining years to 2030 should be five times faster than the evolution in 2012-14. In their reports, the SE4ALL does not provide

(40)

Table 2.1: 2030 scale-up targets for clean cookstoves in Ethiopia, as estimated in Federal Democratic Republic of Ethiopia(2011)

Improved fuelwood LPG Biogas Electric

Urban areas 5% 5% 1% 61%

Rural areas 80% 0% 5% 5%

own estimates of likely future outcomes for fuel penetrations. Instead, the organization uses IEA projections.

In contrast to previously mentioned organizations, the Global Green Growth Institute, which was established in 2012 and promotes a strong and sustainable economic development, does not demonstrate a clear focus on clean cooking (Global Green Growth Institute, 2017). However, this topic is part of their expertise as the organization supports governments of developing countries to strive for a green growth path, which often involves efforts to cap the inefficient biomass use for cooking. An example is provided by the involvement in developing the Climate Resilient Green Economy plan (CRGE) of Ethiopia (Federal Democratic Republic of Ethiopia, 2011).

The CRGE is Ethiopia’s strategy for addressing both climate change adaptation and mitigation objectives, while realizing its ambition of reaching middle-income status by 2025. One of the four strategy pillars is protection and rehabilitation of forests for their economic and ecosystem services. An important goal is to reduce demand for fuelwood via the dissemination of efficient stoves and alternative fuels. Four initiatives for fast-track implementation were selected, one of them being the large-scale promotion of advanced rural cooking technologies. The rationale behind this stress on the switch away from traditional biomass cooking is given by the fact that the forestry sector accounts for 51% of the GHG abatement potential by 2030, the most important lever being an increased uptake of fuelwood efficient cookstoves. According to the report, switching to more efficient technologies and modern cooking fuels has a negative abatement cost, meaning that the benefits outweigh the costs. An exception to this general observation is provided by LPG stoves. With this particularity in mind, ambitious scale-up targets for cooking technologies in urban and rural areas by 2030 were drafted, as visible in table2.1.

Compared to other well-known international bodies, the IEA often follows a more quantitative approach, rather than describing visible trends and policies. The IEA has been paying attention to the topic of energy access for more than a decade. The institution does so by developing energy access databases on electrification rates and traditional biomass cooking, and by creating quantitative analyses to estimate the future situation and investment needs (In- ternational Energy Agency, 2016b). Several annual reports, including the WEO 2013, WEO 2015, and WEO 2016, devoted a chapter to energy access (International Energy Agency,2013, 2015,2016a). The latest IEA number of people in sub-Saharan Africa without access to clean cooking facilities is found in WEO 2016 and equaled 792 million in 2014 (International Energy Agency, 2016a). The topic of energy access was covered more in depth by the IEA in their special energy investment focus (International Energy Agency,2011). This document estimates the number of people lacking access to electricity and clean cooking and provides insights on the level of investments in energy access and the financing sources. In 2009, the aggregated global investments in extending energy access amounted to USD9.1 billion. A large share of

(41)

Section 2.3 National and international organizations

the funding was directed towards electricity projects. International Energy Agency (2015) includes an updated investment estimate, namely USD13.1 billion in 2013. Only 3% of the energy access investments went to extending access to clean cooking facilities. To provide an outlook for energy access to 2030, the IEA considers existing government policies and a cautious implementation of intended policies. This scenario is named the New Policies Scenario. According to these projections, 823 million people in sub-Saharan Africa will lack access to clean cooking facilities in 2030 (International Energy Agency, 2016a). To achieve universal access by 2030, both for electricity access and clean cooking, International Energy Agency (2011) states that an annual average investment of USD48 billion is required. This value represents five times the investment level of 2009 and demonstrates the size of the challenge.

As opposed to other IEA perspectives, International Energy Agency (2014) presents a more detailed view on the cooking fuel mix in sub-Saharan Africa, while offering a comprehensive study of the energy system in Africa. Starting from the 2012 base year, the IEA developed projections to 2040 for cooking energy consumption by fuel for urban and rural areas for the whole of sub-Saharan Africa, population with and without clean cooking access in four sub- regions of sub-Saharan Africa (West Africa, Central Africa, East Africa, Southern Africa), and fuel/technology penetration rates for urban and rural areas in the four sub-regions. These projections state that 653 million people will still cook with biomass in an inefficient way by 2040.

It is important to note that this number is lower than the latest estimate of 708 million in International Energy Agency (2016a). This deviation is an indication for the high uncertainty of the long-term projections. Unfortunately, the methodology behind the IEA projections is not transparent. International Energy Agency (2016b) declares that the dependency on traditional biomass is calculated from panel econometric regressions, but it remains unclear how projections for different fuels are obtained. Moreover, the results are presented on a regional level.

This level of aggregation makes it impossible to know whether differences among countries are present.

(42)

(43)

Chapter 3

Methodology

As described before, the purpose of this report is to provide an outlook for the cooking fuel mix in sub-Saharan Africa to 2050. In order to do so, the following steps were identified:

1. Study literature to understand context and determinants of fuel choice in developing regions

2. Develop calculation logic for the projections 3. Implement defined calculation logic in Excel 4. Process results and derive insights

The next sections go more into detail for each separate step.

3.1 Literature study

The literature study focused on both academic work and publications from international or national organizations that are active in the field of clean cooking. The investigated academic papers covered the identification and quantification of determinants of fuel choice. This allowed to determine key drivers of the switch away from traditional biomass cooking. Moreover, several approaches to model the future cooking fuel mix were analyzed and the differences with the approach behind this report were discussed. In addition, reports from well-known bodies helped to build an understanding of the increased attention for the case of clean cooking and the efforts that are being made to accelerate the uptake of clean cookstoves.

3.2 Develop calculation logic

Before diving into the calculation logic, the reader should understand that the model distin- guishes between five fuel categories and seven cooking technology groups. The connection between fuel groups and technologies is visualized in figure3.1. As can be seen, the main difference between these two terms lies in the fact that technologies differentiate between the traditional and improved use of biomass, which is relevant for the fuel categories of wood and charcoal.

Clean cooking in sub-Saharan Africa: modeling the cooking fuel mix to 2050

Clean cooking in sub-Saharan Africa:

modeling the cooking fuel mix to 2050

Henri Casteleyn

Clean cooking in sub-Saharan Africa:

modeling the cooking fuel mix to 2050

Submitted by Henri Casteleyn

June 2017

Abstract

Acknowledgements

Contents

List of Figures

List of Tables

Abbreviations

Physical Constants

Symbols

To my parents.

Chapter 1

Introduction

1.1 Background

1.2 Purpose of the master’s thesis

1.3 Limitations

1.4 Organization of report

Chapter 2

Frame of reference

2.1 Factors influencing cooking energy choice

2.2 Modeling cooking fuel mix

2.3 National and international organizations

Chapter 3

Methodology

3.1 Literature study

3.2 Develop calculation logic