POTENTIAL OF BIOETHANOL AS A

HOUSEHOLD FUEL FOR MIDDLE-INCOME

URBAN KENYA: CASE OF NAIROBI CITY

JOSEPH MBOTHU MACHANDI

MASTER OF SCIENCE

(Energy Technology)

JOMO KENYATTA UNIVERSITY OF

AGRICULTURE AND TECHNOLOGY

APRIL 2014
 Potential of Bioethanol as a Household Fuel for

Middle-income Urban Kenya: Case of Nairobi City

Joseph Mbothu Machandi

Thesis submitted in partial fulfilment of the degree of

Masters of Science in Energy Technology in the Jomo

Kenyatta University of Agriculture and Technology

2014
 DECLARATION

This thesis is my original work and has not been presented for a degree in any other

University.

Signature: ………………………………….. Date…………..………

Joseph Mbothu Machandi

This thesis has been submitted for examination with our approval as University

Supervisors:

Signature: ………………………………….. Date…………..………

Dr. John M. Kihoro

JKUAT, Kenya

Signature: ………………………………….. Date…………..………

Dr. Benson B. Gathitu

The Technical University of Kenya

i
 ACKNOWLEDGMENT

My heartfelt gratitude goes to my supervisors Dr. Benson B. Gathitu and Dr.

John M. Kihoro for their time, dedication and constant assistance during the

entire research period.

I am also grateful to the staff of IEET, JKUAT for their academic guidance that

contributed to shaping my life and made me excel to higher levels.

Special thanks go to Technical University of Mombasa for providing the

financial support which went a long way in making my study possible.

Thanks go to Evalyne Wanjiku, Kelvin Machandi, Maurine Cheruto Kirui and

Jenita Muthoni all whose contribution to this research was very valuable.

Finally my gratitude goes to my family for being patient and understanding

during this period.

ii
 TABLE OF CONTENTS

DECLARATION ..................................................................................................... i

ACKNOWLEDGMENT ........................................................................................ ii

TABLE OF CONTENTS ...................................................................................... iii

LIST OF TABLES............................................................................................... viii

LIST OF FIGURES............................................................................................... ix

LIST OF APPENDICES ........................................................................................ x

ACRONYMS AND ABBREVIATIONS ............................................................. xi

ABSTRACT ......................................................................................................... xiii

CHAPTER ONE ..................................................................................................... 1

INTRODUCTION .................................................................................................. 1

1.1 Background to the problem ............................................................................. 1

1.2 The problem statement .................................................................................... 4

1.3 The purpose of the study ................................................................................. 5

1.4 The objectives of the study.............................................................................. 6

1.4.1 General objective ..................................................................................... 6

1.4.2 Specific objectives .................................................................................... 6

1.5 Research questions .......................................................................................... 6

1.6 Theoretical framework .................................................................................... 7

1.7 The Model........................................................................................................ 8

1.8 Expected outputs............................................................................................ 10

iii
1.9 Justification of the study................................................................................ 11

CHAPTER TWO.................................................................................................. 13

LITERATURE REVIEW .................................................................................... 13

2.1 Introduction ................................................................................................... 13

2.2 Household fuel and energy use in developing countries ............................... 13

2.3 Overview of household fuels in Kenya ......................................................... 19

2.4 Bioethanol fuel .............................................................................................. 21

2.5 Chemistry....................................................................................................... 25

2.6 Types of bioethanol ....................................................................................... 26

2.7 Overview of bioethanol production process and technology ........................ 28

2.8 Major steps in production process ................................................................. 30

2.8.1 Fermentation ........................................................................................... 30

2.8.2 Distillation............................................................................................... 31

2.8.3 Dehydration............................................................................................. 31

2.8.4 Denaturing............................................................................................... 32

2.9 Bioethanol programs...................................................................................... 32

2.9.1 Brazil ....................................................................................................... 33

2.9.2 Germany.................................................................................................. 35

2.9.3 India ........................................................................................................ 35

2.9.4 Thailand................................................................................................... 36

iv
 2.9.5 United States ........................................................................................... 36

2.9.6 Ethiopia ................................................................................................... 37

2.9.7 Malawi .................................................................................................... 37

2.9.8 Nigeria..................................................................................................... 38

2.9.9 South Africa ............................................................................................ 39

2.9.10 Sudan..................................................................................................... 39

2.10 History and current status of bioethanol in Kenya ...................................... 40

2.11 Research gap................................................................................................ 43

CHAPTER THREE.............................................................................................. 44

METHODOLOGY ............................................................................................... 44

3.1 Introduction ................................................................................................... 44

3.2 Research design ............................................................................................. 44

3.3 Population and sample................................................................................... 45

3.4 Data collection............................................................................................... 47

3.5 Data analysis.................................................................................................. 47

CHAPTER FOUR ................................................................................................ 51

RESULTS AND DISCUSSION........................................................................... 51

4.1 Introduction ................................................................................................... 51

4.2 Modern technology........................................................................................ 51

4.3 Ranking of fuels............................................................................................. 52

v
 4.3.1 Ignoring fuel cost .................................................................................... 53

4.3.2 In order of cost ........................................................................................ 53

4.3.3 In order of risks ....................................................................................... 53

4.4 Characteristics of fuels .................................................................................. 55

4.5 Household fuel consumption ......................................................................... 57

4.6 Binary logistic regression .............................................................................. 58

4.6.1 Cooking (Charcoal)................................................................................. 58

4.6.2 Cooking (Kerosene) ................................................................................ 59

4.6.3 Cooking (LPG)........................................................................................ 61

4.6.4 Cooking (Electricity)............................................................................... 62

4.6.5 Water heating (electricity)....................................................................... 64

4.6.6 Water heating (charcoal) ......................................................................... 65

4.6.7 Space heating (electricity)....................................................................... 66

4.7 Bioethanol supply .......................................................................................... 70

4.7.1 Bioethanol feedstock.............................................................................. 70

4.7.2 The availability of bioethanol feedstock ................................................ 72

4.8 Energy balance .............................................................................................. 75

4.9 Bioethanol fuel and stove .............................................................................. 77

4.10 Bioethanol co-products................................................................................ 80

vi
4.11 Employment and income ............................................................................. 85

4.12 Environmental and social impacts............................................................... 86

4.13 Roadmap to bioethanol production.............................................................. 87

CHAPTER FIVE .................................................................................................. 89

CONCLUSION AND RECOMMENDATIONS................................................ 89

5.1 Introduction ................................................................................................... 89

5.2 Conclusion ..................................................................................................... 89

5.3 Recommendations ......................................................................................... 94

REFERENCES ..................................................................................................... 96

LIST OF APPENDICES .................................................................................... 101

vii
 LIST OF TABLES

Table 4.1 Percentages of households indicating that a fuel has that characteristic . 56

Table 4.2 Percentages of households using a particular fuel for a given purpose.. 57

Table 4.3 Characteristics of sweet sorghum and sugarcane .................................... 72

Table 4.4 Status Quo Feedstock Production Scenario............................................. 73

Table 4.5 Potential Feedstock Production Scenario ................................................ 74

Table 4.6 Optimised Potential Feedstock Production Scenario .............................. 74

Table 4.7 Energy Balance of Sugarcane Bioethanol Production in Brazil.............. 76

Table 4.8 Energy Balance for Bioethanol Production for different countries......... 77

Table 4.9 New products from the sugarcane agro- industry……………………… 84

viii
 LIST OF FIGURES

Figure 2.1 Energy ladder......................................................................................... 16

Figure 2.2 Distribution of world bioethanol production in 2004 ............................ 22

Figure 2.3 Distribution of world bioethanol production in 2006 ............................ 23

Figure 2.4 Distribution of world bioethanol production in 2007 ............................ 23

Figure 2.5 Types of feedstocks for bioethanol production...................................... 27

Figure 2.6 Technological routes for bioethanol production……………………….30

Figure 2.7 International trade in bioethanol in 2006 ............................................... 35

Figure 4.1 Barchart showing proportion of households possessing modern gadgets/

equipments .................................................................................................................. 52

Figure 4.2 Bar chart for the means of household fuels ignoring fuel cost. ............. 54

Figure 4.3 Bar chart for the means of household fuels in order of cost. ................. 54

Figure 4.4 Bar chart for the means of households fuels in order of risks ............... 55

Figure 4.5 The Domestic two burner CleanCook stove .......................................... 78

Figure 4.6 The SuperBlue stove.............................................................................. 78

Figure 4.7 The Gel fuel stove and a bottle of fuel................................................... 79

Figure 4.8 Sugar and sugarcane-based bioethanol production flowchart ............... 82

Figure 4.9 Bioethanol production and by-products from sweet sorghum ............. 83

ix
 LIST OF APPENDICES

Appendix I Questionnaire for households ................................................. 101

Appendix II Categorical variable coding ................................................... 104

Appendix III Variable coding ...................................................................... 105

Appendix IV Cooking with charcoal............................................................ 106

Appendix V Cooking with kerosene........................................................... 107

Appendix VI Cooking with LPG.................................................................. 108

Appendix VII Cooking with electricity......................................................... 109

Appendix VIII Water heating with electricity ................................................ 110

Appendix IX Water heating with charcoal ................................................... 111

Appendix X Space heating with electricity ................................................ 112

Appendix XI Lighting with electricity ........................................................ 113

Appendix XII Letter of research authorization from NCST ......................... 114

Appendix XIII Letter from Provincial Administration (Nairobi Area) ......... 115

x
 ACRONYMS AND ABBREVIATIONS

BNDES - National Bank of Economic & Social Development

CO - Carbon Monoxide

CO2 - Carbon Dioxide

C2H5OH - Ethanol (Bioethanol)

C6H12O6 - Carbohydrate sugar

CAP - Centre for American Progress

CGEE - Centre for Strategic Studies and Management in Science,

Technology and Innovation

E100 - Neat bioethanol fuel

ECLAC - Economic Commission for Latin America and Caribbean

EPI - Earth Policy Institute

ESMAP - World Bank Energy Sector Management and Assistance Program

GDP - Gross Domestic Product

GHG - Greenhouse gases

GoK - Government of Kenya

GTZ - German Technical Cooperation

FAO - Food and Agriculture Organization

FFVs - Flex-fuel vehicles

ICRISAT - International Crops Research Institute for the Semi-Arid

Tropics

IEA - International Energy Agency

IIED - International Institute for Environment and Development

xi
IIASA - International Institute for Applied Systems Analysis

KARI - Kenya Agriculture Research Institute

KEFRI - Kenya Forestry Research Institute

KEPHIS - Kenya Plant Health Inspectorate Services

KNBS - Kenya National Bureau of Statistics

LPG - Liquid Petroleum Gas

LSMS - Living Standard Measurement Survey

MDGs - Millennium Development Goals

MSW - Municipal Solid Waste

MTBE - Methyl Tertiary Butyl Ether

NACOSTI - National Commission for Science, Technology and

Innovation

REN21 - Renewable Energy Network for 21st century

RFA - Renewable Fuels Association

WAEMU - West African Economic and Monetary Union

UN - United Nations

UNDP - United Nations Development Program

UNEP - United Nations Environment Program

UNF - United Nations Foundation

WB - World Bank

WHO - World Health Organization

xii
 ABSTRACT

Provision of modern energy services is essential to improving the livelihood of

households in the developing world. The objectives of the study were; to

investigate the distribution of fuel types and fuel uses in middle-income

households, to investigate how the social and demographic factors influence

household’s fuel choices and to establish the benefits associated with bioethanol

production and use. The study used a questionnaire to collect data from

households and also secondary data from paper publications. The primary data

which was obtained from households addressed the demand aspect of fuels while

the secondary data addressed the supply aspect of bioethanol. The primary data

was analyzed quantitatively using descriptive statistics and binary logistic

regression, and the secondary data was reviewed and analyzed qualitatively. The

primary data was collected from 313 middle-income households within Nairobi.

The study findings of the households sampled, 7.3% used firewood as a fuel,

43% used kerosene, 60% used charcoal, 81% used liquid petroleum gas (LPG)

and 94% used electricity. Households were also found to be using more than one

fuel. More than 97% of the households that used LPG and electricity indicated

them as sources of energy that produce no smoke, and were clean and fast to

work with. Less than 47% of the households that used charcoal and kerosene

indicated them as fuels that produce no smoke, and were clean and fast to work

with. More than 90% of the households that used charcoal, kerosene and

electricity indicated them as fuels that were available. Only 52% of the

xiii
households that used LPG indicated it as a fuel that was available. More than

80% of the households that used kerosene, LPG and electricity indicated them as

expensive. Less than 20% of the households that used charcoal indicated it as an

expensive fuel. From the research findings, LPG was the preferred fuel for

cooking indicated by 81% of the households. For water heating, lighting and

space heating, electricity was the preferred source of energy indicated by 71%,

94% and 67% of the households respectively. For a particular use, households

were found to be using one or more than one fuel. Even though LPG and

electricity were indicated as expensive, they were the preferred sources of

energy. This was ascertained by a high proportion of households indicating them

as the most preferred fuels when ignoring the cost of fuels. A high proportion of

households used LPG for cooking (81%) and electricity for lighting (94%).

Cooking and lighting are the two important uses of fuels for households in rural

and urban areas. The research findings ascertained that some of the social and

demographic factors that were in the models generated using the binary logistic

regression were not significant (p > 0.05) and thus do not influence the models

when the other factors are held constant. It was also noted from the study that

even some of the factors that were significant (p < 0.05) do not concur with the

theoretical expectation. Household income, occupation and education of both the

household head and spouse were strongly expected to strongly influence

households’ fuel choices, but this was not the case. This was an indication that

social and demographic factors were not very significant in influencing

xiv
households’ fuel choices. The researcher proposed sugarcane and sweet sorghum

as the feedstock for bioethanol production due to their climatic suitability, crop

cycles per year and also they were not expected to compete with the food supply

chain. The research found out that if these two energy crops are grown and

expanded in areas with suitable climate, 60-560 million litres of bioethanol could

be produced from sugarcane and over 30 billion litres of bioethanol from sweet

sorghum. The energy balances for the production of bioethanol from either sugar

cane or sweet sorghum in different countries calculated from different approaches

all have positive values indicating that more energy is produced than is consumed

in its production. Bioethanol production has the potential to generate incomes and

employment in the rural areas by boosting the agriculture sector thus bringing

social-economic development. Bioethanol is a renewable, clean and efficient fuel

similar in characteristics to LPG as it emits heat similar in temperatures to the

heat output of LPG cookers and also burns in a smokeless flame similar to that of

LPG. Bioethanol is a green fuel and therefore a sustainable form of energy with

many benefits over LPG. Bioethanol is thus an alternative and clean energy

solution to replace LPG. The future of the country and of the world is in adopting

renewable energy and thus this study has undertaken a step to develop affordable

energy solution for this emerging market.

xv
 CHAPTER ONE

INTRODUCTION

1.1 Background to the problem

Energy is essential to and a hallmark of societal development. Countries with

access to abundant and affordable modern energy have significantly larger gross

domestic product (GDP), higher per capita income levels, longer life

expectancies, increased literacy rates, and greater educational attainment.

Without electricity and other modern energy services, it will be impossible to

raise rural productivity, alleviate poverty and achieve the United Nations

millennium development goals (MDGs) (UNF, 2008). The MDGs were

developed to address the problems of poverty, hunger, diseases, illiteracy,

environmental degradation and discrimination against women (WHO, 2004).

MDGs aim to promote a comprehensive approach and to allow many problems to

be tackled simultaneously across the wide range of spectrum concerning energy

(Mcnube, 2011). Modern household energy services are linked to many of the

MDGs as solution to these problems (Mcnube, 2011).

Alternative fuels, also known as non-conventional fuels or advanced fuels are any

materials or substances that can be used as fuels, other than conventional fuels.

Conventional fuels include fossil fuels (petroleum, coal and natural gas), and

nuclear materials such as uranium. Some well-known alternative fuels include

biodiesel, bio-alcohols (methanol, ethanol and butanol), chemically stored

electricity (batteries and fuel cells), hydrogen, non-fossil methane (biogas),

1
vegetable oil and other biomass sources. The main purpose of fuel is to store

energy, which should be in a stable form and can be easily transported to the

place of consumption. Almost all fuels are chemical fuels.

There are a number of problems associated with fossil fuels, most of which stem

from the by-products formed when they are burned to create energy. Chief

among those by-products are carbon dioxide and nitrous oxide and these are

greenhouse gases (GHG’s) that are the major contributors to global warming.

Largely because of coal and petroleum combustion, the amount of carbon dioxide

and nitrous oxide in the air today are 35% and 18% higher respectively than they

were before the industrial era (The American Coalition of Ethanol, 2008). Other

by-products of fossil fuel combustion include sulphur oxides and nitrogen oxides

both of which contribute to acid rain, and hydrocarbons which react with nitrogen

oxides to form smog. In addition to their environmental harm, the by-products of

burning fossil fuels can cause health problems for humans. Nitrogen oxides, for

instance, irritate lungs. Particulate matter such as soot and dust contribute to

respiratory illness and cardiac problems, including arrhythmias and heart attack.

Fossil fuels dependence also damages the economic health of our nation as all the

fossil fuels currently used in Kenya are imported from other countries. Kenya

recently discovered oil deposits in Turkana and feasibility studies are ongoing.

Despite the fact that fossil fuels are limited resources that cannot be replaced, the

Ministry of Energy acknowledges that fossil fuels usage in Kenya is likely to

grow. This means that unless we dramatically change the way Kenya consumes

2
energy, our dependence on foreign sources of fossil fuels will also grow, and

increasingly threaten our economic stability, business and daily life. Historically,

inflation in Kenya has been mainly caused by increases in oil prices (Wanambwa,

2005). Global oil consumption is projected to increase by 36% by 2030, and to

double in Africa at the same time (GTZ & GoK, 2008).

Efforts to substitute fossil fuels for alternative fuels are gaining attention in a

world threatened by climate change, rural economic decline, and instability in

major oil producing countries (EPI, 2005). Replacing petroleum with bio-fuels

can reduce air pollution, including emissions of fine particulates and carbon

monoxide. Bio-fuel production can also improve rural economies by creating

employment and raising farm incomes. When locally produced, bio-fuels have

the potential to diversify energy portfolios, lower dependence on foreign oil and

improve trade balances in oil-importing nations (EPI, 2005).

Educating the public about the benefits of clean burning, affordable, renewable

and alternative sources will be the first step to improving and protecting the

environment. Each alternative fuel is an option to replace fossil fuels and to

reduce the emissions caused by the fossil fuels. The European Union (EU), by

setting targets of carbon emissions reduction and increase in the use of renewable

energy before 2020, made an extraordinary attempt to achieve more sustainable

energy production in the EU, with some reactions also being noticed in Asian,

African and American countries (Watanabe, 2009; Gnansounou & Dauriat,

3
2004). A sustainable energy future depends on an increased share of renewable

energy, especially in the developing countries (Goldemberg, 2007).

1.2 The problem statement

Roughly three billion people living in developing countries, Kenya included, rely

on traditional fuels such as firewood and charcoal for their daily cooking, lighting

and heating needs (WHO, 2006). In recent decades the unsustainable exploitation

of biomass, for both household and industrial use, has been causing increased

concern due to the damage done on the environment as well as to human health.

At a local level, deforestation can lead to increased flooding and decline of soil

quality. After combustion of biomass, most of carbon dioxide (CO2) is absorbed

back into plants during photosynthesis, however if the plants are unsustainably

harvested, this absorption effect is largely negated. The use of open fires and

stoves makes household energy usage a large source of pollution, with the

burning of biomass indoor, often in poorly ventilated houses, being strongly

linked to acute respiratory infections and other health problems (WHO, 2006).

Compared to traditional fuels, modern energy sources such as kerosene and liquid

petroleum gases (LPG) offer increased efficiencies, reduced emissions and are

more user-friendly. However they release fossil derived GHG’s, and they are

more expensive to both the national economy and user, with poverty being one of

the main barriers to their uptake. Kenya spends nearly US$ 1 billion on oil

importation (GTZ & GoK, 2008). Thus there is need to seek alternatives to high

dependence on imported oil and the concomitant outflow of foreign currency.

4
Rising costs of fossil fuels compromise the ability of many developing countries

to broaden access to energy, even as the use of such fuels worsens global climate

change (UNF, 2008). Developing countries are uniquely vulnerable to climate

change, which will change weather patterns and disrupt agriculture.

Bio-fuels could provide many attractive opportunities to reduce this dependence

while reinvesting in the country’s sustainable development. Liquid biomass

cooking fuels combine some of the advantages of both traditional and new fuels

and bioethanol is beginning to emerge as a viable household fuel, with options

based around either a liquid or a “gel fuel” form. This research will determine the

potential of bioethanol as a household fuel in middle-income urban Kenya. This

will go along with proposing a fuel that will be sustainable as well as reducing

the use of firewood, charcoal, kerosene and LPG.

1.3 The purpose of the study

The purpose of the study was to determine the potential of bioethanol production

and use as a fuel for households in middle-income urban Kenya. The study

investigated the energy use in these households. The study reviewed the benefits

associated with bioethanol production and use. Thus the research study

determined the production and use of bioethanol as a household fuel from an

economic, environment and social point of view.

5
 1.4 The objectives of the study

1.4.1 General objective

To investigate the potential of bioethanol as a household fuel for middle-income

in urban Kenya.

1.4.2 Specific objectives

i. To investigate the distribution of fuel types and fuel uses by households.

ii. To investigate how the social and demographic factors influence

household’s fuel choices.

iii. To establish the value chain for bioethanol production and use.

1.5 Research questions

The following research questions have been asked in this study:

i. What are the perceived costs and benefits of using traditional and modern fuels

(firewood, charcoal, kerosene, LPG and electricity) as household fuels?

ii. What are the potential crops that could be used to produce bioethanol in

Kenya?

iii. What are the costs and benefits of production and use of bioethanol?

iv. What is the current status and future potential of bioethanol as a household fuel

in urban Kenya?

6
1.6 Theoretical framework

A household’s fuel choice consumption decision can be understood by analyzing

its decision in constrained utility maximization framework. This is subject to a

set of economic and non-economic constraints. The economic factors include fuel

price and household income. Non-economic factors include gender, age,

education, occupation of household head and spouse, household size, ownership

of the dwelling unit and the type of the dwelling unit.

Energy is provided by a multiplicity of sources. Each energy source is a

commodity with multiple attributes and multiple purposes (Njong & Tabi, 2011).

The purposes include cooking, heating, lighting etc. The attributes include energy

content, convenience, safety, speed of cooking, quality of light and smoke

emitted when burned. Energy sources are intermediate inputs into the utility

function (Njong & Tabi, 2011). Utility is derived from the final goods such as

cooked food, heat, lighting and entertainment which energy sources help to

produce.

Following Pundo & Fraser (2006), the household choice is a function of a set of

social factors. For this study the social factors to be considered are gender, age of

household head and spouse, level of education of household head and spouse,

occupation status, household size, ownership of the dwelling unit and whether or

not the dwelling unit is modern or traditional.

7
1.7 The Model

The study used binary logit model to estimate the significance of the factors

thought to influence household fuel choices. The model examined fuel choices

from a set of mutually exclusive household fuels such as firewood, charcoal,

kerosene, LPG, and electricity. The probability that a household chooses one type

of fuel is restricted to lie between zero and one. The model assumes no changes

in fuel prices or fuel attributes, and also that households make fuel choices that

maximize their utility.

In binary logit model, the dependent variable is a dichotomous outcome or event.

For this study the dependent variable is fuel choice. Households can either

choose a particular fuel or not choose it. Data is entered into the analysis as 0 or 1

coding for dichotomous outcome, continuous values for continuous predictors,

and dummy coding (e.g. 0 or 1) for categorical predictors. The model can be

expressed as follows;

(1)

Where:

Pr (Yj = 1) is the probability of choosing fuel j

j is an indicator of either firewood, charcoal, kerosene, LPG or electricity

Xj is the jth social factor(variable)

Bj is the estimated parameter corresponding to the jth social factor

8
NB: X (predictor) can be categorical or continuous but Y (response) is always

categorical.

Rearranging the equation above using algebra and taking the natural logarithm,

the following regression equation is obtained;

(2)

The estimated parameters (or coefficients) of the logit model are estimated using

the maximum likelihood method i.e. maximizing log-likelihood with respect to

parameter vector. This is done by differentiating partially the above log equation

with respect to each of the parameter vectors, equating the results to zero and

then solving the equations. In maximum likelihood, the aim is to maximize the

log-likelihood function (Gujarati, 2003).

The quantity is called the odds ratio and is the logit or log odds.

The equation expresses the logit (log odds) as a linear function of the

independent factors (Xs). The equation allows the interpretation of the logit

weights for variables in the same way as in the linear regressions. The estimated

coefficients refer to the degree to which the probability of choosing a fuel would

change with a unit change of explanatory or predictor variables. The estimated

coefficients tell the relationship between the independent variables and the

dependent variable, where the latter is on logit scale. The dependent variable is

the fuel choice (firewood, charcoal, LPG, kerosene or electricity). Estimated

coefficients measure the estimated change in the logit for one unit change in the

9
predictor variable while the other predictor variables are held constant. When the

estimated coefficient is positive, larger (or smaller) X values are associated with

larger (or smaller) logit of Y. Conversely, if estimated coefficient is negative,

larger (or smaller) X values are associated with smaller (or larger) logit of Y.

1.8 Expected outputs

In this study, the social and demographic factors to be considered are; the age of

the household head and that of the spouse, the level of education of the household

head and that of the spouse, the occupation of the household head and that of the

spouse, gender of the household head and that of the spouse, marital details of

household head and that of the spouse, religion of the household head and that of

spouse, household size, household income, whether or not the house hold owns

the dwelling unit, whether the dwelling unit is modern or traditional type house

and whether the kitchen is internal or external.

Theoretically, the above social factors are expected to influence household fuel

choice in the following manner. The older the household head or the spouse, the

more likely the household will continue to use the traditional fuels i.e. firewood

and charcoal. The level of education of household head and spouse is expected to

have positive effects on traditional fuel alternatives i.e. kerosene, LPG and

electricity. The level of education improves knowledge of fuel attributes and

preference for better and cleaner fuels. Occupation of wife or husband is

expected to have a positive effect on traditional fuel alternatives. This is due to

improvements in income and thus can afford the modern fuels. The household

10
size is expected to negatively affect choice of traditional fuel alternatives. The

ownership of dwelling unit is expected to be critical in household fuel choice. If

the households dwelling unit is rented, the household is more likely to use

traditional fuel alternatives. The disadvantage of traditional fuels especially

firewood is that it produces smoke and stains walls and roofs, making it less

preferred in rented houses. Also if the dwelling unit is modern type house, the

household is more likely to use cleaner fuels.

The level of education of wife or husband, occupation of wife or husband,

whether or not the household own the dwelling unit and the type of dwelling unit

all are expected to have a positive effect on traditional fuel alternatives. This will

be an indicator of household’s willingness to switch to clean modern fuels. Thus

introduction of bioethanol as a household fuel is likely to be accepted by many

households. Bioethanol is one of the modern clean cooking fuels and is part of

the Global Clean Cooking Fuel Initiative (GCCFI) which addresses social and

developmental issues. The price and availability of bioethanol will be very

crucial in its acceptance as a household fuel for many households.

1.9 Justification of the study

Bioethanol is an energy source derived from renewable biomass. Bio-fuels

generally emit less toxic air pollutants and GHGs than petroleum–based fuels and

can be produced where sufficient biomass feedstock can be grown. At a time of

record high oil prices and growing concern over global warming, bio-fuels

present a valuable opportunity to reduce dependence on volatile global oil

11
markets, create local economic opportunities in agriculture and industry, and

improve the environment. Kenya has suitable climatic conditions for growing

bio-fuel crops and thus could limit the shock of high oil prices by developing its

own supply of domestically produced bio-fuels.

1.10 Limitations and assumptions

The foreseen limitation in the study will be on data collection. There isn’t a lot of

data documented on bioethanol production and use in Kenya. The data available

is scanty to give the real situation in Kenya. People might be unwilling to give

information especially those working in the two bioethanol producing plants in

Kenya in the fear not to reveal company secrets and also when collecting data in

households. The logit model used in this study assumes a linear relationship

between logit of the independent variables and the dependent variables, and the

independent variables are not linear functions of each other.

12
 CHAPTER TWO

LITERATURE REVIEW

2.1 Introduction

This section reviews the academic literature on household fuels and energy uses

in developing countries as well as an overview of household fuels in Kenya. This

study also reviews literature on bioethanol fuel; its development in the recent

past, brief chemistry of bioethanol, world bioethanol programs especially the

success story of Brazil and the history and current status of bioethanol fuel in

Kenya.

2.2 Household fuel and energy use in developing countries

Energy and fuel use are important for the welfare of households in developing

countries. Using an energy source for lighting and cooking is essential to human

life. Many households remain dependent on traditional biomass fuels for cooking

and on inefficient and costly sources of light such as candles and kerosene.

Improving access to modern energy sources and clean cooking technologies is an

important development goal.

Rural households rely more on biomass fuels than those in urban areas, but well

over half of all urban households in sub-Saharan Africa rely on fuel wood,

charcoal or wood wastes to meet their cooking needs (IEA, 2006). Nearly two

million metric tonnes of fuel wood and charcoal are consumed daily in

developing countries. Fuel wood, charcoal along with dung and agricultural

residues provide over half of the total energy consumed in some 60 – 70

13
developing nations. These fuels supply as much as 95% of the domestic energy in

these countries (Wood & Baldwin, 1985). It has been estimated that about 2.5

billion people in developing countries rely on biomass fuels to meet their cooking

energy needs (IEA, 2006). In these countries about 90% of total household fuel is

biomass. Without new policies, the number of people that rely on biomass fuels

is expected to increase to 2.6 billion by 2015, and 2.7 billion by 2030 due to

population growth (IEA, 2006). Heavy reliance on biomass fuels mainly fuel

wood, charcoal and dung, contribute to deforestation, forest degradation and land

degradation (Njong & Tabi, 2011). Malawi is one of sub-Saharan Africa’s most

densely populated countries. This has put the environment under great pressure

with extreme poverty, the demand for farmland and overdependence on

traditional biomass fuels, in both the agro-industrial and domestic sectors,

resulting in increasing deforestation rate and an estimated loss of 500 square

kilometres of forest per year (Robinson, 2006). Recognizing the adverse effects

of the use of traditional biomass fuels, UN millennium projects recommend

halving the number of households that depend on traditional biomass for cooking

by 2015, which involves about 1.3 billion people switching to other fuels (IEA,

2006). Biomass-based traditional fuels are being steadily replaced with modern

fossil fuels and electricity in cities of developing countries as the latter energy

sources are more convenient and efficient to use, and produce less local

pollutants (Alam et al., 1998).

14
Clean cooking fuels are important for combating the high levels of indoor air

pollution wherever traditional solid fuels are used for cooking and heating. The

use of clean cooking fuels can also have positive effect on the external

environment by reducing outdoor air pollution from venting of kitchen smoke as

well as by preventing degradation through collection of wood for firewood or

charcoal production (ESMAP, 2001). Usage of all modern fuels in developing

countries increases with income except kerosene which tends to be neither

progressive nor regressive (WB, 2003).

The main challenges facing the global energy sector are two-fold; namely, to

increase access to affordable, modern energy services to poor countries that lack

them, and to find the mix of energy sources, technologies, policies and

behavioural changes that will reduce adverse environmental impacts of providing

the necessary energy services (Abebaw, 2007).

Households in developing countries practice fuel switching which is either partial

or complete. Household partial fuel switching is where a household switches to

another fuel but continues to use the old fuel, whereas with complete switching

the old fuel is not used. These are energy transitions and can be presented on an

energy ladder which rates the quality of household fuels. Figure 2.1 represents

fuel types that might be used by households as their prosperity increases.

Households climb the ladder as their income increases (Smith et al., 1994;

Meikles & Bannister, 2003; WHO, 2002). Urban households are more likely to

use cleaner technology as their income increase (Hosier & Dowd, 1982;

15
Campbell et al., 2003). The choice of a fuel is determined by a particular

household’s personal choice and income (Leach, 1992; Smith et al., 1994). Fuel

switching in developing countries is that households ascend the energy ladder

from an inefficient fuel and energy end-use equipment to a more efficient fuel

with increasing income levels and urbanisation (Farsi et al., 2004). The research

in energy consumption of households in developing countries shows that the

energy ladder is simplistic. There are many factors other than income that

determine fuel choice i.e. culture, social, desirability and security of supply

(Davis, 1992; Bernett, 2000).

Increasing cleanliness, efficiency, cost, convenience

Electricity

LPG

Kerosene

Charcoal

Wood

Crop residue

Increasing prosperity
Figure 2.1: Energy ladder
Source: WHO (2002)

In the developing countries, there is a very large variation across countries in the

composition of households’ energy expenditure. In the poorest countries, biomass

16
and kerosene often feature more among the cooking fuels, LPG and kerosene

tend to be where most of the fuel budget is spent (WB, 2003). The urban poor are

the most exposed to energy price fluctuations as they often consume a mix of

electricity, purchased wood or charcoal and kerosene. The energy budget share of

households decrease with income and with household size, reflecting the fact that

energy is a basic good and there are economies of scale in household size (WB,

2003). Studies (Njong & Tabi, 2011; Mekonnen & Kohlin, 2008; Jiang &

Pachauri, 2008; Ouedraogo, 2005; Farsi et al., 2004) indicate that the level of

education, household income, and household size, whether or not a household

owns the dwelling unit, and whether or not the dwelling unit is traditional or

modern type are important factors that influence household fuel choices. For

example, the study by (Mekonnen & Kohlin, 2008) show households in urban

areas in Ethiopia tend to increase the number of fuels they use as their income

rises instead of completely switching from the consumption of traditional fuels

(such as wood) to modern ones (such as kerosene and electricity). This is perhaps

due to a number of factors, including preferences, taste, dependability of supply,

cost, cooking and consumption habits, and availability of technology.

Policy interventions targeting cooking fuels and cooking practices were earlier

motivated by a desire to control deforestation but are presently also motivated by

concerns regarding indoor pollution. Indoor air pollution has been estimated to be

the world’s 4th largest killer, causing perhaps 2.5 million premature deaths a year,

(WHO, 2006). To reduce indoor pollution, policies now focus on either inducing

17
a healthy fuel choice or making biomass use cleaner and safer by using improved

stoves.

The agriculture sector in most sub-Saharan countries is dominated by subsistence

farming with very low investment levels and yields. The development of modern

bio-energy systems offers opportunities for investment and infrastructure

improvements with the promise to diversify agriculture production and thus to

stimulate socio-economic development (Janssen et al., 2009). Bio-fuels afford

important opportunities for sustainable development both globally and

domestically. Bio-fuels can help to tackle climate change and improve rural

employment and livelihoods, diversify energy portfolios, ameliorate trade

balances and improve air quality (Dufey, 2006). Access to and use of cleaner

energy remains the major issue in discussions about sustainable economic

development and environment (Abebaw, 2007). An estimate by IEA revealed that

about 2.4 billion people in developing countries were still dependent on bio-fuels

as their main energy source (IEA, 2002).

Household sector experiences the most pronounced changes in its pattern of fuel

use overtime. Household fuel transition often accompany changes in economic

activity and lifestyles and are central focus of national and state government

policy (Alam et al., 1998). Household fuel and energy use survey data are mainly

sourced from living standard measurement surveys (LSMS). The major energy-

related pieces of data that can be extracted from standard LSMS surveys relate to

the ability to identify patterns of energy and fuel usage and spending at the

18
household level. This information is cross-tabulated with household expenditures

and other household characteristics to gain insight into the distributional profile

of particular fuels as well as the determinants of fuel use. Such analysis is helpful

for evaluating energy subsidies and taxes (Heltberg, 2004). The subsidies and

taxes are forms of government interventions in household fuel markets to

improve the availability of modern fuel at affordable prices (Goldemberg, 2007;

Alam et al., 1998). In India, energy supplies such as LPG, kerosene and

electricity are highly subsidized, the aim being to target the poor even though it is

the rich who benefit the most from the subsidies (Alam et al., 1998).

2.3 Overview of household fuels in Kenya

Firewood and agricultural residues are the predominant energy sources. These are

usually not purchased and are multiple purpose fuels used mainly for heating and

lighting. A Large proportion of the population use firewood for cooking. This is

more in rural regions and less in urban regions.

Charcoal is derived from wood and is the fuel of choice in the urban household.

In rural areas it is used to a lesser extent. Charcoal is a high energy density fuel,

easy to transport, use and store. Production of charcoal in Kenya is not

sustainable and this reduces its environmental appropriateness as a fuel, even

though it has a higher efficiency (15-35%) than wood. Productions of more

efficient charcoal stoves have greatly improved its use in developing countries.

Kerosene is a fossil liquid fuel distilled from petroleum. It is a fuel used for

heating and lighting and produces soot and other particulates when combusted.

19
Kerosene is mainly used for lighting but has a market as a cooking fuel,

especially in the urban areas mainly due to its availability. Kerosene is a

relatively costly fuel in Kenya, this due to high excise duty on it, a move by

government to prevent its use in adulteration of petrol and diesel.

For decentralized use, LPG is transported and stored as compressed liquid in

canisters ranging in size from 3kg to 50kg or larger. For household use, LPG gas

canisters and regulators are sold and refilled by petroleum distributors. Kenya has

a well developed LPG market compared to its neighbours. LPG appears more

expensive compared to fuels like charcoal but the important feature of LPG is the

efficiency during use which is higher than that of charcoal i.e. 50% compared to

20% (ESD Ltd). LPG stoves can also be controlled more precisely to the power

required by the user. A 1994 welfare survey showed that only 3% of Kenyans use

LPG for cooking (Karekezi et al., 2002)

Electricity is an efficient and clean source of energy. Electricity is not popular as

a cooking fuel but is mainly used for lighting both in rural and urban areas.

Electricity has difficulties in penetrating the cooking market mainly due to

combination of low consumer incomes, high electricity tariffs and high appliance

cost. Further, access to electricity is low at about 15% of the total population

(Ministry of Energy, 2004). This is due to the high costs of consumer connections

and network extension, particularly in rural areas.

20
2.4 Bioethanol fuel

Bioethanol fuel is ethanol derived from biomass. It is mostly used as a motor

fuel, mainly as a bio-fuel additive for gasoline. World bioethanol production

tripled between 2000 and 2007 from 17 billion to more than 52 billion litres

(UNEP, 2009). In 2009 worldwide bioethanol fuel production reached 73.9

billion litres (RFA, 2010). Since 1978, the government of Brazil has made it

mandatory to blend bioethanol and gasoline. In 2008, the EU proposed a

mandatory target for bioethanol blending of 10 percent by 2020 (Lisboa, 2011).

Some states in US have mandated the use of 10 percent bioethanol blends with

gasoline (WI &CAP, 2006). Presently, Brazil have a fleet of more than 10 million

flexible fuel vehicles regularly using neat bioethanol fuel known as E100 (Green

Car Congress, 6th March 2010).

Bioethanol is widely used in Brazil and in USA, and together both countries

produced 89 percent of the world’s bioethanol fuel production in 2009 (RFA,

2010). In 2004, Brazil led world bioethanol production by producing 15 billion

litres primarily from sugarcane, the US produced 13 billion litres exclusively

from corn, China produced 3.8 billion litres from wheat and corn, India produced

1.9 billion litres from sugarcane and France, the front-runner in bioethanol

production in EU, produced 0.8 billion litres from sugar beet and wheat (EPI,

2005). Data on bioethanol show important trends of expansion and

diversification. In 2007, the US became the leader in global production, with an

output of 26 billion litres of corn-based bioethanol, followed by Brazil with

21
approximately 20 billion litres of sugarcane-based bioethanol (REN21, 2007).The

main producers in Asia in 2007 being China and India which produced 3.7 billion

and 2.3 billion litres respectively, and the largest producer in EU being France

which produced 1.2 billion litres. Bioethanol production from all the Asian

countries reached 7.4 billion litres in 2007, while in the EU bioethanol production

rose to 2.3 billion litres in 2007 from 1.6 billion litres in 2006 (Licht, 2007).

Figure 2.2, 2.3 and 2.4 shows the distribution of world bioethanol production in

2004, 2006 and 2007 with total share of USA and Brazil increasing with the

years respectively.

Figure 2.2: Distribution of world bioethanol production in 2004

Source: Dufey, (2006)

22
Figure 2.3: Distribution of world bioethanol production in 2006
Source: BNDES & CGEE, (2008)

Figure 2.4: Distribution of world bioethanol production in 2007

Source: REN21 (2008) cited in BNDES & CGEE (2008)

23
Bioethanol is a form of renewable energy that can be produced from agricultural

feedstock. It can be made from very common crops such as sugarcane, potato,

sugar beet, sorghum, wheat and barley among others. Cellulosic bioethanol offers

promise because cellulose fibres, a major and universal component in plant cell

walls can be used to produce bioethanol. According to the IEA, cellulosic

bioethanol could allow bioethanol fuels to play a much bigger role in the future

than previously thought. Several reasons such as energy security, environmental

concerns, foreign exchange savings, and socio-economic issues related to the

rural sector justifies bio-fuels like bioethanol to be considered as relevant

technologies by both developing and industrialized countries (Demirbas, 2008;

Gnansounou & Dauriat, 2004).

Even though production of bioethanol has impacted on ecosystems and food

supplies through destruction of forests and use of food crops, it can be produced

in an environmentally and socially beneficial way if the right crops and models of

production are prioritized (GTZ & GoK, 2008). In order to avoid potential

negative environmental and social-economic impacts, effective mechanisms have

to be put in place (Janssen et al., 2009). There is need to formulate policies and

development plans to guide the development and ensure the sustainability of

bioethanol production. To ensure food security, African governments have

implemented several initiatives such as agro-ecological zoning to identify land

available for food and for bio-energy production (Janssen et al., 2009).

24
2.5 Chemistry

Glucose, a simple sugar is synthesized in the plant by photosynthesis from carbon

dioxide and water in presence of sunlight.

LIGHT
6 CO2 + 6H2O C6H12O6 + 6O2 (3)

During bioethanol fermentation, glucose is decomposed into bioethanol and

carbon dioxide

C6H12O6 2C2H5OH + 2CO2 + HEAT (4)

During combustion, bioethanol reacts with oxygen to produce carbon dioxide,

water and heat

C2H5OH + 3O2 2CO2 + 3H2O + HEAT (5)

The net reaction for the overall production and consumption of bioethanol is just

light to heat.

When bioethanol is combusted, the heat produced can be used to drive the piston

in the engine by expanding heated gases. It can be said that sunlight is used to run

the engine (as is the case with any renewable energy sources, as sunlight is the

only way energy enters the earth).

Fructose is another simple sugar in plants that undergoes fermentation. Three

other compounds namely starch, cellulose and sucrose in the plant can be

fermented after breaking them up by hydrolysis into the glucose or fructose

molecules that compose them. The energy to create fructose in the plant

ultimately comes from the metabolism of glucose created by photosynthesis, and

25
so sunlight also provides energy generated by fermentation of these other

molecules.

2.6 Types of bioethanol

There are two main types of bioethanol:

Starch sugar – produced from carbohydrate rich materials such as corn, potatoes,

sugarcane, sugar beet and oats among others. This type of bioethanol is referred

to as first generation bio-fuel.

Cellulosic sugar – produced from woody materials such as corn Stover (the

waste dried stalks and leaves) and wood processing waste. This type of

bioethanol is referred to as “second generation” or “advanced bio-fuels”.

Figure 2.5 shows the different feedstock that can be used to produce bioethanol.

Much of the current commercial production of bioethanol is from sugar cane,

maize and sugar beet, as all these represent fairly high yield, readily accessible

sugar sources. The technology to produce bioethanol from lignocelluloses

feedstock is yet to become economically competitive. Lignocelluloses biomass is

readily available and often waste material, particularly from the food processing

and forest product industries. This makes it an ideal candidate for those seeking

locally-sourced low cost feed stock.

26
 Root crop Sugar beet

Sugarcane
Sugar
Stalk crop
Sweet sorghum

Corn

Barley

Cereals Rye

Wheat
Starch Sorghum green

Potatoes
Root crops
Cassava

Forest residues
Willows

Energy crops Poplar

Municipal Solid Switch grass

Cellulose Waste
Straw
Agricultural
waste Corn Stover
Paper waste
Bagasse

Figure 2.5: Types of feedstock for bioethanol production

Source: Rutz & Janssen (2008) cited in Tekle (2008)

27
2.7 Overview of bioethanol production process and technology

The manufacturing process for bioethanol from sugar crops and grains is nearly

identical to that of producing portable alcohol. It involves first separating the

sugars from the biomass and then fermenting them in the presence of yeast to

produce alcohol and carbon dioxide. Producing bioethanol from cellulose is not

yet commercially viable but is receiving significant attention in research

laboratories, requiring an additional step that first breaks down the fibres to

extract the sugar. Figure 2.6 shows the flow diagram for production of bioethanol

from sugar biomass, starchy biomass and cellulosic biomass.

There are generally two production methods for converting starch grains into

bioethanol; dry milling and wet milling. Dry milling grinds the entire grain into

flour, which is then made into slurry with water and enzymes to convert starch to

dextrose, a simple sugar. This process is called saccharification. Ammonia is

added to control pH and as a nutrient of yeast and the mixture heated to reduce

bacteria. The mash is put in a fermenter where yeast is added to convert the sugar

to bioethanol and carbon dioxide. The resulting mixture is transferred to a

distiller where the bioethanol is dehydrated and concentrated to 200 proofs. The

waste, called “stillage”, is used for animal feed, biogas generation or fertilizer.

The carbon dioxide can be captured and used for carbonation in the beverage

industry. Wet milling first soaks the grain in water and sulphurous acid to

separate out the oil, fibre, gluten and starch. The oil, fibre, and gluten and starch

28
are used for human and animal feed, and other products. The starch is then

fermented into bioethanol in a similar process as in the dry process.

Producing bioethanol from sugarcane and sweet sorghum is a simpler, cheaper

and more efficient process, as the sugars are readily extracted from the crop. The

crop is crushed into sugar juice and crop residue (Bagasse).The sugar juice can be

directly fermented to bioethanol or first processed into crystallized sugar and

molasses. The molasses is then used for bioethanol production, as well as other

agricultural and industrial uses. The Bagasse is burned for energy, which is often

sufficient to operate the plant and produce excess to be sold back into the

electricity grid. In the case of sweet sorghum, the grain can be processed into

food or converted into bioethanol using the saccharification process. Sugarcane

bioethanol plant can increase or decrease the volumes of production of sugar and

bioethanol depending on the variable market demand of each product.

29
 Sugar biomass Starchy biomass Cellulosic biomass
(Sugarcane, beet) (Corn, wheat)

Extraction through Crushing Crushing

pressure of
diffusion

Enzymatic Acid enzymatic

hydrolysis hydrolysis

Fermentable sugar solution

Fermentation

Distillation

Bioethanol

Figure 2.6: Technological routes for bioethanol production.

Source: BNDES & CGEE (2008)

2.8 Major steps in production process

2.8.1 Fermentation

Bioethanol is produced by microbial fermentation of the sugar. Two major

components of plants, starch and cellulose, are both made up of sugars and can in

principle be converted to sugar for fermentation. Currently, only the sugar (e.g.

30
sugarcane) and starch (e.g. corn) portions can be economically converted.

However there is much activity in the area of cellulosic bioethanol, where the

cellulose part of the plant is broken down to sugar and subsequently converted to

bioethanol.

2.8.2 Distillation

For the bioethanol to be usable as fuel, water must be removed by distillation, but

the purity is limited to 95-96% due to the formation of low boiling water –

bioethanol azeotrope. The 95.6% m/m (96.5% v/v) bioethanol, 4.4% m/m (3.5%

v/v) water mixture may be used as a fuel alone, but unlike anhydrous bioethanol,

is immiscible in gasoline, so the water fraction is removed in further treatment in

order to burn in combination with gasoline in gasoline engines.

2.8.3 Dehydration

There are basically a number of dehydration processes to remove the water from

an azeotropic bioethanol/water mixture. The first process, used in many early fuel

bioethanol plants is called azeotropic distillation and involves adding benzene or

cyclohexane to the mixture. When these components are added to the mixture, it

forms a heterogeneous azeotropic mixture in the vapour- liquid - liquid

equilibrium, which when distilled produces anhydrous bioethanol in the column

bottom and a vapour mixture of water and cyclohexane / benzene. When

condensed this becomes a two phase liquid mixture. Another early method, called

extractive distillation, consists of adding a ternary component which will increase

31
bioethanol relative volatility. When the ternary mixture is distilled, it will

produce anhydrous bioethanol on the top stream of the column.

With increasing attention being paid to saving energy, methods have been

proposed that avoid distillation altogether for dehydration. One of these which

have been adopted by majority of modern bioethanol plants is that which uses

molecular sieves to remove water from the fuel bioethanol. In this process,

bioethanol vapour under pressure passes through a bed of molecular sieve beads.

The beads are sized to allow absorption of water while excluding bioethanol.

After a period of time, the bed is regenerated under vacuum or in the flow of inert

atmosphere (e.g. N2) to remove the absorbed water. Two beds are used so that

one is available to absorb water while the other is being regenerated. This

dehydration technology can account for energy saving of 840KJ / litre compared

to earlier azeotropic distillation.

2.8.4 Denaturing

It is necessary to denature the bioethanol i.e. to make it poisonous. This is

optional. This is done to prevent human consumption of the bioethanol as an

alcoholic beverage.

2.9 Bioethanol programs

The perceived benefits of bio-fuels are reflected in the increasing number of

countries introducing or planning to introduce policies to increase proportion of

bio-fuels within their energy portfolio (Dufey, 2006). More than 30 countries

around the world have launched bioethanol fuel programs, with Brazil and the

32
US leading the way. The growth in bioethanol throughout the world has been the

result of favourable fiscal and regulation policies. In 2003 the EU implemented a

policy allowing member states to exempt bio-fuels from taxes (GTZ & GoK,

2008).

2.9.1 Brazil

Brazil was the first country to launch a comprehensive bioethanol program in

1975, called PROALCOOL. The program had the following major elements:

Fixed quota on blending with petrol, incentives for production of engines that

could run on straight bioethanol, Special credit lines for agriculture & industry, a

fixed price for bioethanol below that of petrol and tax exemptions for car owners

purchasing cars that could run on straight bioethanol.

The program was so successful that by 1988 bioethanol had overtaken petrol as

the main source of transport fuel. The drop in global petroleum prices in the early

1990s combined with a rise in global sugar prices, led to a shortage of bioethanol.

The Brazilian government responded by passing a law requiring a 18-26% blend

of bioethanol. By late 1990s the bioethanol industry was liberalized by

deregulating prices. Flex-fuel vehicles (FFVs) that can run on pure bioethanol or

any blend of bioethanol and petrol, now account for about 80% of new vehicles

sold in Brazil.

The Brazilian experience, the PROALCOOL programme to replace imported

gasoline with bioethanol produced from locally grown sugarcane, shows that bio-

fuels can deliver export opportunities and rural development (Dufey, 2006).

33
Brazil meets both its sugar and bioethanol domestic demand and produces

surplus for export depicted in Figure 2.7. The Brazilian bioethanol program

started as a way to reduce the reliance of oil imports, but it was soon realized that

it had important environmental and social benefits (Goldemberg, 2007). This

included phasing out lead additives and MTBE, and reduced sulphur, particulate

matter and CO emissions. It helped mitigate GHG emissions by having a net

positive energy balance (renewable energy output verses fossil fuel input).

Brazilian adoption of mandatory regulations determining the amount of

bioethanol to be mixed with gasoline was essential to the success of the program.

The motivation was to reduce oil imports that were consuming earnings from

export. This policy decision created market for bioethanol and its cost declined as

the production increased at an average rate of 6% (Goldemberg, 2007). The cost

of bioethanol in 1980 was about three times the cost of gasoline, but the

governmental cross-subsidies paid for the price difference at the pump

(Goldemberg, 2007). The subsidies came mostly from taxes on gasoline and thus

were paid by automobile drivers. All fuel prices were controlled by government.

Since 1990’s subsidies have been progressively removed and by 2004 bioethanol

became fully competitive with gasoline on the international markets without

government interventions (Goldemberg, 2007). Subsidies for bioethanol

production are a thing of the past in Brazil, because new bioethanol plants benefit

from the economies of scale and the modern technology (Goldemberg, 2007).

Present policies supporting bioethanol include the following; 20-25% blending

34
mandate, total value added tax, excise and other duties are about half what they

are for petrol and tax breaks for flex-fuel vehicles.

Figure 2.7: International trade in bioethanol in 2006

Source: UNEP (2009)

2.9.2 Germany

Germany launched a bioethanol program requiring a 1.2% blending mandate

(E1.2) which was expected to grow to 3.6% by 2010. According to the program,

suppliers failing to meet the mandate levels were to be penalized, and the share of

bioethanol in each litre of fuel to be fully taxed. In 2006, Germany produced 765

million litres of ethanol, with 75% coming from grains, 22% from non-

agricultural sources, mainly fossil fuels, and 3% from potatoes, sugar beet and

other feed stocks (GTZ & GoK, 2008). The planned main bioethanol feedstock in

the future is sugar beet.

2.9.3 India

In India, molasses and cane juice are the primary feed stocks used for bioethanol

production. In 2006, India adopted an E5 blending policy for part of the country,

which later was expanded nationwide in 2007, and increased to E10 by October

35
2008 due to expected sugarcane surpluses. Fixed prices of bioethanol and

mandatory blending requirements are the main tools employed for achieving the

targets (GTZ & GoK, 2008).

2.9.4 Thailand

Thailand produced about 150 million litres of bioethanol in 2006, mainly from

sugarcane and cassava. Thailand’s sugar industry is one of the worlds’ largest but

suffered under drought in recent years, as well as low yields in terms of

production and sugar content of the cane. Thai government has implemented a

national policy, which oversees the approval of new projects and grants and fiscal

incentives to increase production from domestically produced feedstock. The use

of 10% bioethanol in all government vehicles has been planned, and all producers

of E10 are exempted from certain fuel and excise taxes. The national goal is to

increase gasohol consumption to 20 million litres per day by 2011 (GTZ & GoK,

2008).

2.9.5 United States

The US has overtaken Brazil as the worlds’ leading producer of bioethanol. The

growth in bioethanol production is due to the US government’s subsidies and

incentives. Other factors have spurred the bioethanol production. First, the Clean

Air Act set requirements for oxygenated fuel as a way of combating air pollution.

Bioethanol has replaced MTBE as fuel additive due to impact on water pollution

and it has consequently replaced MTBE as the main oxygenator of petrol in the

US. Secondly, in 2003 the US government launched a major initiative to support

36
corn-based bioethanol. The program set a renewable fuel standard of 28.4 billion

litres by 2012 (GTZ & GoK, 2008).

2.9.6 Ethiopia

There were only three government owned sugar factories in Ethiopia by 2007

with a total annual output of 1.5 million tons of sugar, and with only one factory

producing about 8 million litres per year of bioethanol from molasses available

from its sugar industry (UNDP, 2007). The Ethiopian government has a policy to

blend 5% bioethanol with petrol for transport sector. A UNDP sponsored

program is also looking into using bioethanol as a household fuel for cooking

(GTZ & GoK, 2008).

In response to escalating oil prices, the Government has initiated a strategy for

the development of indigenous fuel production to substitute oil products. One of

the expected actions is the expansion of the bioethanol production and use in the

country and is expected to increase several fold from existing factories and

planned additions. The Government plan is to expand capacity and produce 130

million litres of bioethanol by 2012 (UNDP, 2007).

2.9.7 Malawi

Subsistence agriculture dominates Malawi’s economy and about 84% of the

population lives in rural areas. The country’s sources of energy are biomass

(93%), petroleum products, all imported (3.5%), electricity (2.5%) and coal,

bioethanol and other sources constitute 1% (Robinson J, 2006). Malawi has two

bioethanol plants, one opened in 1982 and the other in 2004, with sugarcane

37
molasses being the feedstock. In 2006, Malawi produced 18.6 million litres of

bioethanol and since early 1980s over 224 million litres have been blended with

petrol (GTZ & GoK, 2008).

2.9.8 Nigeria

Nigeria has been producing bioethanol since 1973 and is in the process of

adopting a 10% blending policy, which would require about one billion litres of

bioethanol per year (Graeme et al., 2010) Nigeria Yeast & Alcohol

Manufacturing PLC, the sole bioethanol plant in the country, produced 30 million

litres of bioethanol in 2006. The Nigerian National Petroleum Company signed

an agreement with Brazils’ Petrobras Company to construct a US$200 million

bioethanol plant. Nigeria is Africa’s largest cassava producer, with an annual

production of over 30 million tons per year, which is a high potential of

producing bioethanol (GTZ & GoK, 2008).

There are several factors constraining bioethanol development in Nigeria, not at

least is its failure to be fully self-sufficient in feeding its teeming population

which is projected to be the 4th largest in the world by 2050 (Graeme et al.,

2010). Due to such food security issues, the use of cassava as energy crops has

attracted criticism. In Nigeria, the agriculture sector is large and sweet sorghum

has been viewed as the promising feedstock for bioethanol production (Graeme et

al., 2010). Bioethanol development will attract private investment to improve

agriculture practices which will accelerate rural development and create wealth.

38
2.9.9 South Africa

South Africa does not have a significant bio-fuel industry. South Africa released

its national bio-fuel strategy in December 2007 (IIED, 2010). This has resulted in

increasing its bioethanol production levels. The strategy sets a target of 8%

blending for bioethanol. The government is encouraging the program with 100%

fuel levy exemption for bioethanol. The policy proposes sugarcane and sugar beet

to be the crops to be used for bioethanol production and excludes the use of

maize due to food security concern. Two large bioethanol plants have a current

production capacity of 97 million litres per year (GTZ & GoK, 2008)

South Africa consumes 0.7% of global petrol, 0.4% of global diesel and 0.3%

global crude oil. Fuel imports costs South Africa an estimated US$ 7 billion per

annum and account 12-20% of all South Africa imports (IIED, 2010). The

destabilizing economic impact of South Africa’s dependence on imported oil is a

key motivation behind the country’s drive to develop a bio-fuel industry.

2.9.10 Sudan

Kenana, the largest sugar company in Sudan, has a 10 year expansion plan to

produce 200,000 litres of bioethanol per day from molasses. 60% of the

feedstock will be produced by the parent company and the remaining 40% by

other sugar mills (GTZ & GoK, 2008).

2.9.11 Uganda

Uganda produces large quantities of sugar and grain that can be used for

bioethanol production but has yet to develop a comprehensive program for

39
harnessing this potential. In Uganda, large quantities of crude bioethanol for

beverage alcohol rather than fuel are produced from molasses, cassava, sorghum

and millet. The country lacks a comprehensive government policy and thus

development of a bio-fuel industry in Uganda has been slow (GTZ & GoK,

2008).

2.10 History and current status of bioethanol in Kenya

As early as the 1980’s, Kenya produced and blended bioethanol and petrol to

produce gasohol. Fifteen years later they abandoned the programme. The record

oil prices of 1970s and 1980s made the government to initiate the gasohol policy.

The policy mandated a 10% bioethanol blend but due to production limitation,

this was only achieved in the Nairobi market. Agro-Chemical and Food Company

(ACFC) based in Muhoroni (Western Kenya) produced all of the bioethanol used

in the programme, thus it had to be transported to Nairobi. The gasohol

programme became uneconomical by the late 1980s due to a number of factors,

including a drop in global oil prices, a surge in the price of bioethanol for

alcoholic consumption in exports markets and a deterioration of bioethanol

production. To bring gasohol to the same retail price as petrol, the Government

had to reduce the customs tariff on gasohol. Even with this subsidy, the

production of gasohol was still not viable (GoK, 2004). The gasohol programme

ceased even though bioethanol is still produced for the alcohol beverage market

for the East African Countries. Presently there are two bioethanol plants, ACFC

and Spectre International. ACFC and Spectre International use molasses as the

40
bioethanol feedstock, producing 60,000 and 65,000 litres per day respectively.

Most of the bioethanol is used by the alcohol beverage markets in Uganda and

DRC and some sold for industrial purpose. To fulfil the full capacities at the two

plants, they will require all the molasses from all the sugar processing companies.

This is not possible since almost half of the molasses currently produced is sold

to farmers and small scale brewers in Uganda who offer higher purchase prices

than the bioethanol plants. The current sugar productivity is not at its optimum,

as they need to introduce fast maturing types so as to meet the demand for the

molasses.

Privately owned Mumias Sugar Company is planning to begin producing

bioethanol from its molasses. The integrated production of sugar, bioethanol and

power that Mumias is planning is a more efficient and sustainable model of

production. Spectre International plans to increase its production capacity from

its current 65,000 litres per day to 230,000 litres per day. Mumias is also

planning a large sugar and bioethanol factory near the Tana River that will use

irrigation to increase yields. Mumias goal is to produce a globally competitive

sugar product while also producing bioethanol for domestic and export market.

With the limitations on available land and competition with food production, the

planned bioethanol production cannot be supplied by sugarcane alone. Even with

increasing sugar productivity or a wholesale shift from sugar to bioethanol

production, about 80% of the planned and current capacity will have to be met

with new feedstock. Spectre International is planning to meet this increased

41
capacity with sweet sorghum and other crops and have rolled out experimentation

on this programme (GTZ & GoK, 2006). ACFC is also thinking of sweet

sorghum as an alternative to molasses. The greatest potential benefits with sweet

sorghum is that it can thrive in drier and marginal agricultural areas than

sugarcane and can compete economically with the cheaply available molasses

due to its high-value grain production (Roman et al., 2010). ICRISAT has a

sweet sorghum program and their offices in Nairobi are willing to provide

agronomy research and development assistance to both bioethanol producing

companies. There are certain important concerns that need to be addressed for a

project to make a feedstock as a source of bioethanol production. These include

the market, productivity of the feedstock production, cost of the feedstock

compared to other feedstock sources, potential production areas, production

technologies, feedstock supply arrangements between feedstock producers and

processing plants, incentives for industry players, and impact on the environment

(Ranola et al., 2009).

In its Agenda for Action in Kenya’s sessional paper No.4 on Energy, the

Government outlined nothing in its short term and long term implementation plan

on bioethanol. The need to commission a study to revisit the viability of power

alcohol has only been mentioned briefly in the sessional paper’s middle term

implementation plan. Thus the Government in the sessional paper didn’t

formulate a policy on how to develop bioethanol (power alcohol) as an

alternative source of energy.

42
2.11 Research gap

In urban and rural areas, energy is needed for the same end uses cooking, water

heating, space heating and lighting, but their energy carrier carriers tend to be

very different. Access to clean, affordable and renewable household energy

sources has been a challenge in the developing countries. Two billion people

worldwide particularly in developing countries lack access to affordable

household energy (Mcnube, 2006). Households in developing countries rely on

traditional fuels (firewood and charcoal) for their cooking and heating needs.

Firewood and charcoal use leads to environmental degradation. With increase in

their incomes, households switch to fossil fuels (kerosene and LPG) and

electricity. Kerosene is not a clean fuel and is not affordable. Despite LPG being

not affordable, Kenya lack storage facilities of LPG and this compromise it

availability. Prices of crude oil also keep on rising. This study seeks to address

this gap by introducing an energy source that is clean, renewable and convenient

as a household fuel.

43
 CHAPTER THREE

METHODOLOGY

3.1 Introduction

This section focuses on the research design and methodology procedures used in

this study. A research design was selected that best achieve the objectives of the

study. Sample size was also carefully selected so as to obtain data that was

representative of the population of the study. Data collection instruments were

prepared to obtain data that when analysed met the objectives of the study.

Primary data was analysed using descriptive statistics and binary logistic

regression to address the demand aspects of household fuels. Secondary data was

analysed to address the supply aspect of bioethanol.

3.2 Research design

The research design used in this study was a survey design. The study first

investigated the different energy and fuel choices available to middle-income

households within Nairobi city and the different factors that could have affected a

household’s probability of choosing a fuel for particular use. This was to address

the demand aspect for household fuels. The study also addressed the supply

aspects of bioethanol fuel. This was done so as to examine the possibility of

introduction of bioethanol as a household fuel.

On the demand aspect, the study determined the household’s fuel consumption

for different household uses and also how the social and demographic factors

influence the choice of a household fuel for a particular use. On the supply

44
aspect, the study reviewed the feedstock resource potential, evaluated key factors

that influenced cost of bioethanol production, and assessed the social and

economic value of bioethanol production. Since the study is a measurement of

characterization of large population, survey study is the appropriate method for

carrying out the research.

3.3 Population and sample

The population from which the sample was drawn was from the middle-income

households within Nairobi city. Convenient sampling was done within these

households even though they could have differed in their socio-economic

characteristics such as age, level of education, kitchen practices, gender, income,

ownership of dwelling unit, type of dwelling unit, area of residence etc. The

study aimed to identify the preferred fuels for the various household uses in

middle-income urban Kenya though sometimes the choice of a given source of

energy will depend on the specific socio-economic characteristics of each

household.

The data was collected from six estates in Nairobi, Githurai, Kahawa, Umoja,

Buru Buru, Satellite and Ngumo. These areas comprise mainly of the middle

income households where comparatively majority of city residents live. The

research avoided households in the extremes i.e. slums and high income

households. The slums were avoided due to security reasons, and also due to

poverty in slums, people were expected not to respond positively as they will

want something that will directly impact on the improvement of their livelihood.

45
Those areas for the high income households were avoided as they are

inaccessible. The gates of the homes of the high income households are manned

by guards from contracted security firms who are under instructions not to allow

anybody in unless authorized by owners of these homes.

Since this is a social science research and the target population is greater than

10,000, the sample size is determined by the following formula;

(6)

Where:

n = the desired sample size

z = the standard normal deviate at the required confidence level

p= proportion in the target population estimated to have characteristic

being measured

q = 1- p

d= the marginal error of p

The proportion of a target population with the characteristic of interest is

recommended to be 0.50 if no estimate is available (Mugenda & Mugenda,

1999). For this study the proportion of the target population estimated to have

characteristic of interest is taken to be 0.75 and a 95% confidence is required,

and therefore the z-statistic is1.96 and the accuracy desired is at 0.05 levels. Thus

the sample size is therefore;

46
In the study, secondary data was also obtained from the websites of stakeholder

in the energy industry, ICRISAT, sugar processing plants, bioethanol producing

plants and government papers.

3.4 Data collection

The data used in the study was collected through questionnaire that aimed to

obtain information on use of various fuels by households. Respondents were

asked about their perceptions, attitudes, behaviours and values about these fuels.

The questionnaire was semi-structured. Secondary data was collected from the

websites of stakeholders in the energy industry, ICRISAT, sugar processing

plants and bioethanol producing plants, books, journals and government

publications.

3.5 Data analysis

The study used a qualitative approach to determine bioethanol fuel supply. The

study proposed the energy crops for bioethanol fuel production, and for each of

the energy crops the study looked at the crop production, harvesting, processing,

and the products and by-products produced. The production balance of each of

the energy crops considered viable for bioethanol production was qualitatively

analyzed from secondary data. Thus the supply aspect reviewed the feedstock

potential, evaluated factors that influenced the cost of production and assessed

by-products of bioethanol that have social and economic value to the community.

For the demand for household fuels, data was analysed quantitatively using SPSS

software. First the study used descriptive statistics to analyse data collected from

47
households. This aimed to establish fuels popular with most households in

cooking, lighting, water heating and space heating. The descriptive analysis was

also used to determine the most preferred fuels when ignoring costs of the fuels,

when considering costs of the fuels and when considering risks they pose. This

also established the perceptions households have about the fuels they use i.e. if a

fuel is inexpensive, available, clean to work with, produce no smoke, fast or safe.

The analyzed data was presented in form of tables and bar-graphs where

applicable.

The study used binary logistic regression (binary logit) analysis to model

relationships between a dependent variable and a number of predictor variables.

Logistic regression chosen since the predictor variables are a mix of continuous

and categorical variables. The logistic regression was used to determine whether

a group of variables together predict a given dependent variable. For this study,

binary logistic regression was used to determine the influence of age, gender,

household size, education, marital status, whether or not the household owns the

dwelling, type of the dwelling unit, and financial status on the type of fuel used in

households. The models created using this analysis were used to ascertain certain

relationships amongst household fuel choices, evaluating factors such education,

income, gender of household head, house ownership and household size, on the

probability of choosing a particular fuel (Ouedraogo, 2005). The statistical

software was used to determine how the explanatory or predictor variables (social

and demographic factors) influenced the dependent variable (household fuel

48
choice). Estimated coefficients for the predictor variables were obtained for the

dependent variable. Estimated coefficient measure the estimated change in the

logit (log odds) for a unit change in the predictor variable while the other

predictor variables are held constant. The coefficient estimates tell the amount of

increase (or decrease, if the sign of the coefficient is negative) in the predicted

log odds of choosing a fuel that would be predicted by a unit increase (or

decrease) in the predictor, holding all other predictors constant. Logistic

regression predicts the log odds of the probability that an observation will occur

to the probability that it will fail to occur. Exponentiating both sides of equation 2

in section 1.7, we obtain;

(8)

This expression defines a multiplicative model for the odds i.e. if we change the

k-th predictor by one unit while holding all other predictor variables constant, we

would multiply the odds by exp (Bk). For example, if the estimated coefficient

(Bk) of a predictor is 1.69 into choosing a particular fuel, then the odds of

choosing the fuel is exp 1.69   5.42 times greater than that of not choosing the

fuel. If the independent variable has dummy variables x, y and z with z as the

reference, then if the estimated coefficient of x is 1.69, then a household is exp

(1.69) = 5.42 times more likely to use the fuel when the dummy variable is x than

when it is z.

The statistical software also determines the p-value of a predictor. The p-value

indicates whether or not a change in the predictor significantly changes the logit

49
at the acceptance level. If the p-value is greater than the significance level of

0.05, then the estimated coefficient is not statistically significant i.e. there is

insufficient evidence that a change in the predictor affects the choice of the

response. If the p-value is less than the significance level of 0.05, then the

estimated coefficient is statistically significant and therefore change in the

predictor affects the choice of the response.

50
 CHAPTER FOUR

RESULTS AND DISCUSSION

4.1 Introduction

This section outlines the results obtained from the data collected. Results

obtained from descriptive statistics and binary logistic regression was intended in

the study to show household’s perspective towards traditional fuels and modern

energy sources. Results from secondary data were intended to address the supply

aspect of bioethanol. The study also discusses the benefits associated with

bioethanol production and use.

4.2 Modern technology

In this study, 313 households were sampled of which 99.7% of the households

had mobile phones, 94.6% had TV sets, 94.2% had radios, 59.4% had computers

and 31% had smokeless (fireless) jikos as depicted in Figure 4.1. Thus a high

proportion of urban households have modern technology gadgets and/or

equipments. Mobile phones are fast and convenient means of communication.

Radios and television sets make households informed of recent news and

developments. Computers can be used for communication as well as sources of

information. The use of fireless jikos makes households to save in energy, thus

lowering energy costs. Mobile phones, radios, TV sets, computers and fireless

jikos improve the livelihood of households. This is an inference that urban

households are likely to accept introduction of new technologies that will

improve their livelihoods.

51
 Figure 4.1: Bar chart showing proportion of households possessing modern
equipment

4.3 Ranking of fuels

The sample size was 313 households and six types of fuels were considered.

These were firewood, charcoal, biogas, kerosene, LPG and electricity. The fuels

were ranked (a) ignoring fuel cost (b) in order of fuel cost (c) in order of risks

posed by fuels. The fuels were ranked 1 to 6 with the most preferred fuel ranked

6 and the least preferred fuel ranked 1. The rest were ranked in between. The

mean of the ranks was calculated and this was used to rank the fuels in each

category. For each category, the fuel with the highest mean is the most preferred

and the one with the lowest mean is the least preferred.

52
4.3.1 Ignoring fuel cost

Ignoring fuel cost, the most preferred fuel was electricity followed by LPG,

biogas, charcoal, kerosene and firewood in that order shown in Figure 4.2.

Electricity and LPG are high in the energy ladder and are therefore clean,

efficient and convenient but expensive. Thus ignoring fuel cost, majority of the

households would prefer to use LPG and electricity than kerosene, charcoal or

firewood.

4.3.2 In order of cost

In order of fuel cost, charcoal was the most preferred fuel, followed by LPG,

biogas, kerosene, firewood and then electricity in that order as shown in Figure

4.3. In urban areas there are no woodlots and this probably makes firewood a

costly fuel. The use of diesel power plants for power generation increases the

tariffs imposed on electricity making it a costly fuel.

4.3.3 In order of risks

In order of risks posed by the fuels, charcoal was the most preferred fuel,

followed by biogas, firewood, LPG, kerosene and electricity in that order. The

means of biogas, firewood, LPG and kerosene were very close as depicted in

Figure 4.4 and thus this ranking was found not appropriate for this study.

53
Figure 4.2: Bar chart for the means of household fuels ignoring fuel cost.

Figure 4.3: Bar chart for the means of household fuels in order of cost.

54
 Figure 4.4: Bar chart for the means of households fuels in order of risks

4.4 Characteristics of fuels

The characteristics of a fuel were sampled from households that use that

particular fuel. It was assumed that households that do not use a particular fuel

could not give the characteristics associated with the fuel. For the households

sampled in the study, very few households used firewood and thus for this part of

the study it was ignored.

For a particular fuel characteristic, the proportion of households that indicated

that a fuel has that characteristic was determined as shown in Table 4.1. These

proportions were based on households that use the fuel. Whether the fuel was

inexpensive, 84% of the households that used charcoal indicated it as an

inexpensive fuel and less than 20% of households that used kerosene, LPG and

electricity indicated them in this respect. Whether the fuel was available, over

90% of households that used charcoal, kerosene and electricity indicated them as

fuels with good availability and 52% of households that used LPG indicated it in

55
 this respect. Whether the fuel was clean to work with, 99% of the households

that used LPG and electricity indicated them as clean fuels to work with and less

than 50% of the households that used kerosene and charcoal indicated them in

this respect. Whether the fuel produce no smoke, over 98% of households that

used LPG and electricity indicated them as fuels that produce no smoke with

45% and 3% of households that used kerosene and charcoal respectively

indicated them in this respect. Whether the fuel was fast when used, 99% of the

households that used LPG and electricity indicated them as fuels that are fast and

40% of households that used charcoal and kerosene indicated them in this

respect. None of the fuels was indicated by more than 70% of the households that

used them as safe

Table 4.1: Percentages of households indicating that a fuel has that

characteristic

Firewood Charcoal Kerosene LPG Electricity

Fuel is inexpensive 74 84 19 6 4
Good fuel availability 56 96 94 52 94
Clean to work with 0 23 46 99 99
No smoke 4 3 45 98 99
Fast 36 40 41 100 99
Safe 81 53 65 50 47
Sample % of
households 7.3 60 43 81 94
using the fuel

56
4.5 Household fuel consumption

For fuel use, the proportion of households using a particular fuel for a given

purpose was determined from the mean. These proportions are shown in Table

4.2. LPG was the fuel used for cooking by most households with charcoal,

electricity and kerosene following in that order. For water heating, most

households used electricity, with charcoal, kerosene and LPG following in that

order. The fuel for lighting for most households was electricity, with very few

households using kerosene. Substantial proportion of households used electricity

for space heating.

Table 4.2: Percentages of households using a particular fuel for a given

purpose
Firewood Charcoal Kerosene LPG Electricity
Cooking 8 59 41 81 45
Water heating 5 36 23 17 71
Lighting 2 2 8 3 94
Space heating 2 5 1 1 67

Although most households indicated LPG and electricity as expensive, majority

of households used LPG for cooking and electricity for lighting. This was

probably because LPG and electricity are clean, fast, efficient and convenient to

use. Bioethanol is a clean and convenient fuel and having similar characteristics

as LPG. Thus bioethanol has potential to replace LPG as a household fuel.

Bioethanol will be expected to have an additional advantage of being available

57
and affordable as it will be domestically produced. LPG is imported and its

supply is unreliable due to lack of storage facilities.

4.6 Binary logistic regression

The research findings established that the fuels for cooking were LPG, charcoal,

electricity and kerosene; for water heating the fuels were charcoal and electricity;

for space heating and lighting the fuel was electricity. Each of these was analyzed

using the binary logistic regression to show how the social and demographic

factors influenced households’ fuel choices. Models were obtained for each using

the forward stepwise (Wald) method and therefore the last step was the one

considered. The study explains only the factors that are significant as they

influence the model after controlling the other factors. The factors that are not

significant do not influence the model after controlling the other factors.

4.6.1 Cooking (Charcoal)

For binary logistic regression results for cooking with charcoal, see Appendix IV.

Inclusion of predictor variables increases the percentage of correct classification

from 62.3% to 64.8%. The model explains 9.8% of the variation in the dependent

variable based on Cox and Snell R square. The regression equation for the model

is:

Logit (Cooking Charcoal) = 20.382 + 0.38*Household size - 0.029*Household

income - 1.694*Spouse gender (1) - 20.144*Spouse occupation (1) -

21.309*Spouse occupation (2) - 20.691*Spouse occupation (3)

(See variable coding in Appendices II & III)

58
The spouse gender for the dummy variable male, spouse occupation for dummy

variables government employed, private/NGO employed and self-employed are

not significant and thus they do not contribute to the model, after controlling the

other factors. The number of persons (household size) and income of households

are significant and thus each contributes to the model after controlling the other

factors.

The odds ratio of household size is 1.463. Therefore when household size

increases by one, the likelihood of households to use charcoal for cooking

increases by a factor of about 1.5. This concurs with the hypothesized influence

and shows that an increase in the household size positively affects a household’s

choice of charcoal. It is sensible to use a cheaper fuel to cook for many people.

The odds ratio of household income is 0.972. Therefore when household income

increases by a thousand, the likelihood of households not to use charcoal for

cooking increases by a factor of 1.3. As a household’s income increases it can

afford the modern fuels which are cleaner and convenient to use for cooking.

4.6.2 Cooking (Kerosene)

For binary logistic regression results for cooking with kerosene see Appendix V.

Inclusion of predictor variables increases the percentage of correct classification

from 60.1% to 80.6%. The model explains 39.3% of the variation in the

dependent variable based on Cox and Snell R square. The regression equation for

the model is:

59
Logit (Cooking kerosene) = 4.638 + 1.642*Spouse education (2) – 0.081*Spouse

age - 0.064*Household income + 0.918*Spouse education (1) –

0.305*Household head occupation (1) - 1.614*Household head occupation (2) –

0.395*Household head occupation (3)

(See variable coding in Appendices II & III)

The spouse education for the dummy variable primary, household head

occupation for the dummy variables government employed, private/NGO

employed and self-employed are not significant and thus they do not contribute to

the model, after controlling the other factors. The spouse education for the

dummy variable secondary, spouse age and income of households are significant

thus each contribute to the model after controlling the other factors.

The odds ratio of spouse education when the dummy variable is secondary is

5.164. Therefore when the spouse education level is secondary, households are

about 5.2 times more likely to use kerosene for cooking than when the education

level is post-secondary. This supports the study’s theoretical expectation as level

of education improves knowledge of fuel attributes and preferences of better

cleaner fuels. Kerosene is above charcoal but below LPG in the energy ladder.

The odds ratio of spouse age is 0.922. Therefore when spouse age increases by

one, the likelihood of households not to use kerosene for cooking increases by a

factor of 1.08. This supports the study’s theoretical expectation that households

will continue to use traditional fuels such as charcoal as the age of the spouse

increases.

60
The odds ratio of household income is 0.938. Therefore when household income

increases by a thousand, the likelihood of households not to use kerosene for

cooking increases by a factor of 1.07. This supports the study’s expectation that

as a household’s income increases, households can afford to switch to fuels

higher in the energy ladder.

4.6.3 Cooking (LPG)

For binary logistic regression results for cooking with LPG see Appendix VI.

Inclusion of predictor variables increases the percentage of correct classification

from 82.1% to 90.5%. The model explains 37% of the variation in the dependent

variable based on Cox and Snell R square. The regression equation for the model

is:

Logit (Cooking LPG) = 15.016 + 3.087*House type (1) – 2.260*House

ownership (1) + 0.104*Household head income - 16.204*Household head

occupation (1) - 15.872*Household head occupation (2) – 17.604*Household

head occupation (3)

(See variable coding in Appendices II & III)

The household head occupation for dummy variables government employed,

private/NGO employed and self employed are not significant and thus they do

not contribute to the model. The house type for dummy variable modern, house

ownership for the dummy variable rental and household income are significant

and thus each contribute to the model after controlling the other factors.

61
The odds ratio of house type when the dummy variable is modern is 21.911.

Therefore when the house is modern, households are about 22 times more likely

to use LPG for cooking than when it is traditional. Households living in modern

dwelling units are wealthier and thus can afford LPG.

The odds ratio of house ownership when the dummy variable is rental is 0.104.

Therefore when the house ownership is rental, households are 9.6 times less

likely to use LPG for cooking than when they own the house. This does not

concur with the theoretical expectation that households living in rented dwelling

units are likely to use cleaner fuels such as LPG, fuels that do not produce smoke

which stains walls and roofs.

The odds ratio of household income is 1.110. Therefore when household income

increases by a thousand, the likelihood of households to use LPG for cooking

increases by a factor of about 1.1. As household income increases, households

can afford to continue using cleaner and convenient fuels such as LPG.

4.6.4 Cooking (Electricity)

For binary logistic regression results for cooking with electricity see Appendix

VII. Inclusion of predictor variables increases the percentage of correct

classification from 50.9% to 77.3%. The model explains 38.9% of the variation in

the dependent variable based on Cox and Snell R square. The regression equation

for the model is:

62
Logit (Cooking electricity) = -9.247 + 2.713*Household head gender (1) +

0.033*Spouse age + 3.605*Spouse occupation (1) + 2.665*Spouse occupation

(2) + 1.918*Spouse occupation (3) + 0.095*Household income

(See variable coding in Appendices II & III)

The spouse occupation for the dummy variable private/NGO employed and self-

employed are not significant and thus they do not contribute to the model after

controlling other factors. The household head gender for the dummy variable

male, spouse age, spouse occupation for the dummy variable government

employed and income of households are significant and thus each contribute to

the model after controlling the other factors.

The odds ratio of household head gender when the dummy variable is male is

15.078. Therefore when the household head gender is male, households are about

15 times more likely to use electricity for cooking than when the gender is

female. Males generally do not cook and therefore they are likely to prefer to use

a fuel or an energy source that is fast, clean and convenient such as electricity.

The odds ratio of spouse age is 1.034. Therefore when spouse age increases by

one, the likelihood of households to use electricity for cooking increases by a

factor of 1.03. This does not agree with the theoretical expectation that

households will continue to use traditional fuels as they become older.

The odds ratio of spouse occupation when the dummy variable is government

employed is 36.780. Therefore when the spouse is employed by government,

households are about 37 times more likely to use electricity for cooking than

63
when the spouse is retired. Government employed spouses have monthly income

unlike retired spouses, and therefore can afford to pay high electricity bills due to

cooking using electricity.

The odds ratio of household income is 1.100. Therefore when household income

increases by a thousand, the likelihood of households to use electricity for

cooking increases by a factor of 1.1. As household income increases, households

can afford to continue using cleaner and convenient fuels.

4.6.5 Water heating (electricity)

For binary logistic regression results for water heating with electricity see

Appendix VIII. Inclusion of predictor variables increases the percentage of

correct classification from 73.6% to 85.7%. The model explains 34.9% of the

variation in the dependent variable based on Cox and Snell R square. The

regression equation for the model is:

Logit (Water heating electricity) = - 1.394 - 1.728*Spouse education (1) -

1.032*Spouse education (2) + 0.127*Household income

(See variable coding in Appendices II & III).

The spouse education for the dummy variables primary and secondary, household

income are significant and thus each contributes to the model after controlling the

other factors.

The odds ratio for spouse education when the dummy variable is primary is

0.178. Therefore when the spouse education level is primary, households are

about 5.6 times less likely to use electricity for water heating than when the

64
education level is post-secondary. The reason for this could be probably

affordability. Persons with primary level of education are normally employed at

lower job grades than persons with post-secondary level of education, thus the

former has lower income than the latter.

The odds ratio of spouse education when the dummy variable is secondary is

0.356. Therefore when the spouse education is secondary, households are about

2.8 times less likely to use electricity for water heating than when the spouse

education level is post-secondary. Persons with secondary level of education are

normally employed at lower job grades than persons with post-secondary level of

education and therefore the former has lower income than the latter.

The odds ratio of household income is 1.136. Therefore when household income

increases by a thousand, the likelihood of households to use electricity for water

heating increases by a factor of about 1.1. As household income increases,

households are likely to afford to pay for electricity bills due to using it for water

heating.

4.6.6 Water heating (charcoal)

For binary logistic regression results for water heating using charcoal see

Appendix IX. Inclusion of predictor variables increases the percentage of correct

classification from 61.5% to 64.5%. The model explains 4.7% of the variation in

the dependent variable based on Cox and Snell R square. The regression equation

for the model is:

Logit (Water heating charcoal) = - 1.927 + 0.298*Household size

65
(See variable coding in Appendices II & III)

The number of persons (household size) is the only factor in the model. It is

significant and thus contributes to the model.

The odds ratio of household size is 1.348. Therefore when household size

increases by one, the likelihood of households to use charcoal for water heating

increases by a factor of about 1.35. This supports the study’s theoretical

expectation that households with many people are likely to use a cheaper fuel

such as charcoal.

4.6.7 Space heating (electricity)

For binary logistic regression results for space heating using electricity see

Appendix X. Inclusion of predictor variables increases the percentage of correct

classification from 67.8% to 76.9%. The model explains 21.5% of the variation in

the dependent variable based on Cox and Snell R square. The regression equation

for the model is:

Logit (Space heating electricity) = -4.709 + 1.694*spouse religion +

1.579*Spouse education (1) + 1.823*Spouse education (2) + 3.21*House type (1)

- 0.652*House ownership (1) + 0.022*Household income

(See variable coding in Appendices II & III)

Spouse religion, spouse education for the dummy variables primary and

secondary, house type for dummy variable modern, house ownership for dummy

variable rental, household income are significant and thus each contribute to the

model after controlling the other factors.

66
The odds ratio of spouse religion when the dummy variable is Christian is 5.443.

Therefore when the spouse religion is Christian, households are about 5.4 times

more likely to use electricity for space heating than when the spouse is of other

religion other than Muslim. There is no good reasonable explanation for this.

The odds ratio of spouse education when the dummy variable is primary is 4.852.

Therefore when the spouse education level is primary, households are about 4.9

times more likely to use electricity for space heating than when the education

level is post-secondary. This is not as would be expected since persons with post-

secondary education have bigger incomes than those with primary education,

hence expecting the former to afford to pay electricity bills due to using

electricity for space heating.

The odds ratio of spouse education when the dummy variable is secondary is

6.190. Therefore when the spouse education level is secondary, households are

about 6.2 times more likely to use electricity for space heating than when the

education level is post-secondary. This is not as would be expected since persons

with post-secondary education have bigger incomes than those with secondary

education, hence expecting the former to afford to pay electricity bills due to

using electricity for space heating

The odds ratio of house type when the dummy variable is modern is 24.806.

Therefore when the house type is modern, households are 25 times more likely to

use electricity for space heating than when the house is traditional. Households

67
living in modern houses are wealthier and thus are likely to afford to pay for

electricity bills.

The odds ratio of house ownership when the dummy variable is rental is 0.521.

Therefore when the house ownership is rental, households are 1.7 times less

likely to use electricity for space heating than when they own the house.

Households living in rental dwelling units have rents to pay and thus will likely

try to avoid unnecessary additional expenses.

The odds ratio of household income is 1.02. Therefore when household income

increases by a thousand, the likelihood of households to use electricity for space

heating increases by a factor of 1.02. As household income increases, households

are likely to afford to pay for electricity bills due to using it for space heating.

4.6.8 Lighting (electricity)

For binary logistic regression results for lighting using electricity see Appendix

XI. Inclusion of predictor variables increases the percentage of correct

classification from 94.9% to 97.1 %. The model explains 21.3% of the variation

in the dependent variable based on Cox and Snell R square. The regression

equation for the model is:

Logit (Lighting electricity) = 11.627 - 0.934*Household size - 0.151*Spouse age

+ 2.897*House type (1) – 4.578*House ownership (1) + 0.168*Household

income

(See variable coding in Appendices II & III)

68
Household size, spouse age, house type for the dummy variable modern, house

ownership for the dummy variable rental, household income are significant and

thus each contribute to the model after controlling the other factors.

The odds ratio of household size is 0.393.Therefore when household size

increases by one, the likelihood of households not to use electricity for lighting

increases by a factor of 1.5. There is no reasonable explanation for this.

The odds ratio of spouse age is 0.860. Therefore when spouse age increases by

one, the likelihood of households not to use electricity for lighting increases by a

factor of 2.4. There is no explanation for this.

The odds ratio of house type when the dummy variable is modern is 18.120.

Therefore when the house type is modern, households are about 18 times more

likely to use electricity for lighting than when the house is traditional.

Households living in modern dwelling units are generally wealthier than those

living in traditional dwelling units, thus the former are more likely to afford to

use electricity than the latter.

The odds ratio of house ownership when the dummy variable is rental is 0.010.

Therefore when the house ownership is rental, households are about 97 times less

likely to use electricity for lighting than when they own the house. There is no

reasonable explanation for. Electricity is cheaper and convenient for lighting than

either kerosene or LPG.

The odds ratio of household income is 1.182. Therefore when household income

increases by a thousand, the likelihood of households to use electricity for

69
lighting increases by a factor of about 1.2. As household income increases,

households are likely to afford to pay for electricity bills due to using it for

lighting.

4.7 Bioethanol supply

Kenya has suitable climatic conditions for growing bioethanol crops and thus

capable of developing its own supply of domestically produced bioethanol fuel.

These can have enormous economic, social and environmental benefits.

4.7.1 Bioethanol feedstock

Bioethanol crops that can be grown in Kenya include potatoes, cassava, wheat,

maize, sugarcane and sweet sorghum. Some of this feedstock such as potatoes,

cassava, maize and wheat were avoided in the study as their use in bioethanol

production would present a threat to food security. The bioethanol crops selected

for this study were sugarcane and sweet sorghum. The two were selected due to

the history of production, climatic suitability, crop cycles per year and also they

are not expected to compete with the food supply chain.

Sugarcane has a high potential for bioethanol production (ICRISAT, 2006).

Kenya grows sugarcane which is primarily for local consumption and is not

enough to meet the local demand and it therefore makes some imports.

Bioethanol is made from molasses at two plants i.e. Spectre International and

Agrochemical Food and Allied Company. Mumias Sugar Company is also

producing bioethanol by utilizing its own molasses.

70
Sweet sorghum which is similar to grain sorghum with sugar rich stalks and is a

water-use efficient crop has a very good potential for bioethanol production.

Sweet sorghum is a crop similar to sugarcane but with a lower sugar-product

yield and higher tolerance to warmer and drier conditions. ICRISAT considers

sweet sorghum as a smart crop as it produces food, feed, fodder and fuel, without

significant tradeoffs in any of these uses in the production cycle. Sweet sorghum

growing period (about 4-5 months) and water requirement (8000m3 over two

crops) are 4 times lower than those of sugarcane (12-16 months duration and

36000m3 of water per crop) as shown in Table 4.3. Though sugarcane is a good

feedstock for bioethanol production than sweet sorghum, its cost of cultivation is

relatively higher as indicated in Table 4.3.

71
Table 4.3: Characteristics of sweet sorghum and sugarcane
Parameter Sweet sorghum Sugarcane
Crop duration 4 months 12-16 months
Water requirement 400m3 3600 m3
Grain yield (tonnes/Ha) 2.0 -
Bioethanol from grain (litres/Ha) 760 -
Green stalk cane yield (tonnes/Ha) 35 75
Bioethanol from stalk cane juice
1400 5600
(litres/Ha)
Stillage (tonnes/Ha) 4 13.3
Bioethanol from residue (litres/Ha) 1000 3325
Total bioethanol (litres/Ha) 3160 8925
Cost of cultivation (US$/Ha) 220 995
Cost of cultivation & irrigation
239 995
(US$/Ha)
Bioethanol cost per kilolitre (US$) 75.3 111.5
Source: Adopted and modified from ICRISAT (2006)

4.7.2 The availability of bioethanol feedstock

Land availability, agricultural practices and demand from competing uses are the

main factors that determine the supply and price of bioethanol feedstock. The

research reviewed a study conducted by German Technical Cooperation (GTZ)

and the Government of Kenya (GoK) in 2008 and the findings were that land

availability and agricultural practices determine the supply of bioethanol

feedstock. The study considered three scenarios to determine potential

availability of bioethanol feedstock. The first scenario considered the status quo

production of feedstock and found that enough sugarcane was being produced for

72
 49 million litres of bioethanol if only molasses was being used and 345 million

litres if all sugar cane went to bioethanol instead of sugar as depicted in Table

4.4.

The second scenario considered potential production of feedstock that could be

grown at current yields if half of all the suitable areas were planted with the

energy crops. The production of each feedstock was based on suitability mapping

done by GIS experts at ICRAF. Excluding land that was being used to grow other

crops, enough sugarcane could be grown to produce 30-270 million litres of

bioethanol depending on whether molasses or cane juices was used as shown in

Table 4.5. Over 8 billion litres of bioethanol could be produced from sweet

sorghum.

The third scenario considered optimized production of feedstock if half of

suitable lands were used at optimal yields for each crop. This scenario considered

production using high yielding varieties under irrigation. Higher values were

obtained than those in the second scenario as depicted in Table 4.6. The

“optimized scenario” show what difference higher yields would make in terms of

bioethanol feedstock.

Table 4.4: Status Quo Feedstock Production Scenario

Non-food competing Food competing
Hectares Yield Production Bioethanol Production Bioethanol
(‘000) (T/HA) (‘000 tons) (‘000 litres) (‘000 tons) (‘000 litres)
Sugarcane 147.7 33.4 4,933 49,330 4933 345,310
Source: Adopted and modified from GTZ & GoK (2008).

73
Table 4.5: Potential Feedstock Production Scenario
New Farm Lands Existing Farm Lands

Yield Land Production Bioethanol Land Production Bioethanol

T/Ha MHa* Mtons* Mlitres* MHa* Mtons* Mlitres*
Sorghum 35 5.90 206.50 8,260 11.06 387.10 15,484

Sugarcane 33.4 0.09 3.01 30 0.83 27.72 277

*MHa refers to 106 Ha, Mtons refers to 106 tons and Mlitres refers to 106litres.
Source: Adopted and modified from GTZ & GoK (2008).

Table 4.6: Optimized Potential Feedstock Production Scenario

New Farm Lands Existing Farm lands
Yield Land Production Bioethanol Land Production Bioethanol
T/Ha MHa Mtons Mlitres MHa Mtons Mlitres
Sorghum 70.00 5.90 413 16,520 11.06 774 30,968
Sugarcane 68.84 0.09 6.15 61.5 0.83 56.72 567
Source: Adopted and modified from GTZ & GoK (2008).

The research findings established that available land and increased yields are

both important factors in producing adequate supplies of bioethanol feedstock to

support a domestic industry. High yielding varieties of both sugarcane and sweet

sorghum have been developed through the efforts of Kenya Sugar Board and

ICRISAT respectively. Land is available in Kenya to grow enough sugarcane and

sweet sorghum to produce enough bioethanol to be used as a fuel in urban

households. The calculation in the two potential scenarios was based on 50% of

the available suitable land and therefore it is possible to supply more of the

feedstock for bioethanol production. If Kenya embarks on an expansion plan to

74
supply bioethanol feedstock, enough bioethanol fuel for urban households will be

available. The technical potential of bioethanol is therefore enormous.

There are other inputs such as electricity, labor, transport and chemicals (yeast

and enzymes) required for bioethanol production. Electricity and transport are the

most critical as these can drive up the cost of production. The cost of electricity

can be brought down by the processing plants by utilizing Bagasse from both

sugarcane and sweet sorghum to generate electricity. This is a lesson learned

from Brazil which is a success story in bioethanol production. The transport cost

can be lowered by ensuring that bioethanol plants are located where the feedstock

is produced. The available cheap labor and the potential to co-generate electricity

from the Bagasse can help to offset the added costs of the non-feedstock inputs.

4.8 Energy balance

The energy balance for the production of bioethanol is the ratio of the energy

output (bioethanol, electricity and by-products value) to the input (fuel, water,

fertilizers, chemicals, machinery and labour). Energy balances need to consider

the entire life cycle from feedstock production to final consumption. Assessments

should also include energy paybacks associated with co-products. Energy

balances vary depending on the type of the feedstock, method of cultivation and

the conversion technology. The energy balance also depends on the methodology

used in assessment studies.

Brazilian sugarcane-based bioethanol is one of the most energy efficient forms of

bioethanol. From an assessment study (Macedo et al., 2008) in the south region

75
 of Brazil, for each fossil energy unit used to produce sugarcane bioethanol, more

than nine renewable energy units are produced in the form of bioethanol and

surpluses of electric power and Bagasse as shown in Table 4.7. Different studies

in different countries give different energy balances for each crop as depicted in

Table 4.8. This indicates there are a variety of methods used to obtain energy

balance of bioethanol. Each of the approaches however indicates a positive

energy balance of bioethanol and therefore more energy is produced than

consumed.

Table 4.7: Energy Balance of Sugarcane Bioethanol Production in Brazil

Energy balance component 2005/2006 2020 scenario
Sugarcane production and transport 210.2* 238.0*
Bioethanol production 23.6* 24.0*
Fossil input (total) 233.8* 262.0*
Bioethanol 1,926.0* 2,060.0*
Bagasse surplus 176.0* 0.0*
Electricity surplus 82.8* 972.0*
Renewable output (total) 2184.8* 3,032.0*
Energy production/consumption
Bioethanol + Bagasse 9.0 7.9
Bioethanol + Bagasse +electricity 9.3 11.6
*Units are Megajoules per ton (MJ/ton)

Source: Macedo et al., 2008 cited from BNDES, CGEE (2008).

76
Table 4.8: Energy Balance for Bioethanol Production for different countries
Sugarcane Sweet sorghum Country (source)
21.3 6.9 Brazil (Da Silva, 1978)
1.9 Zimbabwe (Rosenchein et al.)
9.2-11.2 Brazil (Macedo, 1996)
3.4-6.1 Spain (Fernandez,1998)
0.9-1.1 USA/Europe (Santos,1997)
3.5-7.9 USA (Worley et al., 1991)
Source: Zuzarte F. (2007).

4.9 Bioethanol fuel and stove

Bioethanol can be utilized as a fuel using stoves adapted for their use. The stoves

in use include the CleanCook stove, the SuperBlue stove and the Gel fuel stove

shown in Figure 4.5, 4.6 and 4.7 respectively. The bioethanol burns on these

specialized stoves emitting heat similar in temperatures to the heat output of LPG

cookers. Bioethanol is denatured to prevent its consumption as a beverage. The

Clean Cook and Super Blue stoves use lean bioethanol but the Gel stove use a

viscous liquid (gel). Research work is currently on-going to improve on the

design of the bioethanol gel stove (Obara et al., 2012).

The bioethanol is a renewable, clean and efficient fuel similar in characteristics to

LPG. The bioethanol burns in specialized stoves and emit heat similar in

temperatures to the heat output of LPG cookers. The bioethanol also burns in a

smokeless flame similar to that of LPG. Thus bioethanol has potential or

possibility of substituting LPG as a household fuel.

77
Figure 4.5: The Domestic two burner CleanCook stove
Source: UNDP (2007)

Figure 4.6: The SuperBlue stove

Source: Robinson (2006)

78
 Figure 4.7: The Gel fuel stove and a bottle of fuel
Source: Robinson (2006)

The bioethanol is made into a gel by enriching with chemical additives to

enhance its physical and chemical composition for increased efficiency. The

result of the enrichment is a viscous liquid (gel) that burns slowly and emits a

higher heat output. The viscosity of the gel is designed to make it safer by

minimizing cases of accidental spillage. For the Kenyan case, the duty levied on

the production of bioethanol gel is high and thus companies that make the gel are

first export bioethanol to neighboring Tanzania for value addition and then bring

it back. According to a company called Consumer Choice that sell bioethanol gel

and stove, the bioethanol gel retailed at recommended price of Ksh.160 per litre

and the stove for Ksh.1800.(Business Daily, 2011). Producing the gel in Kenya

would double the price of bioethanol gel. Thus there is need for the government

to revise and lower the duty and make the bioethanol gel affordable.

79
4.10 Bioethanol co-products

The cost of production of bioethanol can be improved through the use and sale of

co-products obtained during various production processes. The feedstock outputs

are similar; crop residues, carbon dioxide and residues from fermentation and

distillation processes. The crop residues can be used for cogeneration in boilers to

produce steam or electricity to be used in the plant. In industrial processing of the

feedstock, thermal energy is required for heating and concentrating processes,

mechanical energy for milling and mechanically driven systems and electrical

power for powering pumping, control systems and lighting. Bioethanol plants can

produce these energies by using crop residues (Bagasse) as their fuel. This lowers

the production cost of bioethanol significantly. The excess power can be offered

for sale to the national electricity grid.

Crop residues can be used as animal feed especially with the sweet sorghum.

Also with sweet sorghum, after fermenting the grains and distilling the by-

product through dehydration, DDG (Dried distiller grain) is produced. This DDG

is of value as an animal feedstock. The DDG can also provide valuable manure

for use as fertilizer to enhance production of the feedstock. Figure 4.8 and 4.9

shows the flow charts for production balance of sugarcane and sweet sorghum

respectively.

Other than molasses, Bagasse, yeast, filter cake and Vinnase which are the

traditional co-products of sugarcane a growing and varied list of new products

have been developed in Brazil. The list ranges from flavor enhancers for the food

80
industry to packing plastics. The Brazilian sugarcane agro-industry is diversified

with other industrial sectors such as the food and chemical sectors incorporating

sugarcane by-products as raw materials as depicted in Table 4.9. It should be

noted that these new products add significant value and the necessary

investments in plant infrastructure are relatively minor in the context of the

overall cost of the bioethanol plant.

81
 Sugarcane

Bagasse
Milling Boilers

Chemical
Treatment

Filtration Filter cake

Evaporation Molasses Fermentation

Distillation Vinnase
Cooking

Centrifugation
Rectification

Drying

Bioethanol
(Hydrated) Dehydration
Sugar

Bioethanol
(Anhydrous)

Figure 4.8: Sugar and sugarcane-based bioethanol production flowchart

Source: Seabra (2008) cited in BNDES & CGEE (2008).

82
 Harvesting
C Biomass Animal feed
Silo

Sweet sorghum
planting
Poultry and animal
feed

Harvesting stalk, grain

and leaf separately Hydrolysis Centrifugation

Juice extraction Filter

Cake

Fermentation Alcohol

Animal feed

Figure 4.9: Bioethanol production and by-products from sweet sorghum.

Source: Almodares & Hadi (2009)

83
 Table 4.9: New products from the sugar cane agro-industry
Family Feedstock Products

Biotechnology: Molasses a. Citric acid

Materials produced on b. Amino acid: lysine
the biological function c. Agrochemicals: Growth regulator or
of living organisms phytoregulators (indolacetic acid,
jasmonic acid), pesticide (bio-fungicide,
biological controller, biological
insecticide, biological pesticide)
d. Nitrogen fixer
e. Silage inoculums
Chemical: Products Molasses, a. Industrial inputs (technical dextran,
resulting from Bagasse and calcium gluconate, mannitol, sorbitol and
chemical reactions Vinnase biodegradable surfactants).
carried out with or b. Furfural (xylose liquor, furfural, furfural
without catalysts. alcohol, furano – epoxy compounds, wood
preservative, casting resin).
c. Plastics (PHB and PHB/hl, PHA mcl/PHB
hpe).
d. Inputs for the industry of paper and
cellulose (corrugating means,
chemothermomechanic pastes, filtering
materials).

Veterinary drugs: Molasses and a. Anti-diarrheic syrup

chemical, biological, Bagasse b. Ferrous- dextran complex
biotechnological c. Probiotic
substances or
manufacturing
preparations, given
directly or mixed to
the food, to prevent
and treat animal
diseases.

Food Molasses, a. Yeast, fructose and glycols by-products

Bagasse and b. Fructo-oligosaccharides
Vinnase c. Inverted syrups by enzymatic pathway
d. Edible mushrooms of the species
Pleurotusostreatus.
Biologics Bagasse a. Fertilizing compound
Structural: Materials Bagasse b. Bagasse/cement pellets
whose materials make
them useful in c. MDF pellets
structures, machines or
consumable products.
Source: BNDES & CGEE (2008)

84
4.11 Employment and income

To produce adequate supplies of bioethanol feedstock to meet domestic

requirement, this will require expanded agricultural production of these

feedstock. This will provide farm jobs and opportunities for rural farmers to

expand production into new cash crops. Sweet sorghum can grow in semi-arid

areas, thus its introduction as a cash crop in such areas is an economic activity

that will improve the livelihood of the rural population in these regions. An

additional production of bioethanol will yield revenue to the national economy.

The additional revenue could be used to invest in irrigation and better agriculture

practices which will go along in increasing yields of all crops, even those grown

for food. New jobs will also be created in the manufacturing and transport sectors

by the planned bioethanol production.

Domestically produced bioethanol will reduce importation of petroleum products

such as kerosene, LPG and fuel oil leading to saving in foreign currency. This

can be reinvested in the economy and this will go along in creating new jobs and

new opportunities especially for the underdeveloped rural areas.

The potential reduction in GHG’s emissions from bioethanol will provide a

revenue stream through carbon credits. There are the mandatory markets created

through the Kyoto Protocol of UNFCCC and the market for voluntary credits.

The former is more stringent and restrictive but yields higher price per tonne of

carbon. The latter is more flexible and easier to gain compliance with, but fetches

lower price. There will be carbon funding through the Clean Development

85
Mechanisms (CDM) of the Kyoto Protocol. This enables countries to meet their

emission obligations by funding projects that reduce emissions in other parts of

the world, especially developing countries that are not bound to reduce emissions

under the Climate Convention.

4.12 Environmental and social impacts

Bioethanol production and use can have both positive and negative

environmental and social impacts depending on type of feedstock used, methods

of production as well as scale of production. The environmental and health

benefits of bioethanol are quite significant. Both fossil and bioethanol fuels when

combusted emit carbon monoxide, sulphur oxides, nitrogen oxides, aromatics,

volatile organic compounds and particulate matter but the emissions from

bioethanol are lower than those from fossil fuels. Some of these air emissions are

carcinogenic, others cause lung and heart diseases and thus their reduction will

yield large public health benefits.

Another benefit is the reduction in carbon dioxide emissions when sugarcane and

sweet sorghum plants use Bagasse to generate electricity to power their

operations rather than use fossil fuels. The plants can also produce biogas from

biomass residues of its operation, which can also be used to offset the use of

fossil fuels. Carbon dioxide is a GHG largely caused by combustion of fossil

fuels in power plants, automobiles and industrial facilities. Thus bioethanol

production and use as a fuel presents an opportunity to help mitigate climate

change by reducing carbon dioxide emissions from fossil fuels.

86
Bioethanol production utilizes large quantities of water and produces significant

wastewater effluents. In Brazil, 1000-2000 litres of water are used to process a

tonne of sugarcane into sugar and bioethanol (GTZ & GoK, 2008). Therefore,

bioethanol plants require efficient treatment plants to process this wastewater

before its discharge into the environment.

In selecting feedstock for bioethanol production, the feedstock chosen should not

compete with those grown for food and to supply animal feed. Maize, wheat

potatoes and rice are normally grown to meet the demand for food and animal

feed. Kenya is not self sufficient in food and makes imports of the grains

especially maize, wheat and rice. Thus these grains cannot be selected to be

grown for bioethanol production. Even though Kenya is not self sufficient in

sugar, molasses a by-product of sugar production from sugarcane is a good raw

material for bioethanol production. Therefore by increasing acreage of land

grown with sugarcane makes it an ideal feedstock for bioethanol production.

Sweet sorghum produces grain which can be used to meet food demand and its

stalk which contains brix that can be extracted for bioethanol production.

4.13 Roadmap to bioethanol production

Through government, private sector and development partners support, Kenya

can be transformed to become a bioethanol producing country like Brazil.

Bioethanol production will improve the economies of the population living in

areas whose climatic conditions are suitable for sugarcane and sweet sorghum.

The government through the ministry of energy and ministry of agriculture need

87
to develop policies that will promote bioethanol production. These policies

should aim to achieve certain targets.

Sugarcane is presently grown in Kenya and therefore for its expansion program

what is required is how much of the suitable land should be cultivated with

sugarcane per year to meet the target. Also the fast maturing varieties need to be

introduced. Sweet sorghum is a new crop and thus its expansion program need to

be well coordinated through the ministry of agriculture’s research and extension

services with the collaboration of ICRISAT who can play a big role in its

development. ICRISAT has been researching on sweet sorghum in India and

Thailand and have developed a number of varieties.

To develop a value chain of bioethanol, support from different stakeholders will

be required. Bioethanol producing plants should be ready to support out growers

who grow sugarcane and sweet sorghum; especially the latter which is a new

crop and farmers should be assured of its market. Institutions such as KARI,

KEFRI and KEPHIS should play roles in training farmers, providing extension

services, and testing, certifying and distributing seeds.

88
 CHAPTER FIVE

CONCLUSION AND RECOMMENDATIONS

5.1 Introduction

This section gives the summary of the research findings and the main conclusions

drawn from the results. The researcher also offer recommendations for

overcoming the challenges and hindrances to bioethanol production and hence its

use as a household fuel.

5.2 Conclusion

The findings of the research established that of the 313 households sampled,

59.4% had a computer, 94.2% had a radio, 94.6% had a TV, 99.7% had a mobile

phone and 31% had a “fireless jiko “. It was noted that most households had the

modern gadgets or equipments for communication and also for making them

informed of new developments. Some households were also aware of a way to

make saving in energy by using a “fireless jiko”. Thus this was an indication that

urban households are likely to accept new technologies and ventures that impact

positively on their way of livelihood.

In ranking of the fuels, the study findings were that ignoring fuel cost the most

preferred fuel was electricity with a mean value of 5.27, followed by Liquid

petroleum gas (LPG ) with 5.18, biogas with 4.00, charcoal with 2.75, kerosene

with 2.36 and firewood with 1.48. Considering cost of fuel, the most preferred

fuel was charcoal with a mean value of 4.44, followed by LPG with 3.81, biogas

with 3.49, kerosene with 3.31, firewood with 3.09 and electricity with 2.86. In

89
order of risks, charcoal had a mean of 3.94, followed by biogas with 3.59,

firewood with 3.51, LPG with 3.5, kerosene with 3.46 and electricity with 3.05.

In order of risks, the means of biogas, firewood, LPG and kerosene are very

close.

The findings of the research were that of the households sampled, 7.3% used

firewood as a fuel, 43% used kerosene, 60% used charcoal, 81% used liquid

petroleum gas (LPG) and 94% used electricity. Thus a small proportion of the

households used firewood as a fuel. Households were found to be using one or

more than one fuel. Whether the fuel was inexpensive, the findings of the study

were that 84% of the households that used charcoal indicated it as an inexpensive

fuel with less than 20% of households that used kerosene, LPG and electricity

indicated them in this respect. Whether the fuel was available, over 90% of

households that used charcoal, kerosene and electricity indicated them as fuels

with good availability and 52% of households that used LPG indicated it in this

respect. Whether the fuel was clean to work with, 99% of the households that

used LPG and electricity indicated them as clean fuels to work with and less than

50% of the households that used kerosene and charcoal indicated them in this

respect. Whether the fuel produce no smoke, over 98% of households that used

LPG and electricity indicated them as fuels that produce no smoke with 45% and

3% of households that used kerosene and charcoal respectively indicated them in

this respect. Whether the fuel was fast when used, 99% of the households that

used LPG and electricity indicated them as fuels that are fast and 40% of

90
households that used charcoal and kerosene indicated them in this respect. None

of the fuels was indicated by more than 70% of the households that used them as

safe.

The study further confirmed that LPG was the preferred fuel for cooking being

used by 81% of the households, followed by charcoal with 59%, electricity with

45% and kerosene with 41% in that order. For water heating, 71% of the

households used electricity, followed by charcoal with 36%, kerosene with 23%

and LPG with17% in that order. For lighting, 94% of the households used

electricity, followed by kerosene with 8%, LPG with 3% and charcoal with 2% in

that order. For space heating, 67% of the households used electricity, followed by

charcoal with 5%, LPG with 1%, and kerosene with 1%. For a particular use,

households were found to be using one or more than one fuel.

The results indicated that LPG and electricity were expensive but were clean and

efficient sources of energy. Even though LPG and electricity were found out to

be expensive, they were the preferred sources of energy. This was ascertained by

a high proportion of households indicating them as the most preferred fuels when

ignoring the cost of fuels. A high proportion of households used LPG for cooking

(81%) and electricity for lighting (94%). Cooking and lighting are the two

important uses of fuels for households in rural and urban areas.

It was established from the research findings that some of the social and

demographic factors that were in the models generated using the binary logistic

regression were not significant (p > 0.05) and thus do not influence the model

91
when the others are held constant. The study also established that even some of

the factors that are significant (p< 0.05) they do not concur with the theoretical

expectation. Further the study established that even in some cases where the

social and demographic factors concurred with the theoretical expectation, it was

not as strong as expected. Household income, occupation and education of both

the household head and spouse were strongly expected to strongly influence

households’ fuel choices, but this was not the case. This was an indication that

social and demographic factors are not very significant in influencing

households’ fuel choices.

The study proposed sugar cane and sweet sorghum as the preferred feedstock for

bioethanol production in Kenya. The two were selected due to climatic

suitability, crop cycles per year and also they are not expected to compete with

the food supply chain The study established that if these two energy crops are

grown and expanded in areas with suitable climate, 60-560 million litres of

bioethanol could be produced from sugar cane and over 30 billion litres of

bioethanol from sweet sorghum. The findings of the study were that Kenya has

suitable climate and land is available to grow sugarcane and sweet sorghum to

supply bioethanol to be used as household fuel.

The energy balances for the production of bioethanol from different countries

calculated from different approaches all have positive values and greater than

unity. Therefore more energy is produced than is consumed in production of

bioethanol from either sugar cane or sweet sorghum. Bioethanol production has

92
the potential to generate incomes and employment in the rural areas by boosting

the agriculture sector thus bringing social-economic development. The cost of

production of bioethanol from sugar cane and sweet sorghum feedstock can be

improved through the use and sale of crop residues and co-products obtained

during various production processes. Crop residues such as Bagasse can be used

in steam and power generation and others such as filter cake as animal feeds.

Others by-products such as molasses, Vinnase and Bagasse produce co-products

that are used in food and chemical industries.

Bioethanol is a green fuel and therefore a sustainable form of energy and has

many benefits over fossil fuels. The bioethanol is a modern, clean and efficient

fuel similar in characteristics to LPG. The bioethanol burns in specialized stoves

emitting heat similar in temperature to the heat output of LPG cookers. The

bioethanol also burns with smokeless flame like LPG. Thus bioethanol has

potential or possibility of substituting LPG as a household fuel. This is made

possible if bioethanol is available to households at affordable price.

The rising prices of fossil fuels are affecting the economies of most countries

mainly in the developing world. The pollution effects due to fossil fuel usage are

also a global concern. Climate change is now a concern to both developed and

developing countries. Fossil fuels sources are getting depleted and search for

alternative sources of energy has become a global issue. This study has

investigated household’s energy use of Kenyan urban society. The research will

go along in contributing to national strategies of developing alternative sources of

93
energy and meeting the increased energy demand for a middle income economy

envisaged in the Kenyan Vision 2030. The research is expected to generate

interest among scholars of energy technology and management and will lead to

further studies on developing alternative sources of energy. The future of the

country and of the world is in adopting renewable energy and thus this study has

undertaken a step to develop affordable energy solution for this emerging market.

5.3 Recommendations

Bioethanol feedstock production and hence its use as a household fuel requires

various interventions to overcome the barriers and negative implications in order

to stimulate progress while also addressing sustainability. The Kenya government

through the Ministries of Energy, Agriculture and Finance need to take a number

of interventions to back the bioethanol development. These interventions should

control the initial stages of bioethanol development but later the prices of

bioethanol should be determined by market prices. There is need of strong

government commitment and bioethanol production should be given a priority

development agenda so as to expand its production and use. The government

should establish institutional framework with targets and defined tasks. These

tasks will assist the creation of appropriate institutions.

The government should provide supportive policies such as tax incentives, low

interest borrowing options and investment on research and development. These

policies should focus on creating a predictable market through provision of

economic incentives for bioethanol industry by offering loans with low interest

94
rate and high productivity, hence making the bioethanol attractive by bringing its

cost down.

An environment impact assessment (EIA) should be done so as to identify the

social and environmental impacts associated with bioethanol development in the

new areas where the bioethanol crops will be grown. This is a requirement

whenever a major project is being undertaken and is done in consultation with

National Environment Management Authority (NEMA).

Future research should be carried out to determine the Kenyan bioethanol energy

balances, the net GHG emissions for sugarcane and sweet sorghum and on

bioethanol equipment and machinery. These will ensure the bioethanol

development is done in sustainable way.

95
REFERENCES

Abebaw, D (2007). Household Determinants of Fuel wood Choice in Urban

Ethiopia: A Case Study of Jimma Town. The Journal of Developing Areas, Vol. 41,
No. 1. pp 117- 126

Alam, M., Sathaye, J., Barnes, D. (1998). Urban Household energy use in India:
Efficiency and Policy Implications. Energy Policy. Vol. 26. Issue 11, pp 885-891

Almodares, A., Hadi, M. (2009). Production of Bioethanol from Sweet Sorghum: A

Review.

BNDES. CGEE (2008). Sugar-Based Bioethanol: Energy for Sustainable

Development. Rio de Janeiro: BNDES 2008.

Bernett, A (2000). Energy and the Fight against Poverty, Department for
International Development (DfID), Livelihood Sector Report No.6.United Kingdom.

Business Daily (2011). Unique Bio-fuel Warms its Way to Kitchen as Search for
Green Energy Hots Up. Available at http:/allafrica.com/stories/201108040243.html.
Accessed on 10th November 2011

Campbell, B; Vermeulen, S; Mangono, J; Mabugu, R. (2003). The Energy Transition

in Action: Urban Domestic Fuel Choice in a Changing Zimbabwe, Energy Policy,
31(6) pp 553-562

Davis, M. (1998). Rural Household Energy Consumption: The Effects of Access to

Electricity-Evidence from South Africa. Energy Policy 26(3), pp 207-217

Demirbas, A. (2008). The Importance of Bioethanol and Biodiesel Energy Sources.

Part B: Economics, Planning and Policy, 3:2, pp 177-185

Dufey, A. (2006). Bio-fuel Production, Trade and Sustainable Development:

Emerging Issues. International Institute for Environment and Development. London
(UK).

EPI. (2000). Ethanol’s Potential: Looking Beyond Corn. Available at D:\ethanol –

plan B. Html. Accessed on 11th November 2010

ESMAP. (2001). Sustainable Wood fuel Supplies from the Dry Tropical Woodlands.
ESMAP Technical Paper 13, Washington DC.

Farsi, M., Filippini, M., Pachauri, S. (2004). Fuel Choices in Urban Indian
Households: Environment and Development Economics 12 (6): pp 757-774

96
Gnansounou, E., Dauriat, A. (2005). Energy balance of Bioethanol: A Synthesis,
Lasen, Ecole Polytechnique, Lausanne, Switzerland.

Goldemberg (2007). Ethanol for Sustainable Energy Future. Science 315,808(2007).

DO1: 10. 1126/Science.1137013. www.sciencemag.org. Accessed on 5th April 2011.

Graeme, W., Nasidi, M., Akunna, J., Deeni, Y., Blackwood, D. (2010). Bioethanol in
Nigeria: Comparative Analysis of Sugarcane and Sweet Sorghum as Feedstock
sources. Energy and Environmental Science. DO1: 15.1039/c0ee00084a.
www.rsc.org/ees. Accessed on 20th May 2011.

Green Car Congress (2010). http:/www.greencarcongress.com/2010/03/anfavea-

20100306 html# more=. Accessed on 15th November 2011.

GTZ & GoK. (2008). A Roadmap for Bio-fuels in Kenya: Opportunities and
Obstacles.

Gujarati, D. (2003). .Basic Econometrics. 4th Edition. McGraw-Hill / Irwin

Heltburg, R. (2004). Fuel Switching: Evidence from Eight Developing Countries.

Energy Economics 26(5): pp869-887

Hosier, R.; Dowd, J. (1987). Household Fuel Choice in Zimbabwe-An Empirical Test
of Energy Ladder Hypothesis. Resources and Energy, 9, pp347-361

IEA. (2002). World Energy Outlook 2002.Energy and Poverty. Paris.

IEA. (2006). World Energy Outlook 2006. Making Energy Poverty History In Africa,
Paris, France

IIED. (2010). Bio-fuel Production, Trade and Sustainable Development. Dufey A,

Grieg-Gran M. (Eds). International Institute for Environment and Development.
London. (UK).

Janssen, R., Rutz, D., Helm, P., Woods, J., Chavez, D.(2009). Bio-energy for
Sustainable Development in Africa: Environment and social Aspects. Renewable
Energies. http://www.competebioafrica.net/publication/pub/otherpubl.html. Accessed
on 25th June 2011

Jiang, L., Pachauri, S. (2008). The Household Energy Transition in India and China.
International Institute for Applied systems Analysis (IIASA). Energy Policy.
Luxemburg, Austria.

97
Karekezi, S; Khamarunga, G; Waeni, K. (2002). Improving Energy Services for the
Poor in Africa – A Gender Perspective “ENERGIA” News 5:4 pp 16-19

Lisboa, C., Mauder, M., Kiese, R., Bahl, K. (2011). Bioethanol Production from
Sugarcane and Emissions of GHG-Known and Unknowns. GCB Bio-energy (2011)
doi: 10.1111/j.1757-1707.2011.01095.x

Macedo, I. Seabra, E. Silva, E. (2008). Green House Gases Emissions in the

Production and Use of Ethanol from Sugarcane in Brazil: 2005/2006 Averages and a
Prediction for 2020. Biomass and Bio-energy, 32, pp 582-595

Mcnube, D (2006). Household Survey on Energy Consumption Patterns in

Johannesburg Township: A Case Study of Diepkloof, Soweto

Meikle, S; Bannister, A. (2003). Energy Poverty and Sustainable Urban Livelihoods.

Department for International Development (DfID). United Kingdom

Mekonnen, A.,Kohlin, G. (2008). Determinants of Households Fuel Choice in Major

Cities in Ethiopia. Environment for Development DP pp8-18

Ministry of Energy (2004). Sessional Paper No.4 on Energy, Nairobi, Government of

Kenya.

Mugenda, A., Mugenda, O. (1999). Research methods: Quantitative and Qualitative

Approaches. Act Press, Nairobi, Kenya.

Njong, A., Tabi, J. (2011). An Analysis of Domestic Cooking choice in Cameroon.

European Journal of Social Sciences, Vol.20, No 2(2011).

Ouedraogo, B. (2005). Household Energy Preferences for Cooking in Urban

Ouagadougou, Burkina Faso. Energy Policy Vol.35, Issue 18.

Pundo, M., Fraser, G. (2006). Multinomial Logit Analysis of Household Fuel Choice
in Rural Kenya: The Case of Kisumu District. Agrekon, Vol. 45 No.1
RFA .(2010). Ethanol Industry Outlook. Climate of Opportunity.
Ranola, R., Demafelis, R., Rosario, E., Bataller, B. (2009). Enhancing the Viability of
Cassava Feedstock for Bioethanol in the Philippines. J.ISSAAS Vol.15.No 2, pp147-
158

REN21 (2008) “Renewables 2007 Global Status Report” (Paris: REN21 Secretariat
and Washington DC: World Watch Institute)

Robinson, J. (2006). Bioethanol as a Household Cooking Fuel: A Mini Pilot study of

the SuperBlu Stove in Peri-Urban Malawi.

98
Roman, M., Adriana, G., Chiatoanu, M., Senila, L., Luca, E., Naghiu, A., Irimie, D.,
Roman, C. (2010). Sweet Sorghum: A Good Solution for Bioethanol. Agriculture
Practice and Science Journal, Vol.73, No 1-2(2010).

Smith, R.., Michael, A., Ma Yuqing, Wathana, W., Ashwin, K. (1994): ‘Air Pollution
and the Energy Ladder in Asian Cities’, Energy,19 (5), pp 587-600.

Tekle, G. (2008) Local Production and Use of Bioethanol for Transport in Ethiopia-
Status, challenges and Lessons. MSc Thesis, Lund University, Sweden,
The American Coalition for Ethanol. (2008). Life Cycle Analysis of Greenhouse Gas
Emissions Associated with Starch-Based Ethanol.

UNDP. (2007). Business Plan for Ethanol Cooking Fuel and Domestic Clean Cook
Stove Market Development in Addis Ababa, Ethiopia.

UNEP (2009). Towards Sustainable Production and Use of Resources: Assessing

Bio-fuels.

UNF, (2008). Sustainable Bio-energy Development in UEMOA Member Countries.

WB, (2003). Household Fuel and Energy Use in Developing Countries: A Multi-
country Study. ESMAP Technical Paper, No 42.Washington DC: World Bank.
Available at http:/siteresource.worldbank.org/INTPSIA/Resources/490023. Accessed
on 10th September 2011.

WHO. (2002). Indoor Air Pollution, Health and the Burden of Disease.
www.who.int/indoorair/info/briefing 2.pdf. Accessed on 10th July 2011.

WHO. (2006). Fuel for Life: Household Energy and Health.

WI, CAP. (2006). American Energy: The Renewable Path to Energy Security.
www.americanenergynow.org. Accessed on 10th July 2011

Wanambwa, L. (2005). Ethanol Fuel Production and Use in Kenya for Sustainable
Development. .MSc.Thesis, Lund University, Sweden.

Watanabe, M. (2009). Ethanol Production in Brazil: Bridging its Economic and

Environmental Aspects. Energy Forum Articles. International Association For Energy
Economics, pp45-48.

Wood, T., Baldwin, S.(1985). Fuel wood and Charcoal Use in Developing Countries.
Annual Review of Energy, Palo Alto, 10: 407 – 429

99
Zuzarte, F. (2007). Ethanol for Cooking. Feasibility of Small-scale Ethanol Supply
and its Demand as a cooking fuel: Tanzania Case Study. MSc Thesis, KTH School of
energy and Environment Technology Stockholm, Sweden.

100
 LIST OF APPENDICES

Appendix I: Questionnaire for households

Table1:
Interview date
Enumerator identification
Region/Zone
House number
Number of persons in the house

Table 2: Does the household have the following? Use 1 for Yes and 2 for No
TV set
Computer
Radio
Mobile
Fireless jiko

Table 3: Household head and spouse details

Use the corresponding number for the alternative e.g. 1 for male and 2 for female
Household Head Spouse

Date of birth
Gender
1. Male
2. Female
Religion
1. Christian
2. Muslim
3. Others
Marital details
1. Married living with
spouse
2. Married living without
spouse
Education
1. Primary
2. Secondary
3. Post-Secondary

101
Occupation
1. Government employed
2. Private/NGOs
3. Self employed
4. Retired

Table 4: Housing.
Use the corresponding number for the alternative e.g. 1 for modern and 2 for
traditional
House type
1. Modern
2. Traditional
House ownership
1. Rental
2. Owner Occupier
Kitchen
1. Internal
2. External

Table 5: Rank the fuels in order of preference starting with 1 with the most preferred
and no ties
Ignoring costs In order of costs In order of risks
Firewood
Charcoal
Kerosene
LPG (Gas)
Biogas
Electricity

Table 6: Characteristics associated with household fuels. Use 1 for Yes and 2 for No
Firewood Charcoal Kerosene LPG Electricity
Fuel is
inexpensive
Good fuel
availability
Clean to work
with
No smoke
Fast
Safe

102
Table 7: Fill as appropriate with respect to fuel consumption. For fuel uses 1 for yes
and 2 for no.
Firewood Charcoal Kerosene LPG Electricity
Cooking
Water heating (bathing)
Lighting
Space heating
Monthly fuel
consumption (Kshs.)

103
Appendix II: Categorical variable coding
Parameter coding
Frequency (1) (2) (3)
Q3.6b Spouse Occupation 1. Government employed 52 1.000 .000 .000
2.Private/NGO's 75 .000 1.000 .000
3. Self employed 144 .000 .000 1.000
Retired 2 .000 .000 .000
Q3.6a Household Occupation 1. Government employed 43 1.000 .000 .000
2.Private/NGO's 113 .000 1.000 .000
3.Self employed 103 .000 .000 1.000
Retired 14 .000 .000 .000
Household head Education Primary 8 1.000 .000
Secondary 62 .000 1.000
Post-secondary 203 .000 .000
Spouse Education Primary 28 1.000 .000
Secondary 89 .000 1.000
Post-secondary 156 .000 .000
Spouse Gender Male 7 1.000
Female 266 .000
Household head Religion Christian 262 1.000
Muslim 11 .000
spouse Religion Christian 261 1.000
Muslim 12 .000
Household head Marital details Married living with spouse 265 1.000
Married living without spouse 8 .000
Spouse Marital details Married living with spouse 265 1.000
Married living without spouse 8 .000
House ownership Rental 160 1.000
Owner occupier 113 .000
House type Modern 253 1.000
Traditional 20 .000
Kitchen Internal 267 1.000
External 6 .000
Household head Gender Male 269 1.000
Female 4 .000

104
Appendix III: Variable coding

Q3.2a Household gender (1) Male

Q3.2b Spouse gender
(2) Female

Q3.3a Household head religion (1) Christian

Q3.3b Spouse religion
(2) Muslim

Q3.4a Household head marital status (1) Living with spouse

Q3.4b Spouse marital status
(2) Living without spouse

Q3.5a Household head education (1) Primary

Q3.5b Spouse education
(2) Secondary

(3) Post- secondary

Q3.6a Household head occupation (1) Government employed

Q3.6b Spouse occupation
(2) Private/NGO’s

(3) Self employed

(4) Retired

Q4.1 House type (1) Modern

(2) Traditional

Q4.2 House ownership (1) Rental

(1) Owner occupier

Q4.3 Kitchen type (1) Internal

(2) External

Q8 Age of household head in years

Q9 Age of spouse in years

Q12 Household income in thousands

105
Appendix IV: Cooking with charcoal
Block 0: Beginning Block

Classification Tablea,b
Predicted
Cooking (charcoal)
Observed No Yes Percentage Correct
Step 0 Cooking (charcoal) No 0 103 .0
Yes 0 170 100.0
Overall Percentage 62.3
a. Constant is included in the model.
b. The cut value is .500

Block 1: Method = Forward Stepwise (Wald)

Omnibus Tests of Model Coefficients

Chi-square df Sig.
Step 4 Step 4.277 1 .039
Block 28.034 6 .000
Model 28.034 6 .000

Model Summary
Cox & Snell R
Step -2 Log likelihood Square Nagelkerke R Square
4 333.812c .098 .133

Classification Tablea
Predicted
Cooking (charcoal)
Observed No Yes Percentage Correct
Step 4 Cooking (charcoal) No 34 69 33.0
Yes 27 143 84.1
Overall Percentage 64.8
a. The cut value is .500

Variables in the Equation

B S.E. Wald Df Sig. Exp(B)
Step 4d Q12 -.029 .010 7.881 1 .005 .972
Q1b .380 .107 12.722 1 .000 1.463
Q3.2b(1) -1.694 .872 3.775 1 .052 .184
Q3.6b 8.832 3 .032
Q3.6b(1) -20.144 2.636E4 .000 1 .999 .000
Q3.6b(2) -21.309 2.636E4 .000 1 .999 .000
Q3.6b(3) -20.691 2.636E4 .000 1 .999 .000
Constant 20.382 2.636E4 .000 1 .999 7.108E8

106
Appendix V: Cooking with kerosene
Block 0: Beginning Block
Classification Tablea,b
Predicted
Cooking (kerosene)
Observed No Yes Percentage Correct
Step 0 Cooking (kerosene) No 164 0 100.0
Yes 109 0 .0
Overall Percentage 60.1
a. Constant is included in the model.
b. The cut value is .500

Block 1: Method = Forward Stepwise (Wald)

Omnibus Tests of Model Coefficients
Chi-square df Sig.
Step 4 Step 13.306 3 .004
Block 136.132 7 .000
Model 136.132 7 .000

Model Summary
Cox & Snell R
Step -2 Log likelihood Square Nagelkerke R Square
4 231.170a .393 .531

Classification Tablea
Predicted
Cooking (kerosene)
Observed No Yes Percentage Correct
Step 4 Cooking (kerosene) No 144 20 87.8
Yes 33 76 69.7
Overall Percentage 80.6
a. The cut value is .500

Variables in the Equation

B S.E. Wald Df Sig. Exp(B)
Step 4d Q3.5b 20.259 2 .000
Q3.5b(1) .918 .594 2.389 1 .122 2.504
Q3.5b(2) 1.642 .368 19.919 1 .000 5.164
Q3.6a 12.480 3 .006
Q3.6a(1) -.305 .938 .106 1 .745 .737
Q3.6a(2) -1.614 .931 3.006 1 .083 .199
Q3.6a(3) -.395 .867 .207 1 .649 .674
Q9 -.081 .019 17.433 1 .000 .922
Q12 -.064 .013 23.078 1 .000 .938
Constant 4.638 1.345 11.898 1 .001 103.339

107
Appendix VI: Cooking with LPG
Block 0: Beginning Block

Classification Tablea,b
Predicted
Cooking (LPG)
Observed No Yes Percentage Correct
Step 0 Cooking (LPG) No 0 49 .0
Yes 0 224 100.0
Overall Percentage 82.1
a. Constant is included in the model.
b. The cut value is .500

Block 1: Method = Forward Stepwise (Wald)

Omnibus Tests of Model Coefficients

Chi-square df Sig.
Step 4 Step 15.833 3 .001
Block 126.125 6 .000
Model 126.125 6 .000

Model Summary
Cox & Snell R
Step -2 Log likelihood Square Nagelkerke R Square
4 130.831c .370 .607

Classification Tablea
Predicted
Cooking (LPG)
Observed No Yes Percentage Correct
Step 4 Cooking (LPG) No 33 16 67.3
Yes 10 214 95.5
Overall Percentage 90.5
a. The cut value is .500

Variables in the Equation

B S.E. Wald Df Sig. Exp(B)
Step 4d Q3.6a 14.004 3 .003
Q3.6a(1) -16.204 9.986E3 .000 1 .999 .000
Q3.6a(2) -15.872 9.986E3 .000 1 .999 .000
Q3.6a(3) -17.664 9.986E3 .000 1 .999 .000
Q4.1(1) 3.087 .957 10.404 1 .001 21.911
Q4.2(1) -2.260 .792 8.133 1 .004 .104
Q12 .104 .024 18.199 1 .000 1.110
Constant 15.016 9.986E3 .000 1 .999 3.323E6

108
Appendix VII: Cooking with electricity
Block 0: Beginning Block

Classification Tablea,b
Predicted
Cooking (electricity)
Observed No Yes Percentage Correct
Step 0 Cooking (electricity) No 139 0 100.0
Yes 134 0 .0
Overall Percentage 50.9
a. Constant is included in the model.
b. The cut value is .500

Block 1: Method = Forward Stepwise (Wald)

Omnibus Tests of Model Coefficients

Chi-square df Sig.
Step 4 Step 4.243 1 .039
Block 134.308 6 .000
Model 134.308 6 .000

Model Summary
Cox & Snell R
Step -2 Log likelihood Square Nagelkerke R Square
a
4 244.059 .389 .518

Classification Tablea
Predicted
Cooking (electricity)
Observed No Yes Percentage Correct
Step 4 Cooking (electricity) No 109 30 78.4
Yes 32 102 76.1
Overall Percentage 77.3
a. The cut value is .500

Variables in the Equation

B S.E. Wald Df Sig. Exp(B)
Step 4d Q3.2a(1) 2.713 1.334 4.140 1 .042 15.078
Q3.6b 17.838 3 .000
Q3.6b(1) 3.605 1.523 5.604 1 .018 36.780
Q3.6b(2) 2.665 1.496 3.176 1 .075 14.371
Q3.6b(3) 1.918 1.486 1.665 1 .197 6.808
Q9 .033 .016 4.179 1 .041 1.034
Q12 .095 .014 46.354 1 .000 1.100
Constant -9.247 2.232 17.162 1 .000 .000

109
Appendix VIII: Water heating with electricity
Block 0: Beginning Block

Classification Tablea,b
Predicted
Water heating (bathing) (electricity)

Percentage
Observed No Yes Correct
Step 0 Water heating (bathing) (electricity) No 0 72 .0
Yes 0 201 100.0
Overall Percentage 73.6
a. Constant is included in the model.
b. The cut value is .500

Block 1: Method = Forward Stepwise (Wald)

Omnibus Tests of Model Coefficients

Chi-square df Sig.
Step 2 Step 13.532 2 .001
Block 117.302 3 .000
Model 117.302 3 .000

Model Summary
Cox & Snell R
Step -2 Log likelihood Square Nagelkerke R Square
2 197.701a .349 .510

Classification Table
Predicted
Water heating (bathing) (electricity)
Observed No Yes Percentage Correct
Step 2 Water heating (bathing) (electricity) No 49 23 68.1
Yes 16 185 92.0
Overall Percentage 85.7
a. The cut value is .500

Variables in the Equation

B S.E. Wald Df Sig. Exp(B)
Step 2b Q3.5b 12.964 2 .002
Q3.5b(1) -1.728 .550 9.864 1 .002 .178
Q3.5b(2) -1.032 .384 7.230 1 .007 .356
Q12 .127 .018 47.772 1 .000 1.136
Constant -1.394 .453 9.475 1 .002 .248

110
Appendix IX: Water heating with charcoal
Block 0: Beginning Block

Classification Tablea,b
Predicted
Water heating (bathing) (charcoal)
Observed No Yes Percentage Correct
Step 0 Water heating (bathing) (charcoal) No 168 0 100.0
Yes 105 0 .0
Overall Percentage 61.5
a. Constant is included in the model.
b. The cut value is .500

Block 1: Method = Forward Stepwise (Wald)

Omnibus Tests of Model Coefficients

Chi-square df Sig.
Step 1 Step 13.138 1 .000
Block 13.138 1 .000
Model 13.138 1 .000

Model Summary
Cox & Snell R Nagelkerke R
Step -2 Log likelihood Square Square
1 350.650a .047 .064

Classification Tablea
Predicted
Water heating (bathing) (charcoal)
Observed No Yes Percentage Correct
Step 1 Water heating (bathing) No 155 13 92.3
(charcoal)
Yes 84 21 20.0
Overall Percentage 64.5
a. The cut value is .500

Variables in the Equation

B S.E. Wald Df Sig. Exp(B)
Step 1a Q1b .298 .085 12.274 1 .000 1.348
Constant -1.927 .440 19.182 1 .000 .146

111
Appendix X: Space heating with electricity
Block 0: Beginning Block

Classification Tablea,b
Predicted
Space heating (electricity)
Observed No Yes Percentage Correct
Step 0 Space heating (electricity) No 0 88 .0
Yes 0 185 100.0
Overall Percentage 67.8
a. Constant is included in the model.
b. The cut value is .500

Block 1: Method = Forward Stepwise (Wald)

Omnibus Tests of Model Coefficients

Chi-square df Sig.
Step 5 Step 4.265 1 .039
Block 65.950 6 .000
Model 65.950 6 .000

Model Summary
Cox & Snell R
Step -2 Log likelihood Square Nagelkerke R Square
5 277.278b .215 .300

Classification Tablea
Predicted
Space heating (electricity)
Observed No Yes Percentage Correct
Step 5 Space heating (electricity) No 37 51 42.0
Yes 12 173 93.5
Overall Percentage 76.9
a. The cut value is .500

Variables in the Equation

B S.E. Wald Df Sig. Exp(B)
Step 5e Q3.3b(1) 1.694 .700 5.858 1 .016 5.443
Q3.5b 23.671 2 .000
Q3.5b(1) 1.579 .624 6.399 1 .011 4.852
Q3.5b(2) 1.823 .393 21.544 1 .000 6.190
Q4.1(1) 3.211 .726 19.573 1 .000 24.806
Q4.2(1) -.652 .319 4.170 1 .041 .521
Q12 .022 .011 3.949 1 .047 1.022
Constant -4.709 1.061 19.705 1 .000 .009

112
Appendix XI: Lighting with electricity
Block 0: Beginning Block
Classification Tablea,b
Predicted
Lighting (electricity)
Observed No Yes Percentage Correct
Step 0 Lighting (electricity) No 0 14 .0
Yes 0 259 100.0
Overall Percentage 94.9
a. Constant is included in the model.
b. The cut value is .500

Block 1: Method = Forward Stepwise (Wald)

Omnibus Tests of Model Coefficients

Chi-square df Sig.
Step 7 Step 8.780 1 .003
Block 65.541 5 .000
Model 65.541 5 .000

Model Summary
Cox & Snell R
Step -2 Log likelihood Square Nagelkerke R Square
7 44.900c .213 .641

Classification Tablea
Predicted
Lighting (electricity)
Observed No Yes Percentage Correct
Step 7 Lighting (electricity) No 8 6 57.1
Yes 2 257 99.2
Overall Percentage 97.1
a. The cut value is .500

Variables in the Equation

B S.E. Wald Df Sig. Exp(B)
Step 7f Q1b -.934 .326 8.224 1 .004 .393
Q4.1(1) 2.897 1.060 7.471 1 .006 18.120
Q4.2(1) -4.578 1.523 9.039 1 .003 .010
Q9 -.151 .057 7.078 1 .008 .860
Q12 .168 .060 7.910 1 .005 1.182
Constant 11.627 3.383 11.816 1 .001 1.121E5

113
Appendix XII: Letter of research authorization from NCST

114
Appendix XIII: Letter from Provincial Administration (Nairobi Area)

115