PRIVATEdeecomm.blogspot.com
A Feasibility Study for
Improving Uganda’s Water
to Drinkable Standards:
Lessons from Kampala
[Type the document subtitle]
FRANCIS WASSWA                                                                                  
204029899
In fulfillment of the Requirement for the degree of
MAGISTER COMMERCII
In the department of Economics & Economics History
Nelson Mandela Metropolitan University
December 2007
Supervisor: Prof. S.G. Hoskingii
List of Acronyms
ADB  African Development Bank
ADF  African Development Fund
BGS  British Geological Survey
BOU  Bank of Uganda
CBA Cost‐Benefit Analysis
CRC  Cooperative Research Center
CVM  Contingent Valuation Methodology
DALY Disability Adjusted Life Year
DWD  Directorate of Water Development
FAO Food and Agriculture Organisation
ICWE International Conference on Water and the Environment
ITCZ Inter‐Tropical Convergence Zone
JWSSD
Johannesburg World Summit on Sustainable Development
LICs
Lower Income Countries
LL
Log‐Likelihood
MDG Millennium Development Goals
MLE Maximum Likelihood Estimation
MOH  Ministry of Health (Uganda)
MWLE  Ministry of Water, Lands and Environment
NEMA  National Environmental Management Authority (Uganda)
NGOs Non‐Governmental Organisations
NOAA North Oceanic and Atmospheric Administration
NWP National Water Policy (Uganda)
NWSC  National Water and Sewerage Corporation
OM  Operation and Maintenance
PEAP Poverty Eradication Action Plan
PUV  Passive Use Values  iii
QALY  Quality Adjusted Life Year
RWSD  Rural Water Supply Department (Uganda)
UDHHS  US Department of Health and Human Services
UN United Nations
UNBS Uganda National Bureau of Standards
UNCED  United Nations Conference on Environment and Development
UNDP  United Nations Development Programme
UNESCO United Nations Educational, Scientific and Cultural Organisation
UNPAC United Nations Platform for Action Committee
UV Use Values
WAP  Water Action Plan
WHO  World Health Organisation
WMO  World Metrological Organisation
WTP Willingness to Pay
WWAP  World Water Assessment Programme
WWC World Water Commissioniv
Acknowledgments
Cicero once wrote that “gratitude is not only the greatest of virtues, but the parent of all the
others.” Tempted by his sentiment, I am compelled to dedicate this section to all of the people
that have had a hand in my work till its fruition. I must admit I am a little embarrassed by the
litany of names I have to make mention of; then again, it would be a grave injustice if I did not
acknowledge other people’s contribution.
My first token of thanks goes to the academicians that had a direct input in my work. The
fondest of these thanks go to my supervisor Prof. S. G. Hosking HOD Economics NMMU; you
always had that teasing question that prompted me to search for my full potential. You have
afforded me unexampled experience in writing and critical analyses and for this am grateful.
Next, I wish to thank June Elijah Simakani and Gary Sharp of NMMU for their guidance in the
data analysis exercise. Also, I wish to thank Tim Haab of Ohio State University, whose work was
monumentally inspirational and was willing to decipher any queries I had. To Prof. CVR Wait of
NMMU, am thankful for your guidance with regard to the structuring and presentation of my
work. Dr. Ped Winter and Mr. Christopher Parsons of NMMU, you too your help was prodigious.
Then, to Mrs. Noluntu Dyubhele of NMMU, thank you for your input.
The financial support (R20,000) for this research from the NMMU trust fund is gratefully
acknowledged. Being a student in South Africa from a non‐SADC country severely limits one’s
funding options.
To understand the importance of the other category of contributors, permit me to draw from
the muses of another philosopher called Albert Schweitzer. With admirable honesty, he writes
that “At times our own light goes out and is rekindled by a spark from another person. Each of
us has a cause to think with deep gratitude of those who have lighted the flame within us”. My
flame was lit by my parents and family both at home and in South Africa, The Hutton and the
Kakembo families, my colleagues: Moses Mlangeni, Johane Dikgang, Solomon Mdege, Michealv
Sale, Radu Mihailescu, Requier Wait, William Akoto, Balbo Gatogang, Dennis Nyabonga, Dedre
Erasmus, Nasreen Adams, Ronnie Ncwadi, Guinevier Pereira, and Henry van der Westhuizen,
my friends: Sven Jensen, Emelie Höglund, Sofia Wahlstedt, Martin Lutaaya and Baijuka Rufus.
To all of you I am grateful.  vi
Abstract
An enthusiastic global campaign on intervention in water in the Lower Income Countries (LICs)
was launched by the UN at the International Conference on Water and the Environment
(ICEW), in Rio de Janerio, in January of 1992. In June of the same year, in Dublin, a plan of
action was devised and a commitment to the water related goals highlighted in Rio de Janerio
was made. Close to fifteen years on, there is little to show by way of success in the intended
countries.
Over 1.1 billion people in the LICs lack safe water. The direct impact of this is a higher risk of
waterborne diseases. The waterborne diseases claim 42,000 lives every week in the LICs. By any
standards this is a serious depletion of the human capital stock. Looked at in light of the fact
that these countries still heavily rely on labour in production, amplifies the need to preserve
health. The inherent danger posed by the poor quality water‐ as can be drawn from the above
statistics‐  seems to suggest that improving the quality of water would go a long way in
improving and preserving societal health in the LICs. By implication this would improve the
productivity of the workers. Other benefits include cost mitigation, improved investor
confidence as well as increased tourists’ confidence‐  all of which are vital for LICs’ growth
prospects. It begs the question of why these countries have not improved their water quality.
With specific reference made to Uganda, this research is bent on answering this question.
In Uganda, there is consensus among scientists that the ground and open water sources are
degraded to dangerous levels.   Water quality parameters like turbidity, coliform count, and
colour are all above the WHO minimum specifications for potable water and are on the rise in
the country. This is indicative of water quality deterioration and it heightens the risk of
waterborne diseases to the users. The waterborne burden of disease in Uganda is on the rise
with a high fatality rate of 440 lives every week. The need to improve water quality in the
country has been acknowledged. However, attempts to address the problem have only been
undertaken on a small scale, most notable of these being the PuR home water treatmentvii
program. There is evidence in the country that the water quality would have apparent benefits.
Strong correlations have been found between improved health in HIV patients and improved
water quality in the country. In the economics of health, improving societal health inherently
improves workers’ performance and productivity, leading to higher growth of the economy.
There is an economic imperative therefore, as to why countries like Uganda should improve
their water quality. In spite of this, even the country’s most urbanized setting‐ Kampala‐ lacks
potable water. This study therefore investigates why, in a time when not only the global agenda
is more supportive than ever and when the country’s water resources have been found to be
risky to use, Uganda has not improved water quality. Kampala is used as the model district for
this study. The district accounts for three quarters of users of treated water in the country.
The problem is investigated by assessing the efficiency case of such a project (a water quality
improving project) in the country; the methodology employed to this end is the Cost Benefit
Analysis (CBA). The methodology compares the costs and benefits of a project, in monetary
terms, in the same analysis, over its useful life. In the application of CBA one allows for the time
value of money by using the discount rate to make the costs and benefits of the project
occurring in different years comparable. In principle, the methodology is simple to apply‐ only
that issues arise in the quantification of benefits and the determination of the discount rate.
Benefits of the Kampala water quality, improving project include non‐market values and for this
reason a non‐market valuation technique, the Contingent Valuation Method (CVM), was
employed in their quantification. The CVM technique estimates the benefits by measuring the
individuals’ willingness to pay for the improved scenario‐ in this case the scenario was one with
a water quality‐improving project.
The application of the CVM across many disciplines has invited a lot of criticism over the
reliability of its estimates as a measure of value. A panel assembled by the North Oceanic and
Atmospheric Administration (NOAA) to investigate the reliability of the CVM resolved that as
long as the CVM was well conducted, the generated results would reliably predict non‐market
values. The Kampala CVM, for the benefits’ quantification, was conducted with the NOAA
guidelines in mind. The final value of the project’s benefits was the WTP predicted for theviii
median respondent namely Ushs 385.07/= per cubic meter of water. The discount rate was
deemed to be the social opportunity cost of capital in the country, viz 12 percent, this being
that rate of return foregone by investing in another sector.
The project’s costs were arrived at through liaison with water engineers and consulting past
data from Uganda’s Water suppliers. From this, the project’s fixed costs were predicted to be
Ushs 1451/= per cubic meter of water and the operation and maintenance costs predicted to
be Ushs 591.7/= per cubic meter of water. The project’s useful life was deemed to be the
average life of a Ugandan, namely 52 years; this choice reflecting the belief that the benefits
would last over the users’ whole life.
The results of the Kampala water quality‐improving project indicate that the project would not
be feasible. It did not matter what discount rate one employed, the project’s operating and
maintenance (OM) costs exceed the benefits. The results offer an indication as to why water
quality has not been improved in Uganda‐ because the paying population is unwilling to pay for
the entire cost of the project. This deduction is not to suggest that the users do not recognize
the benefits of the project. The unpleasant truth is that the users’ incomes are typically
stretched so thin by other demands that a decision to make more deductions from these
incomes is not an inviting one.  
However, there is a need to improve water quality in LICs like Uganda, as can be deduced from
the analysis of the risks of not doing so and benefits of doing so. Accordingly, such projects
have to be funded by mechanism that does not require the users to cover the whole cost, but
only part of such a cost, with the remainder from other sources like NGOs and foreign aid.ix
List of Tables
TABLE 1.1: THE UGANDA NATIONAL AND REGIONAL ACCESS TO SAFE WATER AS A PERCENTAGE OF THE TOTAL POPULATION ....................5
TABLE 1.2: KAMPALA DISTRICT HOUSEHOLD WATER SOURCE.....................................................................................................6
TABLE 2.1: THE GRAND EXPERIMENT RESULTS .......................................................................................................................29
TABLE 3.1: RAW WATER QUALITY IN TERMS OF TURBIDITY AND COLOUR ....................................................................................53
TABLE 6.1: DESCRIPTIVE STATISTICS FROM THE KAMPALA SURVEY ...........................................................................................107
TABLE 6.2: EDUCATION LEVEL OF THOSE SURVEYED...............................................................................................................108
TABLE 6.3: DICHOTOMOUS VALUATION RESPONSE SUMMARY................................................................................................111
TABLE 6.4: THE VARIABLE DESCRIPTION AND ANTICIPATED COEFFICIENT SIGNS............................................................................114
TABLE 6.5: PARAMETER ESTIMATES ...................................................................................................................................116
TABLE 6.6: THE CASE PROCESS SUMMARY OUTPUT FOR THE KAMPALA CVM MODEL ...................................................................117
TABLE 6.7: THE CLASSIFICATION TABLE ..............................................................................................................................118
TABLE 6.8: THE TURNBULL ESTIMATION OF WTP.................................................................................................................121
TABLE 7.1: OPERATING COSTS FOR KAMPALA FOR THE YEARS 2004 AND 2005 .........................................................................125
TABLE 7.2: KAMPALA CBA COMPONENTS ..........................................................................................................................127
TABLE 7.3: SENSITIVITY ANALYSIS DECISION CRITERIA.............................................................................................................129
TABLE 7.4: SECONDARY SENSITIVITY ANALYSIS DECISION CRITERIA ............................................................................................130x
List of Figures
FIGURE 2.1: THE HEALTH PRODUCTION FUNCTION.................................................................................................................20
FIGURE 2.2: THE WORK LEISURE TIME TRADE‐OFF ..................................................................................................................22
FIGURE 2.3: ALLOCATION OF RESOURCES TO PRODUCE HEALTH AND COMPOSITE GOOD B...............................................................23
FIGURE 2.4: THE MARGINAL EFFICIENCY OF HEALTH CAPITAL....................................................................................................24
FIGURE 2.5: THE COMBINED ANALYSIS OF THE HEALTH DEMAND MODEL .....................................................................................25
FIGURE 2.6: FRAMEWORK FOR THE WATER POVERTY AND HEALTH LINKAGES ................................................................................31
FIGURE 2.7: THE GLOBAL WATER RESOURCES........................................................................................................................34
FIGURE 2.8: ESTIMATED ANNUAL WORLD WATER USE..............................................................................................................35
FIGURE 3.1: MAP OF UGANDA SHOWING THE MAJOR DRAINAGE WATER SUB‐BASINS ....................................................................47
FIGURE 3.2: UGANDA WATER WITHDRAWALS (CONSUMPTIVE USES) .........................................................................................49
FIGURE 3.3: AERIAL SHOT OF THE LAKE VICTORIA ...................................................................................................................50
FIGURE 3.4:  WATER TREATMENT COST (IN SHILLINGS) PER CUBIC METER AT THE GABA II ...............................................................52
FIGURE 3.5:  TYPICAL KAMPALA DRAINAGE CHANNEL ..............................................................................................................55
FIGURE 3.6: STRUCTURE OF MINISTRY OF WATER LANDS AND ENVIRONMENT (MWLE) ...............................................................61
FIGURE 3.7: THE CLARIFIER AND INLET CHANNEL (GABA II).......................................................................................................65
FIGURE 3.8: THE FILTER BED (GABA II).................................................................................................................................66
FIGURE 3.9: THE CHLORINATION ROOM................................................................................................................................68
FIGURE 3.10: CHLORINE STORAGE ROOM AND TREATED WATER TANK ........................................................................................68
FIGURE 3.11: THE PUMP HOUSE AND SURGE VESSEL ..............................................................................................................69
FIGURE 6.1: THE ELICITATION ...........................................................................................................................................105
FIGURE 6.2.: GENDER PARTICIPATION IN SURVEY..................................................................................................................107
FIGURE 6.3: RESPONDENTS’ WATER SOURCE ......................................................................................................................109
FIGURE 6.4: QUALITY PERCEPTIONS...................................................................................................................................110
FIGURE 6.5: KAMPALA WATERBORNE DISEASE BURDEN ........................................................................................................111xi
Table of Contents
Contents                        Page
List of Acronyms .................................................................................................................................... ii
Acknowledgments................................................................................................................................ iv
Abstract ................................................................................................................................................... vi
List of Tables.......................................................................................................................................... ix
List of Figures ..........................................................................................................................................x
Table of Contents.................................................................................................................................. xi
CHAPTER ONE: INTRODUCTION AND BACKGROUND OF THE PROBLEM                        
15
1.0. Introduction.......................................................................................................................... 2
1.2. Background and statement of the problem......................................................................... 5
1.3. Objective: ........................................................................................................................... 10
1.4. Sub‐Objectives:................................................................................................................... 11
1.5. Reasons for the Research:.................................................................................................. 11
1.6. Scope of study:................................................................................................................... 12
1.7. Methodology:..................................................................................................................... 12
1.8. Organisation of the study:.................................................................................................. 14
CHAPTER TWO: LITERATURE REVIEW                                                                                           15
2.0. Introduction........................................................................................................................ 16
2.1. The Economics of Health:................................................................................................... 18
2.1.1: Demand for Health: The Grossman Model .............................................................................................. 19
2.1.2. Critique of the Grossman Model: ............................................................................................................. 26
2.1.3. Applying the Grossman Model to the Ugandan context: ......................................................................... 26
2.2. Beyond the Grossman model:............................................................................................ 27xii
2.3. The water health linkages: the John Snow Model............................................................. 29
2.3.1. Applying the Snow model to the Ugandan scene .................................................................................... 30
2.4. Poverty Water and Health Linkages:.................................................................................. 31
2.5. Changing conditions: The scarcer resource. ...................................................................... 33
2.6. The Choice to Privatise water utilities or not to: ............................................................... 38
2.6.1 The Bottled water Industry experience..................................................................................................... 40
2.7. Conclusion: ......................................................................................................................... 42
CHAPTER THREE: UGANDA WATER                                                                                           45
3.0. Introduction........................................................................................................................ 46
3.1. The Uganda freshwater resources..................................................................................... 46
3.1.1. Open water sources ................................................................................................................................. 46
3.1.2. Ground water sources .............................................................................................................................. 54
3.2. The Uganda Water management structure....................................................................... 56
3.2.1. The Uganda water sector legal Framework ............................................................................................. 56
3.2.2. Uganda Water Legislation on Economic Appraisal .................................................................................. 59
3.2.3. The Uganda Water Institutional Framework: .......................................................................................... 62
3.2.4. Kampala Water Supply ............................................................................................................................. 62
3.3. Kampala Water treatment: The Gaba II plant.................................................................... 63
3.4. Conclusion .......................................................................................................................... 69
CHAPTER FOUR: COST­BENEFIT ANALYSIS THEORY                                                                71
4.0 Introduction......................................................................................................................... 72
4.1. Application of the CBA ....................................................................................................... 73
4.1.1. Defining the Scope of the Project ............................................................................................................ 73
4.1.2. Quantifying the project benefits and costs: ............................................................................................. 74
4.1.3. Discounting the Benefits and Costs .......................................................................................................... 75
4.1.4. Decision criteria: ...................................................................................................................................... 77
4.1.5. Sensitivity analysis:................................................................................................................................... 78
4.2. The case for CBA................................................................................................................. 79xiii
4.3. Conclusion .......................................................................................................................... 79
CHAPTER FIVE: CONTINGENT VALUATION METHOD THEORY                                                  81
5.0. Introduction: ...................................................................................................................... 82
5.1. The CVM Model: Rationale and Critique ........................................................................... 83
5.2. NOAA CVM design Guidelines............................................................................................ 85
5.2.1. Elicitation with the Referendum format .................................................................................................. 85
5.2.2. Addressing the Embedding problem ........................................................................................................ 86
5.2.3. Time considerations and CVM surveys .................................................................................................... 87
5.2.4. Sample size and type ................................................................................................................................ 87
5.2.5. Comprehensive survey instrument .......................................................................................................... 88
5.2.6. Reporting .................................................................................................................................................. 88
5.3. Application of the CVM model........................................................................................... 89
5.3.1. Defining the valuation problem ............................................................................................................... 89
5.3.2. Preliminary decisions about the survey ................................................................................................... 89
5.3.3. Survey Design: .......................................................................................................................................... 91
5.3.4. Administering the survey ......................................................................................................................... 93
5.3.5. Processing of the data .............................................................................................................................. 93
5.3.6. Test for Reliability and Validity of Results ................................................................................................ 97
5.4. Conclusion .......................................................................................................................... 98
CHAPTER SIX: KAMPALA CVM SURVEY                                                                                        100
6.1. Introduction: .................................................................................................................... 101
6.2. Conducting the Kampala CVM ......................................................................................... 101
6.2.1. Defining the Kampala CVM scenario: Questionnaire design ................................................................. 101
6.2.2. Details of the survey .............................................................................................................................. 102
6.3. The Results....................................................................................................................... 106
6.3.1. Social economic characteristics of respondents .................................................................................... 106
6.3.2. Parametric Results ................................................................................................................................. 112
6.3.3. Discussion of Results .............................................................................................................................. 115
6.3.4. Reliability Checks .................................................................................................................................... 120xiv
6.5. Estimating WTP. ............................................................................................................... 120
6.6. Conclusion: ....................................................................................................................... 121
CHAPTER SEVEN: KAMPALA COST­BENEFIT ANALYSIS                                                             123
7.1. Introduction...................................................................................................................... 124
7.2. The Kampala Water Quality‐Improving Project costs:..................................................... 124
7.3. The Kampala Water Quality‐Improving Project benefits................................................. 126
7.4. Discounting the Kampala Water Project.......................................................................... 126
7.5. Project life span................................................................................................................ 126
7.6. Decision Criteria: .............................................................................................................. 127
7.7. Sensitivity Analysis........................................................................................................... 129
7.8. Feasibility studies for improving drinking water quality in Lower income countries...... 131
7.9. Conclusion ........................................................................................................................ 132
CHAPTER EIGHT: CONCLUSION AND RECOMMENDATIONS                                                135
8.1. Conclusion: ....................................................................................................................... 136
8.2. Recommendations ........................................................................................................... 138
REFERENCES                                                                                                                                              140
APPENDICES....................................................................................................................................... 154
Appendix A: The Survey Instrument......................................................................................... 155
Appendix B: Kampala water­Existing flow Zoning............................................................... 165
Appendix C: Kampala Water Quality­Improving project Costs and Benefits Outlay
over 52 years, Adjusted at 12% discount rate....................................................................... 166
APPENDIX C: Complete Kampala CVM Model: The MLE SPSS­output ............................ 168xv
CHAPTER ONE:
INTRODUCTION AND BACKGROUND OF
THE PROBLEM2
1.0. Introduction
In 1992, at the International Conference on Water and the Environment (ICWE) in Dublin,
comforting predictions relating to interventions in water in developing countries like Uganda
were made. Among these predictions was the alleviation of the development challenges in
Lower Income Countries (LICs), such as poverty and ill health, if provision of safe water to
households was increased in these countries (WMO 1992). At the end of the conference, in the
Dublin statement and the conference report, the guidelines to the new water resources’
management paradigm, were adopted. A commitment to the Dublin statement was made by the
heads of state in Rio de Janerio in June of the same year.
The Dublin statement advocated for a change away from irresponsible use and mismanagement
of water resources as they are renewable but finite. In the very fact of the water resources being
finite lies the other important recommendation of the Dublin statement‐  one of recognising
water as an economic good if past mistakes are to be avoided in its management (WMO 1992).
However, the statement cautioned that the recognition of water as an economic good should in
no way impede on the people’s right of access to safe clean water (ibid). The Dublin statement
initiated an enthusiastic interest in water and its implications for growth and development in
LICs.
The global agenda since 1992 has been supportive of the cause of managing water better if
higher growth is to be realized in LICs. More importantly, international fora have repeatedly
advocated increasing the level of access of safe water to users in LICs, many of whom are at risk
of disease. At the second World Water Forum in 2000, the African water vision was born. The
water vision recognizes the potential of improved water management in addressing Africa’s
major challenges of poverty and kick starting the process of sustainable development. At the
Johannesburg World Summit on Sustainable Development (JWSSD) of 2002, it was
recommended that developing nations incorporate the provision of safe water to their3
population if the poverty levels were to be halved by the year 2015 (UN 2002). Also, at the 58
th
UN general meeting in 2002, a resolution was adopted to declare the decade 2005‐2015 the
‘Water for Life Decade’. This resolution, geared toward provision of safe water to the people
currently in need of it, only echoed the water related goals committed to in the 2000 Millennium
Development Goals (MDG).  
Close to fifteen years since the adoption of the Dublin statement and amid the overwhelming
support internationally, the efforts towards achieving the water‐related goals have met with
little success in the intended countries such as Uganda, if at all.    Over 1.1 billion people
(WHO/UNICEF 2005) still lack access to safe water, as was the case in 1992 (Blackman 2001). The
direct consequence of this has been a high mortality rate related to water diseases. It is
estimated that 42, 000 people die every week due diseases related to low water quality and
sanitation, and that 90 percent of these deaths occur among children below the age of 5 in
developing countries (WHO/UNICEF 2005).
The economic costs of not improving the quality of water in the developing countries are grave
and weigh down these countries growth efforts. Since water is a basic need, safe water provision
improves societal health (WHO 2004). Health is an important determinant of growth in the
developing world as the primary sector still plays a major role in these countries’ production and
national income generation. In Uganda, for instance, agriculture, the country’s leading income
earner, contributes 40 percent to GDP and employs 83 percent of the population  (USAID 2005).
The leading role of labour in the country’s production (income) therefore, amplifies the
importance of health to the country’s economic growth. The other argument in favour of
improving water quality in LICs is directly linked to households’ behavior in their use of unsafe
water. Owing to the need to preserve own health, households may treat their water by either
boiling or adding chemicals to make it potable.   Some would even opt to buy still water, all of
which add to the cost of abstraction and use of the water. Improving the quality of water
(publicly) would therefore mitigate the cost of abstracting the water thereby increasing the
household’s real disposable income and hence their welfare.  It would therefore pay to improve
the quality of water in the LICs, which begs the question of why it has not been undertaken yet.4
There is no single universally acceptable definition for water quality (Raymer 2005, Winter
2005). Usually aspects of water quality such as colour of the water, turbidity, conductivity and
the coliform count are looked at to learn about the state of the water’s quality (ibid). All these
aspects (parameters) for Uganda’s raw as well as reticulated water have been deteriorating
(Tumwebaze 2006), indicating water quality deterioration.
There may be many causes of the water quality deterioration in Uganda. However, all of these
are linked to increased pressure on the water resources originating from population increase.
There has been increased pollution of the Lake Victoria, the country’s major source of water,
due to increased waste deposit in the lake (Tumwebaze 2006).    The increased pollution
hastened eutrophication of the lake making the water risky to use without prior treatment
(Klohn &Andjelic 1997). However, the water authorities maintain and many others such as
Howard & Luyima (1999), Oburu (2006) and Oyoo (2006) agree that the water they treat is of
internationally acceptable standards.   Despite these claims the end product remains unsafe to
drink. Part of the problem is that on some occasions treatment processes have been by‐passed
to cope with the greater risk of lower than required quantities of water  (Tumwebaze 2006).
The need to improve the consistency of water quality in Uganda has been acknowledged.
However, efforts to address this have been on a small scale. The most notable effort was the
home water treatment program launched at the close of 2004. The program, managed by the
CDC and the government, extends usage of the PuR purifier to low income households in the
country to treat water at home. Part of the treatment entails adding chemicals that are supplied
in sachets, each treating 10 liters of water (Allgood et al. 2005). This project has the advantage
of optimising treated water and chemicals, since less than 15 percent of household water
demanded is used for drinking and cooking purposes (Hunt 2004). However, this project has the
disadvantage of excluding the very poor that are most at risk to waterborne diseases and who
often would be unable to buy the chemical(s).
Greater benefits for the country are only possible if the country’s water quality is improved on a
larger scale, i.e. improving the quality of reticulated water. The scale of benefits for such a
project is far reaching and the intended beneficiaries are the very poor5
1.2. Background and statement of the problem
Uganda’s Water treatment and distribution are handled by the National Water and Sewerage
Corporation (NWSC), a government parastatal under the Ministry of Water, Lands and
Environment (MWLE). The NWSC was established by WHO in 1972 (NWSC 2004), and has been
operational since. If the corporation’s services to the users were ample in the 1970s, they have
fallen short of doing so since the 1990s.
Safe water access remains a big challenge in Uganda (BGS 2001, Nasinyama et al. 2001). The
table 1.1 summarises the NWSC’s outreach at the national and regional level. NWSC water
poses the least risk compared to other water sources in the country (Howard &Pedley 2004). For
this reason access to safe water in the Ugandan context is taken to mean access to NWSC
water. Kampala lies in the central region. From the table it can be deduced that only 11 percent
of the total population has access to piped water. About 20 percent of the Ugandans have
access to public stand pipes.
Table 1.1: The Uganda national and regional access to safe water as a percentage of the total
population
Source: NEMA (2000)
Over all, access to safe water in the country is estimated at 65 percent of the total population,
without any marked difference across the country’s regions. More conservative estimates
Type of Source National (%)
Central
Region (%)
Eastern
Region (%)
Northern
Region (%)
Western
Region (%)
Piped Water inside the
house
11 10 8 5 10
Piped water at Public
stand Nearby
20 15 18 12 26
Accessibility to Protected
Water sources by
Household
65 64 62 72 666
however, put the overall percentage of people with access to protected and/or safe water
sources in Uganda in a range of 46 to 50 percent (Barungi 2003, UNICEF 1999). Also it is
noteworthy that surveys in Uganda have found that not all regarded to be in the safe water
zones actually use safe water (MWLE 2004). Therefore, the estimates of the households with
access to safe water might be slightly inflated in the various data sets.
In the country’s capital, Kampala, as can be seen in table 1.2 below, access to safe water is still
low with in‐house connections at close to 15 percent and stand pipes catering for close to 9
percent of the total. Wells and boreholes and other sources cater for 72.2 and 4.5 percent
respectively.
Table 1.2: Kampala District Household Water Source
Source of Water  for domestic use Percentage
Piped water within the Household      14.6
Standing Water Pipes (Outside) 8.7
Wells and Boreholes 72.2
Other Sources 4.5
Total   100
Source: Nasinyama et al (2001)
According to the data from the NWSC (2004), 75 percent of Kampala district’s population has
access to piped water, but only 8 percent of the households actually have running water in their
houses. Protected springs cater for 36 percent of the population, while unprotected springs
cater for 11 percent. There are only 38 boreholes in Kampala district and these cater for only 0.5
percent of the district population. Only 1.4 percent of the population percent draw their water
from streams, lakes, rivers ponds and the like (NEMA 1998).
The various data sets on access to safe water are presented to preclude biasing the reader to
any particular data set. Uganda, like many other LICs, is notorious for poor record keeping
especially by the country’s public corporations. Recent and consistent past data on water use in7
the country was hard to find. This weakness of poor record keeping is acknowledged in the
Ministry of water lands and Environment’s sanitation report (MWLE 2004). Consistency
however, has been in agreement that the access to safe water, majorly referring to NWSC
treated water, is still very low, even in the most urban of settings in Uganda, Kampala.
Kampala is Uganda’s most urbanized setting with a population of 1.2 million and growing at a
rate of 5.6 percent (Nyakaana et al. 2007). The city grew from a mere 195 sq Km in 1968 to 839
sq Km in 2004; much of the expansion being through annexation of adjacent urban centers
(Lwasa 2005). Yet, something remains peculiar about Kampala’s growth; the city’s expansion is
driven by demographic shifts brought about by rural‐urban migration. The expansion of Kampala
has been so fast that planning forever has remained behind it. As a result the expansion has
been at the detriment of the city’s services and infrastructure. The number of the urban poor is
on the rise in the city and inadequacy of the services, infrastructural and environmental
sanitation problems are a common occurrence in the city (ibid).
The city’s reticulated water is one such service that has deteriorated to a state deserving
attention. So grave has been the deterioration, that during the 1998 cholera episode (a
preventable waterborne disease); the city accounted for 62 percent of the deaths in the country
(MOH 1999). At its peak the disease had an incidence of 9.4 per 10 000 people in the district‐
during the first quarter of 1998 (WHO 2001). While some sources suggest that the quality of the
water treated at Gaba I and Gaba II, the city’s major water treatment plants, is of internationally
acceptable safety standards and carries a lower risk compared to other water sources in the
country (Howard &Luyima 1999, NEMA 1998), the end product (at the user level) is not safe for
use without prior treatment. Households have to boil NWSC (tap) water before using it for
drinking (Howard &Pedley 2004).   Households that boiled water carried a lower risk of diarrhea
and other waterborne diseases in Kampala (Nasinyama et al 2001). Tourists and foreign visitors
are advised not to use the water, even for brushing their teeth. Bottled water, or boiling the
water first are recommended before use, as all water in Uganda remains regarded a potential
health risk (Sutherland 2006, UTB 2007).8
According to a risk assessment survey for piped water in Kampala by Howard & Pedley (2004),
users of the piped water are at a risk of getting infected with   diseases associated with E.coli
O157:H7, Cryptosporidium pavrum (highest risk), and Rotavirus pathogens (Howard & Pedley
2004). The three aforementioned pathogens are the most significant in posing risk to human
health and their control is essential. If they are eradicated, it increases confidence levels that all
bacterial, protozoan and, viral pathogens would have been effectively controlled in the water
supply system (WHO 2004).
It is thought by some that the piped water problems are rooted in the quality deterioration of
the open and ground water sources (Tumwebaze 2006). A survey by the British Geological
Survey (BGS) on groundwater countrywide revealed a high faecal coliform count (BGS 2001). The
survey also found fluoride, iron, manganese and nitrate to be in excess of the WHO minimum
acceptable levels for drinking water (BGS 2001). Another study by Kulabako et al (2004)
concluded that due to the numerous pit latrines in the Kampala area and the huge amounts of
solid waste carelessly disposed,    ground water was contaminated to dangerous levels, if
consumed without  treatment. The open water sources’ quality too has been deteriorating. Lake
Victoria; the major source of water in the district and country, has had all physical,
bacteriological and chemical parameters increasing, which is indicative of deteriorating quality
(Tumwebaze 2006). Stream water in Kampala tested positive for Escherichia Coli, with the
Chromocult Coliform Agar (CCA) monitoring technique, at all six randomly selected sampling
points in the district (Byamukama et al. 2000). The pathogen is a leading indicator of
bacteriological pathogens in the water (Bird 2005). Questions of how the water resources in a
country, famed for its naturalness, came to such a lamentable state can in part be answered by
looking at the water management structure that was in place in the period preceding the water
resources’ quality deterioration. This permitted irresponsible use of the country’s water
resources.
Until 1995, Uganda had no clear policy on water services’ management. More importantly, there
was no framework in place to oversee the monitoring of the water quality levels (MWLE 2004).
The 1995 constitution saw the enactment of The Environmental Statute of 1995 and The Water
Statute of 1995. The Water Statute of 1995 stipulates the code for the use, protection and9
management of water resources. It highlights the right to safe water for all. The same statute
vests all rights to manage the water of Uganda in the government and the rights are
administered by the minister and director of water (Ntambirweki &Dribidu 1999).
In 1998, the need to realign the many users of the water resources in the country and their
multi‐dimensional uses was envisaged in the enactment of The National Water Policy of 1998.
This legislation governs the use and management of all water resources in the country today.
The National Water poicy of 1998 amalgamated all the legislations that were governing water in
the country before 1998. These were: The Uganda constitution (1995), The Local Government
Act (1997), The Ugandan Plan of Action for Children (UNPAC) (1992), The Water Statute (1995),
The National Water and Sewerage Cooperation Statute (1995), The National Environment
Management Policy (1994) and The National Environmental Management Statute (1995).
Progress on the legislative front was followed by institutional reforms of decentralisation and
liberalisation, among others in the year 2002. Such reforms saw the role of government being
restricted to those of: policy formulation, setting standards, regulation, developing capacities of
local government to deliver services and formulation of guidelines (IRC 2003). The reforms
occasioned the diversification in most of the country’s institutions. In the water sector a new
institution under the MWLE, The Directorate of Water Development (DWD), was born. This
institution is responsible for the sustainable provision of water and the management of the
water services in the country at the urban and rural level. Other stakeholders in the water
services besides the MWLE include international organisations like DFID (UK) and the WHO, CDC,
and RCPEH.
After remarkable advances made institutionally in the water sector, since 1995, and amid the
enthusiastic global campaign to improve access to safe water, one would expect Uganda’s access
to safe water to have improved. However, this has not been the case. Households country‐wide
lack potable water and are at risk of waterborne diseases. Even in the most urban of settings like
Kampala, where the user population has the capacity to pay for better services, the quality of
water remains poor. This forms the problem of this research. Against this background therefore,
this research will seek to answer the question of why the water quality levels in Uganda remain10
poor. The Kampala district accounts for the consumption of over 70 percent of reticulated water
in Uganda (NWSC 2005) and so is used as the modal district for the investigation of the country’s
water quality problems.  
Improving Uganda’s water quality to drinkable standards would certainly have positive net
benefits on the economy‐  in the form of improved health gains due to the reduced risk of
disease. Already, surveys in the country have established strong correlations between
improvement of health of HIV patients and availability of safe drinking water in the country (Lule
et al. 2005). Other benefits could be: avoided health spending cost in the treatment of the
diseases caused by poor quality water (Howard & Pedley 2004, Nasinyama et al 2001). There are
also likely to be benefits in the form of avoided costs in achieving desired water quality at an
individual or a societal level (for example, energy saved in boiling water, avoided expenses in
buying bottled water). Further, there could be benefits like gained labour days after reducing the
mortality and morbidity of these health risks, given that labour is the more abundant factor in
the country’s production.
The demand for better quality water in the country and in Kampala is high. An urban water
surveillance program reported that between 85.6 to 92 percent of all households in Kampala
boil their water prior to drinking (Howard &Luyima 1999). The Kampala district report attributed
the growth of industries based on treating water to the need for better quality of public water
services, inter alia (NEMA 1998).
The potential benefits and the existing demand for improved water quality in Uganda discussed
above make the investigation of why water quality levels in the country remain poor all the
more imperative.
1.3. Objective:
The main objective of this research is to establish the efficiency case for reticulating better
quality (safe) water in the Kampala district of Uganda. The benefits, according to the
encountered literature are mainly health gains and avoided costs that are bound to come with11
the improved water quality. These benefits will be compared with the costs of providing the
improved quality water to conclude about the project’s feasibility. Other benefits, like increased
investment and tourism growth are acknowledged but not quantified. ‐
1.4. Sub­Objectives:
The study investigates why the current water supplied by the NWSC is not safe to
drink in the city of Kampala and what ought to be done to make it drinkable.
The study investigates the water situation in Uganda with an aim of identifying the
infrastructural development necessary to ensure that water reticulated is safe for
drinking
The study investigates anticipated benefits from the treatment of water levels where
it is safe to drink.
The study also investigates the costs associated with improving water quality in the
country and hence in the Kampala district.
1.5. Reasons for the Research:
The study investigates the economic benefits vis‐à‐vis the costs of improving the water quality
for Uganda. There can be no doubt that improving water quality improves public health in the
country since the risk of disease in turn, shall have been lowered as acknowledged by:
Nasinyama et al (ibid), WHO (ibid), Howard & Pedley (ibid). Also, according to Howard & Luyima
(1999), improved water quality has spill over benefits that would accrue to other sectors in the
economy. Understanding why improving the quality of the water has not happened in the
country will offer insights into what needs to be done and how it is to done to harness the
benefits water offers.  12
1.6. Scope of study:
The study investigates the feasibility of improving water quality in the Kampala district of
Uganda. The feasibility of water quality improvement is assessed by comparing the benefits and
costs of undertaking such a project in the district. The benefits of such a project would accrue to
the users of Kampala’s reticulated (NWSC) water. By implication therefore, the benefits are
quantified according to the Kampala users’ valuation. The costs are estimated in liaison with the
suppliers of reticulated water in Kampala‐ NWSC.  
The Kampala district was chosen for research because it is the country’s capital and the only
district designated as an urban municipality. The district is home to three quarters of the NWSC
user population. The level of service provision in the district sets the trend for other districts
and/or regions. The district suffers the problems of rapid urbanisation and population growth;
phenomena thought to be at the root of water and sanitation problems in developing countries
(Wall 2000). The risk of water quality deterioration is highest in this district. By implication the
risk of waterborne diseases is highest in Kampala as proved the 1998 cholera episode.
1.7. Methodology:
The methodology employed in investigating the efficiency case of the proposed project is Cost
benefit analysis (CBA).    CBA is a project appraisal tool employed to determine the economic
viability and/or feasibility of proposed or competing projects. The project costs and benefits are
compared. Where these occur in different time periods, a discount rate is used to convert all the
values to their present values for meaningful comparisons to be made, as money has time value
(Constantinides 2000). A CBA requires the quantification of the benefits and costs and the
determination of the discount rate.
From a theoretical standpoint, CBA is aptly suited for assessing the efficiency case of the project.
Efficiency in welfare economics is that situation that leaves no potential to make one individual13
better off at the expense of another. In reality, the actual compensation of the losers is rarely
bothered about‐  just that they potentially could be compensated (Brent 2003). If the analysis
finds the project to have positive net benefits, it implies there are enough benefits to
compensate the losers; this qualifies the project as (potentially) efficient (ibid).
The costs of a project to improve water quality include chemical purchases, personnel payment,
water plant upgrade and reticulation‐accessories’ expenditure‐  information obtained through
interviewing the water engineers. The estimation of the project’s benefits is rather more
difficult, as the benefits from improved water quality may include non‐market values such as
Passive Use Values (PUVs) that have no behavioral traits for reference to be made in estimating
their value.
Where these values are involved, the use of non‐market valuation techniques is recommended.
One such technique is the Contingent Valuation Method (CVM). The CVM is a survey based
technique for estimating the value of non‐market resources. It is estimated by measuring the
individuals’ willingness to pay (WTP) for an improved resource scenario, or their willingness to
accept (WTA) compensation for reduced resource service (Arrow et al. 2002).
The main advantage of this method is that it avoids the need to value each individual project‐
benefit separately. The valuation question informs the respondents about the projects benefits
and their WTP/WTA estimates the value of these benefits in their totality. The final value of the
project’s benefits is arrived at by predicting the WTP/WTA for the median respondent (or the
mean of these).
The application of the CVM over time and across many fields of analysis has invited a lot of
criticism, mainly centering on the plausibility of the predicted values estimated by the method.
To this end, a panel consisting of prominent economists was assembled by NOAA to establish
the reliability of the CVM. The panel concluded that any well designed CVM would generate
reliable results (Arrow et al. 2002). A well designed CVM is one that is designed in line with the
guidelines the NOAA CVM panel prescribed.   In this vein, a questionnaire was designed to elicit
improved water quality benefits, with the NOAA panel CVM guidelines in mind. The respondents
were random reticulated users interviewed in Kampala14
1.8. Organisation of the study:
The rest of this thesis is arranged as follows: Chapter two furnishes the theoretical platform on
which the research is premised. The theory ranging from economics, epidemiology and
Geography serves to show that the improvement in the quality of the water for the Kampala
district has economic potential in improving quality of life of the poor. Chapter Three focuses on
the description of the Ugandan water system. Here, a detailed account of the institutional setting
and an investigation into the quality status is made, of water used by the Households household.
The next two chapters lay the theoretical platform for the methodology employed in the
investigation of the research problem. In Chapter Four, CBA application and the issues
surrounding it are furnished and the chapter five furnishes the background for the CVM’s
application. The chapter six presents the results of the CVM survey conducted in Kampala. The
Kampala water quality‐improving project CBA is presented in chapter seven. Finally, in the
chapter eight conclusions are drawn and recommendations made.15
CHAPTER TWO:
LITERATURE REVIEW16
2.0. Introduction
The case for this research, as can be gathered from the preceding chapter, was built on the
premise of improving the country’s health and thereby its human capital stock. In line with the
WWC (2000) prescribing, this study recognizes four water use categories namely:
water for people and industries
water for food and rural development
water for nature
water for energy production
Some level of water quality is essential for all these uses. This study however, focuses only on
the first use. Only this use has universally defined guidelines; the recent stipulation of this is in
the third edition of the WHO (2004) Guidelines for drinking water quality. The rest of the uses
guidelines are subjective and conditioned to the locality of investigation. The water for people
and industries’ use can further be subdivided into four sub categories namely water for basic
needs, for other household needs, for services and for industry (Zehnder et al. 2003).
The basic needs use of water, covers exposure of the user to disease causing organisms because
water acts as the organism‐conduit. The advances in epidemiology by the work of John Snow
(1855) in the 1850s occasioned this awareness. After that, disinfection became an important
part of public water supplies’ treatment, motivated by the need to preserve public health (CRC
2003). An economy’s performance heavily draws from the public’s health status. These linkages
between health and economic performance have attracted interest from various scholars over
the last two decades. For instance, in the work of Schultz (1961), Mushkin (1962), Dasgupta
(1993), Hamoudi & Sachs (1999), van Zon & Muysken (2003) McDonald & Roberts   (2004), to
mention but a few, the general conclusion has been that health is both a cause and consequence
of economic success. Empirical evidence has also been furnished linking health indicators like
child mortality, life expectancy and the fertility rate with economic performance (Hamoudi
&Sachs 1999).17
The lessons learned from focused experiments on workers’ productivity and the history of cross‐
country economic growth lend credibility to findings of the studies earlier noted, relating health
to economic growth. Workers in Sri Lanka treated for iron‐deficiency anemia were found to be
more productive (Selvarantnam et al. 2003). Cross‐country annual GDP per capita examined for
a period of 25 years in the particular ecological zones was found to correlate with a decrease in
the prevalence of malaria (Sachs & Gallup 1998).   A study on the challenges of emerging Asia
attributed the ‘miracle growth’ of the East Asian economies to improved health, and it is
reflected in the decreased child mortality rates and increased life expectancy the economies
experienced since the 1940s (ADB 1997).    The life expectancy in the East Asian economies
improved by a whole 15 life years from the 1960’s to the 1990s i.e. from 56 to 71 (Page et al.
2004). The African Development Bank (ADB) (ibid) study found that the health parameters
contributed approximately 0.5 to 1.3 percent of the accelerated annual growth of the ‘Asian
Tigers’. More generally, cross‐country data reveals that one additional year of life expectancy is
associated with sustained increment of 4 percent in national income (WB 2006).
Furthermore, there are some who have attributed the superior global economic growth of the
Twentieth Century to improvement in health (Hamoudi & Sachs 1999). A look at the global
economics indicators supports Hamoudi & Sachs’ (ibid) findings. It shows that lowering mortality
rates after the 1950’s coincide with a rising global growth (Wikipedia 2007a). Health is therefore
an important precondition for economic performance and should be maintained. The high
prevalence of waterborne diseases in Uganda, as can be gathered from MOH (1999) data,
implies that the provision of improved quality water will go a long way in improving public
health and this may well  translate into positive economic outcomes for the country.  
Changing conditions in the water sector however, pose a challenge to the provision of improved
quality water for the public. Water is getting scarcer (Barlow 2001, Bakker 2002, Hunt 2004,
Kravcik et al 2002, Postel 2002). Its provision therefore, ought to be in the most efficient
manner. This theory in itself has added to the debate, by raising the question of who is better
suited to provide water‐ the public or the private sector? The current failures in the developing
countries are a manifest of the government’s failure to avail the public water services in these
countries, and so the private sector should be allowed a leading role in their provision (Budds18
&McGranahan 2003). However, if left to the private sector, the equity goal of the services’
provision is likely to be sacrificed.   More compelling amid the uncertainty of who is best suited
to provide the services, is the need to improve water quality sooner if the developing countries
like Uganda are to realize the attributed benefits of such an undertaking in the near future. This
chapter will therefore survey the literature surrounding improving water quality, the
implications of such a project on an economy such as Uganda, and debates arising from
experiences around the world.
The chapter is divided into two arranged as follows: the first part presents theoretical platform
for water quality and its implications to an economy, the other covers the issues surrounding the
provision of that water.
2.1. The Economics of Health:
Health in many respects is different from other forms of human capital. Investment in activities
such as higher education skills acquisition and work experience are undertaken for their direct
reward, increased income. For example, an employee in a lower position will invest in higher
skills or higher education to gain a higher productive position that affords them higher pay. The
same motivation would explain why individuals seek added work experience. Health investment,
on the other hand, is more complex. Individuals are not rewarded directly for their investment in
health. In economic theory, such an investment would be deemed irrational; yet individuals
continue to invest in health. The missing incentive by no means diminishes the importance of
health as a part of the production process; rather, it makes the task of delving into the theory of
demand for health more worthwhile.
The most influential model in the analysis of the demand for health is the Grossman (1972)
model. The model is premised in the human capital theory which suggests that individuals invest
in themselves (through either education or health) to increase their productivity. To undertake
these investments, individuals need money income. Part of the potential benefits of the
proposed Kampala water quality‐improving project, were identified to be lowering of mitigation
costs (see section 1.2 chapter 1). The lowering of mitigation costs would increase the individuals’19
money income and thus their investment in health which resultantly increases their health stock
and productivity. Using the Grossman (ibid) model it shall be shown how the various choices the
individual makes combine to influence their health.
2.1.1: Demand for Health: The Grossman Model
Grossman (1972) has argued that health demand is different from traditional demand theory. As
a point of departure, he argues that “a person’s stock of knowledge affects his market and non
market products, while their stock of health determines the amount of time they can spend
earning income” (Grossman 1972). The axioms on which he based his model were:
Demand for health and the other inputs of health is not for their demand parse, but a
derived demand for utility.
Individuals are not passive consumers of health, but are active producers who spend
time and money on its production.
Health is lasting and depreciates over time and can therefore be analysed as a capital
good (Folland et al. 1997).
The model envisages the individual to have a considerable control over their health in their
being able to influence their health‐affecting consumption patterns, their health care utilization
and their environment.
Health, therefore, is demanded mainly for two reasons, consumption and investment. In terms
of its investment role, health is the determinant of the total time available for market and non‐
market activities now and in the future. In terms of consumption, health is demanded for direct
utility, such as the quality of life years it avails. These two qualities qualify health as a factor of
production (Grossman 1972).
With the aid of a production function, a relationship between a factor of production and
changes in its inputs can be investigated. This function for health is shown in figure 2.1 below.
The health production function in the figure 2.1 shows that an individual is born with a minimum
stock of health that defines their initial health stock H0. This health stock increases as the20
individual increases consumption of health care inputs, but at a decreasing rate (Folland et al.
1997).
Figure 2.1: The Health Production Function
Source: Folland et al. (1997)
Other than health care inputs, a host of activities done in our leisure time, like going to the gym,
watching TV, strolling in a park, to mention but a few, also contribute to an individual’s health
stock. These activities, Grossman refers to as composite goods B. The individual’s inter‐temporal
utility function is an aggregate of all these inputs and is defined as seen in the expression 2.1
below.  
 
In the equation 2.1 H0 and Hi are the initial stock of health and stock of health in subsequent
period i. The symbol Φ stands for the number of healthy days per unit of health stock and Bi
is
the consumption level of the composite good B in the period under review.
The composite good B and health care are the health investment inputs.  Health investment may
be defined according to the expression 2.2 below:
Health Stock
Health Care inputs
H0
H1
H2
M0 M1 M2
( ) 0 0 N N 0 N
U = U Φ H .......Φ H ,B .....B ………………..  (2.1)
i 1 i i
δi
Hi
H+
−H =I − ……………….. (2.2)
Health Stock = f(Health care inputs)21
……………….. (2.3)
    ………………. (2.4)
In equation 2.2, Hi and Hi+1 are the current and future health stock of an individual respectively, Ii
is gross investment in health in period i and  δ  is the rate of depreciation of health. The
expression 2.2 implies that the stock of health a person can enjoy between now and in the
future is equivalent to the gross investment undertaken in health adjusted for deterioration in
health in the current period (Grossman 1972).
The model, as earlier noted, is premised in the human capital theory, which envisages
individuals investing in themselves to increase their productivity. These investments are
undertaken on the basis of their relative costs and benefits. The health investment decision is
determined by the resources available to the individual. These are time, health care inputs and
market good inputs (ibid). The functions for investment in health and composite goods are
depicted below:
                                       I=I (M, TH, E)
                                       B=B (X, TB, E)
Health Investment (I) is directly influenced by the amount of medical goods (M) available and
time out of their leisure they devote to health (TH). Likewise, the stock of composite goods is
directly influenced by the amount of market goods
1
(X) and the time out of their leisure they
devote to production of the Composite good B (TB), as expressed in equation 2.4 above.
The parameter E, in the functions 2.3 and 2.4 above, is the factor that is responsible for the
efficiency with which B and I can be increased. This is influenced by the education level of the
individual (Folland et al 1997). The more educated people are, the healthier they are (ibid).
Hamoudi & Sachs (1999) found that improved literacy of women improved child survival. The
welfare gains from a decision such as to increase time (TH) spent on health improving activities
would be such as: reduction in time lost to illness and increasing the individual’s productivity at
work. In efficient markets, higher productivity is translated into higher income. Increasing the
time TB spent on the consumption of the composite good B would have the same effect as TH. it
                                                     
1
 Goods used in the production of the Composite good B 22
follows that devoting more time to health investment is desirable to an individual (Folland et al.
1997).
The individual must choose how they allocate their income this between the consumption of
health care inputs and other goods. This income is determined by how much time they devote to
work at a given wage rate. The individual’s labour‐leisure time trade off is as depicted in the fig
2.2 below
Figure 2.2: The Work Leisure time trade‐off
Source: Folland et al. (1997)
The figure 2.2 above represents the labour‐leisure trade‐off of the individual. On the vertical axis
the income earned is measured, while on the horizontal axis the amount of time available to the
individual is measured. The individual has a 365 day year which he has to allocate between work
time and leisure time. If an individual chooses to have time OT0 for leisure, they leave time T0W
for work. Given the wage rate WW, this earns them income Y0 at equilibrium point E0. This
equilibrium position affords them utility level U0. The individual could choose a different
allocation of their time, such as increasing their time spent on health from TH0 to TH1. If the net
effect from this allocation is gain in available time, then the health investment pays off. As a
Income
Time
365‐TH0‐TL0 365‐TH1‐TL1
Y1
Y0
T0 T1
O
U1
U0
W
W W’
W’
E1
E023
result the individual’s future income (W’W’), leisure time (OT1) and utility (U1) are increased at a
new equilibrium position E1.
The other choice the individual has to make is one of deciding how to allocate the resources
available to produce health and the composite goods B (Grossman 1972). In the figure 2.3
below, the production possibilities’ trade‐off for the two inputs is illustrated
Figure 2.3: Allocation of resources to produce health and composite good B
Source: Folland et al. (1997)
This production possibilities curve (PPC) is different from the usual PPC as there is no real trade‐
off in the allocations between points A to R. Choices moving away from point A up to point R
imply more time being made available for health, and at the same time making more time
available to market goods to produce composite good B (Folland et al. 1997).
If the individual obtains no intrinsic value from health, i.e. health being demanded only for the
production of the composite good B, the individual’s indifference curve would be a vertical line
such as U1. The point of optimization would be point R where the individual has B0 amount of
the composite good and H0 of health. At this point the person does not trade any amount of the
composite good B to gain more health (ibid).   If the individual gets some utility from both the
home good and intrinsic utility from health, then the indifference curve may be shaped like the
A
Composite Good B
Health
O B1
R
N
U2 U1
H0
Hmin
H1
B0
Production possibility curve
Indifference curves24
curve U2 (see figure 2.3 above), and optimal allocation would be such as at point N; here the
individual reduces the total amount of the composite good (B0‐B1) for more health (H0‐H1).
The optimal choice of health investment is where the individual combines these (resources) for
the marginal benefits of health to equal the marginal costs of investing in it. The marginal cost of
health   (MC) – the opportunity cost  ‐ is the sum of rate of return on other investments (r) and
the rate of depreciation of health of health (δ); algebraically we have that MC = r + δ. Both r and
δ  are exogenous. The marginal benefit of health (MB) is the rate of return on investment in
health in which could be realized in both market and non‐market sectors.
The marginal benefit is also the same as the marginal efficiency of health capital (MEC); this
declines with increased investment as can shown in figure 2.4 below.
Figure 2.4: The Marginal Efficiency of health Capital
The MEC is convex indicative of diminishing returns to investment in health. At the equilibrium
such as E0 the marginal costs are equal to marginal benefits.   Changes that increase the rate of
depreciation of health (for instance, from  δ0 to  δ1), such as aging, increase the marginal cost of
capital. Consequently, the demand for health capital declines as shown by a movement along
the MEC1 curve to equilibrium point E1. The demand for health care may increase as health has
an inelastic demand curve (Folland et al. 1997).  An increase of the wage rate and the education
Cost of capital
r + δ0
r + δ1
Health Stock
H1 H0
O
H2
MEC1
MEC2
E0
E1 (aging)
 
Source: Folland et al (1997)
E225
level shift the MEC curve outward. An increase in either the wage rate or the education level
would change the equilibrium position to E2 (figure 2.4 above).
In figure 2.5 below, the different choices the individual has to make to invest in health are
considered at in the same analysis.
Figure 2.5: The combined Analysis of the health demand model
.
If the individual earns wage w0w0, their consumption possibilities frontier is defined by the outer
most curve. The equilibrium point is where their indifference curve is tangent to their
B1
U3
I  [Consumption
possibilities]
III [Budget Constraint]
II [Production Function]
Health Inputs Consumption
Consumption
Health
E0
E1
45
0
IV
M0 M1
H0
H1
U
U
w0
w0
w1
w1
B0
Source: Dolan (2003)26
consumption possibilities frontier‐ this is point E0. This optimal level affords the individual health
inputs’ level M0, composite goods B0, health stock of H0 and optimal utility level of U0. The wage
rate is reduced to w1w1, the consumption possibilities’ frontier shifts inward, implying reduced
levels of health inputs (M1) and composite goods (B1) for the individual and consequently,
reduced health stock (H1) and utility (U1). An increase in the price of goods, aging and long‐term
diseases like cancer, would have similar effects as the reduction in the wage rate. On the other
hand, higher health stock levels and utility such as U3 are attainable if there is an increase in the
individual’s income, the education or through subsidization of health.
2.1.2. Critique of the Grossman Model:
There are three main critique of the Grossman (1972) model. Firstly, the model assumes that
health care is a constant life time investment (Grossman 1972). This is too simplified an
assumption for any investment. The model also ignores the role of insurance markets in health
demand, which is a major determinant, especially in the developed world.
Secondly, some have argued that the Grossman Model ignores the productive feedback of the
population health at the macro level, only emphasising the micro economic demand
perspective. The macro level feedbacks are the reason health services are provided (vanZon
&Muysken 2003). Implicitly under the Grossman model, national health services, in spite of their
benefits, should not be provided in an economy.
Thirdly, the model is overly deterministic, including the individual choice of when to die
(Grossman 1972)
2.1.3. Applying the Grossman Model to the Ugandan context:
The Grossman Model offers some direction in regard to what should be done if policy makers
wish to improve human capital investment stock in terms of health.  27
The model implies inter alia, that the more educated the population is, the more efficient the
production of health shall be. The other implication is that subsidization of health care inputs
can also be beneficial in increasing societal health. Furthermore, the model suggests that an
improvement in the wage levels of the workers persuades them to increase their level of
investment in health. This investment would be reflected by shift of the equilibrium E0 outward
in Figure 2.5 above.
While the first two implications may be applicable, the last is contestable. Wage levels of
workers are not likely to be raised every time improved health is achieved because this is
precluded by financial resources’ constraint in a country like Uganda (which is caught up in
poverty cycles). In light of this fact one could also argue that subsidization of health care services
may not be very frequent in the country.
A project to improve the reticulated water to drinkable standards for the country could
however, indirectly subsidise health care and hence health investment for Ugandans. If
envisaged in the Grossman model perspective, the project lowers the risk of waterborne
diseases, ceteris paribus, this would lower treatment and prevention expenditure (attributed to
waterborne diseases). The reduced expenditure effect increases individuals’ income in real
terms, even for the poor, and resultantly their welfare. This would be exhibited by a shift of the
budget line outward increasing health investment level and hence health stock of the
individuals.
Such a project will be beneficial if the populace most affected by waterborne diseases is
significantly serviced by the reticulation system. Also, benefits attributed to the project are more
likely to be realized if the waterborne diseases formed a significant part of the household’s
expenditure and the per capita cost of the new water project is low.
2.2. Beyond the Grossman model:  
The Grossman Health demand model is arguably still one of the most influential in health
economics (Dardanoni 1986). Later work in this area of health economics has mainly been in the
form of adopting the model. For instance, Dardanoni (1986) sought to present the model in a28
more simplified way, arguing that the technical complexities hindered the model from being
extensively usable. Employing a dynamic analysis of the costs and benefits of health, Dardanoni
(1986) found that his comparative static results concurred with Grossman’s findings. Education
had an ambiguously positive effect on health of an individual. Dardanoni (ibid) further
investigated impacts of a variable absent in Grossman’s model  ‐ inflation. He (ibid) found that
change in capital gain (inflation) will in the short term increase optimal stock of health. In the
long run however, there would be an accumulation effect on the cost of capital, such that
optimal investment decreases.
In the mid 1990s an important contribution in health economics was made by Lofgren &
Johansson (1995). The Grossman model is an archetype of the true market economy, where the
individual choices are made nearly uninhibited and so influence demand and supply and the
economy at large. Further more, the model makes no mention of what the magnitude of optimal
health is set as. These are the inadequacies addressed by Lofgren & Johansson’s (ibid)
contribution.
Lofgren & Johansson (1995) argue that the market economy cannot correctly account for the
positive production externality resulting from health capital. This accounting necessitates
implementation of a subsidy to the individuals. In their analysis, Lofgren & Johansson (ibid)
compare optimal condition of the command economy and the optimal conditions of the market
economy, in order to get an idea of how health externality can be internalized. The individual is
not compensated for their health’s contribution to production. Thus they are not given a proper
incentive to invest in accumulating health capital. To internalize this externality and enable the
optimal allocation of health to be made, Lofgren & Johansson (1995) suggest that individuals
should be given a unit subsidy Ph=fh(.) where fh(.) is the marginal productivity of health capital
along the optimal path of its allocation.
The subsidy may be financed by taxing the firm according to the function Tt=Phh; where Tt
is a
lump sum tax since the firm cannot change the use of health capital. This subsidy can only
happen with government intervention because; the market economy therefore will not provide
the level of optimal health (Lofgren & Johansson 1995).29
Health is demanded, not for its own sake, but as a means to aid production and its most optimal
allocation would be under government’s supervision (subsidization).
2.3. The water health linkages: the John Snow Model.
The link between health and water was recognised by John Snow (1855) the 1850s’ London
cholera episodes.   Snow (1855) established the link by plotting the cholera cases on a map. The
plot revealed that cholera episodes occurred in tightly clustered localities around a water pump
on Broad Street. The pump handle tested positive for the Cholera vibrios. The spread of the
disease was halted by removal of the pump handle (Marsden 1995).
In his Grand experiment, conducted over the period 1853‐54, Snow (1855) established a
causative link between the quality of the water and health of the users. He used the mortality
rate due to Cholera as the proxy to measure health.    Choosing two water sources for the
investigation, Snow sought to correlate water use and health.    The one source was the river
Thomas 20 miles upstream of London. This water was distributed by Lambeth Company. The
second source was the local wells managed by Southwark and Vauxhall Company. All this water
was delivered untreated to the users. The table 2.1 summarises Snow’s results.
Table 2.1: The Grand Experiment results  
As can be seen in the table 2.1 above the low quality water source (local wells) managed by the
Southwark and Vauxhall Company were associated with more cholera deaths. The statistics
indicate that the quality of the water influences health.
Number of
Houses
Cholera
Deaths
Deaths Per 10,000
Houses
Southwark and Vauxhall Company 40,046 1,263 315
Lambeth Company 26,107 98 37
Rest of London 256,423 1,422 59
Source: Marsden (1995)30
The leading scientist of the time William Farr (1852) contended that the major factor responsible
for the disease’s spread were miasma and elevation of the districts. However, a reexamination
of the results of the two scientists in    Bingham et al.(2004), confirms Snow’s (1855) earlier
position that water quality was responsible for the spread of the disease. The transmission mode
applies to all other diseases categorized as waterborne.
2.3.1. Applying the Snow model to the Ugandan scene
The spread of water borne diseases is through a feacal‐oral transmission according to Snow’s
(ibid) theory. The theory establishes the link between health and water and can be employed to
understand the health deterioration episodes related to water. Such linkages could be useful in
understanding the Uganda cholera outbreak during 1997‐98.  
The period was marked by constant flooding resulting from the El Niño rains experienced in the
country at the time. In spite of the national climatic phenomenon, Kampala was most affected
by the epidemic (Hulme et al. 2005). The district accounted for 61.6 percent
2
of the cholera
cases in the country in the outbreak of 1997 (MOH 1999). An investigation into the epidemic
revealed that the brunt of the disease was felt in the slums of the district, particulary in: Kasubi
Kyebando, Kisenyi, Nateete, Luzira, Kisugu, Kalerwe, Bweyogerere and Kibuli (Rees‐Gildea
&Geleta 1999). The slums are characterized by highly densely populated habitations with very
unsanitary conditions. The households are highly impoverished, drawing their water from the
open sources. As earlier mentioned the open water sources, in part due to increased settlement
and negligence in the district, has a high feacal coliform count (Byamukama et al 2000, Howard
& Luyima 1999, Howard & Pedley 2004). The floods’ role therefore, was one of conveying the
pathogens to the next water source. Those that used the water were at risk of infection. Hence
the cholera epidemic in the district. The fact that the highest case of the epidemic were
recorded in the most densely populated districts of Kampala, Mpigi, and    Bugiri (MOH 1997)
                                                     
2
 Calculated form the Ministry of Health’s epidemic Statistics31
serves to show that health deterioration is positively related to the rate of contamination of the
water
3
.
The experience from Uganda reveals one other fact; that the poor (in the slum areas mentioned
above) are always highest at risk of being affected by the waterborne diseases.
2.4. Poverty Water and Health Linkages:
Owing to the fact that water is a necessity of life, when costs pertaining its provision are raised
the poor are bound to be most affected. Such costs increments could arise in the case of water
quality deterioration. Any increased marginal cost of providing potable water, like costs in
changing the source of abstraction and costs of additional treatment, add to cost burden on the
poor.  
The link between water, poverty and health drawn from the literature surveyed is summarized
in the figure 2.6 below:
Figure 2.6: Framework for the water poverty and health linkages
                                                     
3
This is on the assumption that the more densely populated an area is the more the water sources are likely to be
contaminated as was the case for Kampala
Increased Expenditure and
Loss of revenue
ILL HEALTH
HEALTH
Avoided costs translate into benefits
Dependence on Nature leads to
overexploitation
Water Diseases
Natural Hazards
WATER
Social
Economic
Good
Social
Economic
Production
Consumption
Environment
Investment in Water
Management
Economic
Growth
Poverty
Water Reclamation  Water Treatment 32
Water in its raw state could be either a social economic good or a social economic bad
(Abayawardana &Hussain 2003). As a social economic good, water plays an important role in
economic growth of a country. Water’s roles in the production process, maintaining the
ecological balance, and for consumption are all welfare improving. Its consumption contributes
positively toward an individual’s Quality Adjusted Life Years (QALYs). The addition, ceteris
paribus, increases the productivity of workers, resultantly contributing toward economic growth.
As a social economic bad, water stifles economic growth. The importance of water as a bad for
this study is its role in perpetuating poverty. Water diseases’ treatment and prevention for
instance, increases household expenditure. For the poor the increased expenditure reduces their
income and welfare, thereby perpetuating poverty.
The other linkages between water and poverty may not be as direct. For instance, it is argued
that demographic impacts of low levels of health are high fertility rates in a bid to ensure
survival (Hamouid & Sachs 1999). This high fertility rate increases demand for resources by the
households and so reducing income available to satisfy other needs. Ill health also increases the
cost of trade commerce and investment (Hamouid & Sachs 1999), by making harder to reap the
full gains from the activities that would otherwise have improved societal income and/or
welfare. It could be argued further, that the water diseases reduce productivity and
performance of labourers thereby inhibiting their ability to earn.
There are also backward linkages between poverty and social bad consequences of water, as can
be seen in figure 2.6. The prevalence of poverty compels households to subsist on the natural
resources (water). This reliance overtime, is likely to breed overexploitation of these resources,
resulting in social ‘bads’, like waterborne diseases. Scientists employ Murray & Lopez’s (1996)
Disability Adjusted Life Years (DALY) methodology to quantify the approximate the losses of the
burden of disease. The DALY combines morbidity and mortality into a single outcome weighted
by the risk of disease. In monetary terms, the Cost per DALY can be estimated to quantify the
gravity of the burden of disease. Though not extensively used in Uganda, small scale applications
of the methodology in the work of Howard & Pedley (2004) indicate that the losses due to
waterborne diseases are significant.     33
The figure 2.6 also shows that if intervention is made, the social bad consequences of water can
be averted, leading to economic growth. Intervention can be such as water reclamation that
would enable reuse of water and so reducing the strain on natural water resources. Another
intervention, presumed to be less costly than the former, is water treatment. This intervention
would ensure that the waterborne burden of disease is reduced and resources like time, money,
fuel etc, formerly devoted to securing usable water, are put to more productive use. These
interventions’ benefits are more likely to accrue to the poor, who normally are not in a position
to prevent the social bad impacts of water (Abayawardana &Hussain 2003).
2.5. Changing conditions: The scarcer resource.
Common to most analyses regarding water scarcity, has been the notion that overexploitation of
the resources has driven the water resources near the point where they cannot sustain the life
on Earth, let alone the future generation. If allowed to continue, the current water withdrawals
will result in a water stress/scarcity crisis (Barlow 2001, Hunt 2004, Kirby 2004) .
Water stress is a situation that results from an imbalance (deficit) between use and availability
of water resources (WWC 2005). After perpetuated water stress, comes water scarcity, which,
according to Abrams (2001), is the shortage of freshwater availability from renewable resources
to meet the essential demand for water, like domestic use and agriculture. Unless current
withdrawals are lowered or better conservation methods for the water resources found, the
pundits warn that the sustainability of activities like agriculture, on which human life depends, is
not guaranteed (Barlow 2001, Stikker 2002, Hunt 2004, Postel 2002). Some contend that the
current water quality problems are a manifestation of the early signs of the crisis needing
immediate attention (Hunt 2004).
Important to the water stress argument  is the fact that there is a near constant supply, of water
available in a given location (Postel 2002).   This amount is determined by global water supply,
which is also more or less near fixed volumes. In spite of the Earth being 70% water, only 0.26 %
of that water is available for use, 97.5% of the water is saline and 2.24% is locked in the polar
regions as ice (Samson &Charrier 1997, UNESCO 2003).   34
Figure 2.7: The Global water Resources
The estimates in the figure 2.7 above show the percentage of freshwater available and how
much of this freshwater can be withdrawn, that is, groundwater and the water in open water
sources, which is 30.8 and 0.3 percent respectively.
The estimates point to the fact that water is a renewable but finite resource as agrees (Postel
2002). The water cycle process, powered by solar energy, draws water from the water bodies by
evaporation. The drawn water is condensed at great heights and is redistributed all over by
various precipitation forms. Estimates of the World Water Council (WWC) put the amount of
rain water in excess of evaporation on land at 44 000km
3
(WWC 2005). This estimate, in its
strictest form, combined with the total water available in all the aquifers would put per capita
water availability per day at 15,000 litres (Ibid). In theory hence, there should be no talk of a
water scarcity/stress crisis. However, in reality this water is not evenly distributed in space and
time. Most regions receive 100% precipitation for only brief rainy seasons in the years. On
average the amount of water potentially available from the renewal process is 10,000  ‐
12,000Km
3
per year (WWC 2005) and not all of this is withdrawn. In 2000 for instance,
approximately a mere 15 ‐ 30 percent of this was withdrawn which is 2000 ‐ 4000 Km
3
(ibid).
Source:  Kirby (2004)35
Figure 2.8: Estimated annual world water use
Be that as it may, water withdrawals have been (and still are) increasing globally.  Most notably,
increases have been in agricultural and industrial use as depicted by the figure 2.8 above.
The world population tripled in the Twentieth Century, and as a result the water withdrawal
increased six fold (Kirby 2004). Since 1970 the world population grew by 2 billion and the per
capita water availability fell by one third (Hinrichson & Upadhyay 1998). It is estimated that the
global consumption of water doubles every 20 years. At this rate, the demand of water above
the current use will be 56% higher in the next two decades (Barlow 2001). Considering that
already, there are 1.1 billion people presently without safe water (UN 2002, Kirby 2004), the
statistic bodes a grave challenge in the years to come as regards water availability.
Further, there is a quality dimension to understanding the looming water availability problem.
The best quality water is exploited first and the subsequent exploitation is of lower quality
water. For this reason, as more water is required, increased investment in water treatment
becomes necessary.
Source: Ghassemi et al (1995)36
To cope with the fast growing population, technical solutions have been sought. Unfortunately,
some of these have created new problems for water resources for instance, irrigation to
increase the agricultural yield. it has the down side of contaminating water with agro‐chemical
pollutants beyond any other use even for further irrigation (Hunt 2004). By contaminating water,
irrigation use makes more water unusable. The fact that of the total current water withdrawals,
between 66 to 70 percent is committed to the use in irrigation (UN 2002, WWC 2005), indicates
the gravity of problems it is causing.
Another technical solution has been to buffer the lapses in water withdrawals in space and time;
by building dams and making inter‐basin transfers, to mention but a few. Much as inter‐basin
transfer avails water to the water scarce regions, it has the tendency to create new uses and
hence new demand for water in these regions. Inter‐basin transfer therefore compounds
pressure on water resources with increased abstraction.
Additionally, Kravcik et al (2002) in their ‘New Theory of Global warming’ revealed that activities
such as inter‐basin transfers, ground water mining, combined with urbanisation
4
lower the total
accumulation of water in the hydrological cycle and resultantly the amount of available
freshwater on the planet. They argue that the hydrological cycle is ceteris paribus in a stable
equilibrium. If a drop of water evaporates from a water body and falls back either to the water
body, forest, or porous ground it is reintroduced back into the cycle. If it falls elsewhere like in
the desert, or if this water is temporarily held somewhere in a reservoir, due to activity such as
of inter‐basin transfer, the cycle will continue drawing water from water bodies yet returning
less to them, hence precluding what Kravcik et al (2002) refer to as “The Right of domicile” of a
drop of water, resultantly reducing the amount of water available.
The implications of water scarcity are the losses of the benefits that water avails. Consider food
security; 40 percent of the world’s food is from irrigable land, which accounts for only 18
percent of total farm land (Postel 2002). In the face of the ever growing population, the water
scarcity/stress crisis threatens the sustainability of food production. Lessons from the Northern
plain region of China, that extends from the left bank of the Yellow River to the South of Beijing
                                                     
4
This comes with sealing of large expanses of land with concrete and tarmac, and deforestation   37
and Tianjin, give credence to the theory. Due to heavy irrigation in the region, water tables
receded over the years forcing the farmers to abandon the duo to mono‐cropping reducing food
production greatly and also causing serious environmental change (Shimada et al. 2000).
Other serious implications of water scarcity relate to ecosystems. There are many ecosystems
and species that depend on water. If water is reduced there are losses of either regulatory,
aesthetic, or bequest values that were originally availed by the water resources (WWC 2005),
and complications induced such as: organic matter pollution, saline intrusion or eutrophication
5
,
to mention but a few. The deterioration in quality and quantity of water can aggravate the
water‐borne diseases problem. Kirby (2005) argues that poverty would be perpetuated in the
LDCs by the scarcity crisis, since the very activities that are required to extricate these countries
from their position, like industrialisation and agriculture, need ample water supply. Others have
argued that water scarcity is bound to heighten political tensions between countries (WWC
2005, Postel 2002).
Suggested solutions to avert the looming crisis focus on the conservation of water and changing
the current wasteful exploitation habits. Postel (2002) argues that, the societal challenges like:
rigid mindset on the extraction of water, unsustainable water and environmental policies have
to be addressed if the situation is to be brought under control. Best usage methods have to be
sought in scientific investigations. For instance, irrigation should be based on how much
nutrition per unit of water we are getting in return and there should be trans‐boundary co‐
operation in management of water resources (Postel 2002).
However, not all scientists agree that the scarcity crisis is as grave as purported; There is the
argument that talk of a water scarcity crisis is a “more profession of the pessimist’s cynicism
than their expression of scientific argument” (Kirby 2004). There is the view that the global
                                                     
5
Eutrophication is a process whereby water bodies, such as lakes, estuaries, or slow‐moving streams receive excess nutrients
that stimulate excessive plant growth (algae, periphyton attached algae, and nuisance plants weeds). This enhanced plant
growth, often called an algal bloom, reduces dissolved oxygen in the water when dead plant material decomposes and can cause
other organisms to die (USGS 2005)38
population is not likely to exceed nine billion and therefore, with a helping hand from the
scientific solutions, the global system will be able to cope with whatever challenges relating to
water availability (ibid). In agreement, Kirby (2004) argues scientific solutions such as drip
irrigation and virtual water
6
trade will be helpful in averting the crisis‐alleged.
2.6. The Choice to Privatise water utilities or not to:
The privatization of water is always going to be a contentious issue  ‐ partly because allocations
under the privatization model allow for exclusion, which would be ethically unjustifiable for a
necessity like water, and partly because there is no evidence of the model actually being the
panacea. Proponents of the water privatization model argue that the water related goals
committed to by the heads of state in the 2000 United Nation’s MDGs need to be addressed  
with utmost efficiency, which the public sector is incapable of (Budds &McGranahan 2003).
Counter arguments are that privatization is purely driven by a profit motive and is therefore
bound to leave out the very poor (Bakker 2002).
The private/public water management debate is not a new one. In the Nineteenth Century,
water utilities were largely managed by private institutions in the world cities. In the middle of
the century, waterborne diseases broke out, most notably the 1850s cholera and typhoid
episodes (Bakker 2002). It signaled failure on the part of the private sector in the provision of
the service.   In the wake of the increasing acceptance of the germ theory and realization of the
externalities associated with safe water provision, the public sector took over management
(Bakker ibid). The situation under the public sector management, after less than two centuries,
has changed for the worse. With   one in six people lacking safe water to drink and waterborne
diseases killing a child every 15   seconds
7
(UN 2006),   the theory of state failure (rather than
market) is gaining credence.
                                                     
6
Virtual water is the amount of water that is embedded in food or other products needed for its production. For example, to
produce one kilogram of wheat we need about 1,000 liters of water, i.e. the virtual water of this kilogram of wheat is 1,000 liters.
For meat, we need about five to ten times more (WWC 2005)
7
Calculated from the UN 2002 water issues Fact sheet39
No more pronounced has this theory been than in Europe. Since 1980, the privatization model
has gained so much popularity that today reference is made of the ‘British model’ and the
‘French model’. The former model consists of privatising both assets and operation of the assets
while the later model allows for the assets to remain in publicly owned (Wikipedia 2007e).
Neither of the models was adopted globally until the 1990s; this was under the auspices of the
WMO. At the Rio De Janeiro conference on water in 1992, the commoditization of water was
emphasized (WMO 1992). This served to bring privatization to the forefront of international fora
in the intervening years. Many international institutions and governments realigned their
policies to promote increased Private Sector Participation (PSP) (Budds & McGranahan 2003).
Privatisation is normally defined as the shift of management and operation tasks from the public
sector to the private sector. However, such a simplistic definition offers a biased opinion to the
private public water debate. An extended definition which incorporates the various forms of
contracts that allow for partial involvement of the private sector (Budds &McGranahan 2003).
These contracts are such as: lease, affarmage, concessional and BOT  ‐  all of which leave the
public sector as the overseer of all operations and thus constitutes qualified privatisation. At the
extreme end is divestiture or the ‘British Model’ where the government transfers the entire
water business to the private sector through the sale of all or some of its shares to the highest
bidder in the private sector. Such a model has only been adopted by England and Wales (Budds
&McGranahan 2003).
The privatisation critics argue that while the privatisation proponents present efficiency as the
case for PSP, such an argument is not always going to be true for necessities like water (Budds &
McGranahan 2003). Bakker (2002) argues that efficiency requires effective regulation and the
correct incentives. Further, the lowering of costs of services is dependent on the cost of capital.
Therefore, the assertions of efficiency by the privatisation advocates ought to be tested against
the conditions prevailing in a place and at the time. Peculiar to all anti‐privatisation arguments
has been the ideology that water is a human right whose access should not be denied anyone.
The fact that privatisation is commercially driven implies that it does little to cater for
affordability and hence equity, especially where low income consumers are involved (Bakker
2002, Barlow 2001). Barlow (2001) notes that since the privatization of water in the 1990s, in40
France and Britain, water tariffs have increased by 150 and 106 percent respectively. The tariff
hikes in Britain, came with raised disconnections by 50 percent (ibid).
The operation of a full cost recovery business model requires private firms to be selective in the
siting of their operations. This selectiveness, commonly referred to as “cherry picking”,
guarantees the minimization of their risk. Budds & McGranahan (2003) highlight that it is for this
reason that Sub‐Saharan Africa has not attracted many private firms. Where PSP has taken off, it
has been short term in duration. The contracts drawn in US dollars to guard against the
fluctuating local currencies, even though this sets the water tariffs high for the average
consumer in these countries (ibid). Also, water utility services are lumped with other services
like electricity to make the franchise more profitable.
Further, talk of corruption and lack of transparency has not been lost on the water privatization
companies. The UN secretary General, Kofi Annan (2004) remarked that the problems of
collusive bidding for franchises, lack of transparency, inflexible contractual guarantees,
monopolization of essential infrastructure and lack of proper regulation by the governments
have led to the questioning of the merits forwarded by privatisation proponents in the water
sector (Deen 2004, Budds & McGranham 2003). The experience of the South that is in Africa,
Asia and Latin America, where numerous scandals in water privatization have been recorded,
indicate that privatization has been disappointing in many regions where it happened.
2.6.1 The Bottled water Industry experience
Until thus far, the discussion has dwelt on the privatization of the public water reticulation
works. However this has not been the only form of privatization in the water sector. The other
privatization experience in the water sector could be drawn from the bottled water industry
experience the world over. The companies managing water are some of the richest in the world.
For instance, in 2005 France’s Suez, and Germany’s  RWE ranked 79
th
and 78
th
respectively in the41
Fortune’s Global 100‐list
8
(Public Citizen 2005). These are the biggest water companies, and they
capture approximately 40 percent of the world’s water market share (ibid). Other major players
are Vevendi , Nestle, Pepsi, Coca‐Cola, Evian, Fiji waters, Biwater plc, Boouygues/Saur, US water,
Severn Trent and Anglian water. Estimates put the annual value of private water market at US $
1 trillion (Barlow 2001). Seen in light of the fact the private corporations serve only 5 percent of
the world population (ibid), the value serves to show the enormous profit potential these firms
possess. By contrast, estimates put the amount of total investment needed over 25 years (2000‐
2025) to in urban water supply and sanitation at US $ 1.8 trillion (WSSCC 2000).    The irony
presented by Clayton (ibid) and Gleick (2004) is that much of the bottled water sales growth
rates from the late 1990s originate from the developing countries. These countries, that have for
long been financially constrained to put in place improved water quality systems, are opting for
a far more expensive product. Moreover the bottled water product is not totally free from the
flaws that have plagued the public water services.  
A four year scientific studyon various brands of bottled water by the Natural Resources Defense
Council (USA) deduced that one third of the brands tested registered presence of Arsenic and
carcinogenic compounds. These findings were in line with those of an earlier survey at Ohio
State University which revealed that a mere 39 out of 57 bottled water samples to have had
purer water than the tap water and 15 of the samples had significantly high bacteria samples
(Klessig ibid) .
The evidence from other parts of the world has been equally discrediting of the bottled water
industry. The Bureau of Indian Standards, for instance, shut down 200 bottling plants in New
Delhi, among which were the Coca Cola’s Kinly and King fisher, over unsatisfactory quality
performance (ibid). In the Klaten district of Indonesia, concerns by the local farmers over a
lowered water table in the region by PT Tirta Investma (TI), producer of bottled water brand
AQUA‐Danone, were raised (Yahya 2004). The company in harvesting most of the ground water
penalized the farmers with higher costs of farming, since they had to hire pumps and other
superior technologies to avail irrigation waters.
                                                     
8
This is a list of the companies with the largest revenue and net turn over42
The commitment by the heads of states in the UN’s MDGs in 2000 to halve the number of
people without safe drinking water by the next two decades is going to need serious rethinking
of the management of water. The increasing role of the private sector in water management is
one such step in rethinking the management structure. Concerns over equity, often raised
where the private sector is concerned, should be looked at in light of the fact that the resource
needs to be managed as efficiently as possible. The current debate is useful, in the sense that it
highlights the flaws of both private sector and public sector institutions in their management
abilities. The current arguments viewed in light of the 1950s’ experience, reveal that both the
private and public sectors left alone are prone to failure. In the LDCs, where the biggest
challenges of water remain amid high poverty levels, it is important that the public sector plays
the leading role in the provision of water and regulation. If the private sector is providing the
water government should subsidise those that cannot meet water tariffs. This will ensure more
equitable access for the general population.
2.7. Conclusion:
When we review the literature of the last three decades we find that there are three important
contributions made to this research by the body of knowledge that exists. The main
contributions made are:
Health is an important factor of production whose stock depends on the
inputs the consumer invests in its growth.
The current water quality issues in water can be understood by
investigating the broader issue of the decreasing quantity of water on
Earth.
In the face of current failures of the water sector mainly managed by
public sector, the private sector may have a role to play in alleviating the
problems.
Regarding health as a factor of production there are two main ideologies namely:
The Mercantilists’ view
The Statists’ view43
The main theme of the mercantilists view, in the Grossman (1972) model, is that health is
produced by individuals in their leisure time and through investment in more health goods. The
more the investment in health goods the higher the stock of health the individual enjoys. The
Statists’ view by Lofgren & Johansson (1995) asserts that viewing health in a purely market
economy is unrealistic as the market economy cannot fully appropriate the benefits, owing to
the externalities associated therein. They (ibid) suggest that the public sector be involved to
subsidise health in order for the full benefits to be realized, as this would ensure its most
optimal allocation.
Regarding the growing concern of the decreasing quantity of water on Earth there are also two
schools of thought namely:
The pessimists
The Optimists
The main ideas of the pessimists are that: water in the global hydrological system is near fixed
supply. Water resources are renewed by the water cycle, but this process too generates a near
constant amount of water. Moreover, only meager quantities of this (not more than 30%) is
withdrawn annually (Postel 2002). Faced with this stringent supply of the resource, on the
demand side, the global population is growing dangerously fast. Technology has allowed for the
extravagant use of water resources. Strain on the water resources can originate from beyond
borders by means of such as inter‐basin transfers. All these factors have led to more water being
used than is being restored. Some of the current problems in the water sector like increases in
water diseases, are early signs of a looming scarcity crisis. If allowed to continue, the current
situation threatens an aggravated water scarcity crisis in most parts of the world in the next few
decades, which could have more serious repercussions for economic growth. The LDCs are more
likely to bare the brunt of the crisis and so intervention there is more needed now, to avert the
crisis. The intervention required is in improved investment in treatment.
The optimists, on the other hand, argue that the world population will not escalate to
unmanageable levels. They argue that the current problem is uneven distribution, rather than
the available quantity. Technical solutions can be employed to bridge the gaps and find more44
efficient ways to use water. They therefore argue that with human ingenuity the global system
will cope through the crisis the pessimists allege, if there is one.
Finally, concerning the private public water management there are also two competing
ideologies namely:
The privatisation proponents
The privatisation opponents
The proponents argue that the commitment to avail water for all by the next two decades
requires efficiency. The current situation that has been highlighted by the prevalence of
waterborne diseases. It suggests that the public sector has failed and that provision can be
improved through (selective) privatisation.
The opponents of this view, on the other hand, argue that the while privatization is associated
with efficiency, this argument does not hold for goods like water. Water has to be managed in
such a way that recognizes access to it as a human right hence, it must be equitably distributed.
Mainly for this reason the private sector is not entirely suited for water management. In
addition, there is no supporting evidence to the much acclaimed merits of privatization in LICs so
far where it has been undertaken.45
CHAPTER THREE:
UGANDA WATER46
3.0. Introduction
To understand the Uganda’s drinking water quality problems, a thorough analysis should
investigate the state of the raw water sources and the processes involved in the abstraction of
this water, its treatment and its delivery to final user. The legislative structure under which the
water systems are managed determines the allocation and protection of these resources. An
investigation of legislative structure is imperative too. This chapter is an attempt to investigate
these issues. First, the freshwater resources’ endowments of the country are assessed, and
second, major water withdrawals in the country are described. The management structure of
Uganda water is investigated and finally the water treatment process for the city of Kampala is
discussed.  
Uganda is a land locked country laying on the equator in central Africa. The country has a surface
area of 236,040 km
2
with a population approximated at 27 million (Wikipedia 2007d). The
country has 18 percent of its land area covered by water (AQUASTAT 2005, Ndege 1996).
Estimates put the volume of renewable freshwater in the country is 66km
3
(AQUASTAT 2005). By
implication, per capita availability of water is 2800m
3
per annum. In reality however, the
freshwater resources are confined to specific regions of the country and not others. Also, the
micro‐climatic regimes vary across the regions, with the result meaning that only specific region
experience heavy rainfall. In light of the country’s poor (water) infrastructural development
indicate that only part of the total volumes of water available is accessible. To this problem, is
the additional one of  which increases the costs of treating water to make it safe to consume
The country has two major sources of water namely: surface water resources, ground water
resources. The surface water sources encompass all the open water bodies in the country. The
ground water sources define all water beneath the surface, existing in soil pore spaces.
3.1. The Uganda freshwater resources
3.1.1. Open water sources
Precipitation (rainfall) in Uganda has been estimated to vary from averaging at 750mm‐1500mm
or 700mm‐2000mm per annum (AQUASTAT 2005, NEMA 2000). Uganda has an equatorial47
climate characterized by two rainfall maximums March to May and September to November.
These maximums are influenced by the double passing of the inter‐tropical convergence zone
(ITCZ) over the equator (and in turn the country), this is a low pressure belt formed by vertical
accent of warm moist air from the latitudes above and below the equator (Wikipedia 2007b). As
the convective activity of thunderstorms conveys the ITCZ between the latitudes, the rainfall
distribution of the regions it passes are affected. As a result the tropical regions experience only
wet and dry seasons (ibid). Uganda lying on the equator implies that the ITCZ passes the country
twice with the duo rainfall maxima. The Lake Victoria basin, under which Kampala lies, has the
highest rate of runoff of greater than 10liters/s/ km
2
(UN‐HABITAT 2007).
The open water sources – which are the most important sources of water in the country‐ draw
heavily from the precipitation; this ensures their renewal.
Figure 3.1: Map of Uganda showing the major drainage water sub‐basins
Source: WWAP (2006)48
Due to the physical processes of warping and faulting, Uganda’s geographical outline curves out
a depression at the center and an elongated trough demarcated by fault lines along the fringes
of the country called the rift valley. The two features are important in distribution of the
country’s water resources (open water sources).
The depression forms the Nile Basin and the rift valley the rift valley basin. The Nile basin covers
98 percent of the country’s land area, the upper part of the White Nile basin to be specific. The
basin is made of a system of interconnected equatorial lakes and rivers (AQUASTAT 2005). The
lakes and rivers are marked by surroundings of marshes, fen or peatland hydrological features,
generally referred to as Wetlands. Wetlands cover 24000km
2
of the country’s total area, one
third of which are permanently flooded (AQUASTAT ibid). They are a source of water for the
crafters and herdsmen and help in the purification of the water entering the lakes and rivers
(WWAP 2006). The rift valley basin is noted mainly for its lakes ‐ George, Edward, Albert and the
section of the Nile draining out of Lake Albert; the Albert Nile. Combined, the surface water
resources account for a volume close to 39km
3
of water available in the country per annum
(AQUASTAT 2005). Hydrologists define eight sub‐basins in the country; the basins are defined
according to the largest water body in the region serving the users. These are: L. Victoria, L.
Kyoga, R. Kafu, Lakes George and Edward, L. Albert, R. Aswa, Albert Nile and Kidepo Valley
basins as can be seen in the figure 3.1 above.  
Lake Victoria perhaps, is the most important open water source in the country. The lake is the
second largest in the world. It serves the country’s largest cities in terms of population and
industrial activity, that is: Kampala, Jinja and Entebbe. The lake is also the source of the longest
river in the world, River Nile. Being the longest the river has enormous hydro electric power
generation potential. Water uses are categorized into two broad categories: consumptive and
non‐consumptive uses. The former involves abstraction of water from the resource, using some
of it and returning less than was initially abstracted. The non‐consumptive uses on the other
hand involve those uses that do not remove the water from the source. Uganda water
withdrawals currently are at approximately 300 million m
3
per annum. Domestic water uses rank49
highest in importance with their withdrawal levels accounting for a whole 44.7 percent of total
withdrawals (AQUASTAT 2005) (also see figure 3.2 below). Agricultural uses claim 40 percent of
these withdrawals, where as Industrial uses account for the lowest, namely 15 percent (ibid).
Both the consumptive and non‐consumptive uses have equally contributed to the straining of
the Lake Victoria, and contributed to the deteriorating water quality levels
Figure 3.2: Uganda Water Withdrawals (consumptive uses)
Source: AQUASTAT (2005)
Increased demand for water, rooted in increasing population growth in the country has meant
increased withdrawal rates. However, in the riparian region of the lake Victoria Klohn & Andjelic
(1997) note that the population growth rates were above the national rate of 2.8 percent per
annum. In Bukoba, for instance, the growth rate is 4 percent (UN‐HABITAT 2007).  By implication
the volume of water available per capita is being reduced.
The city of Kampala accounts for 70 percent of reticulated water consumption (NWSC 2005). The
consumption has been increasing, necessitating construction of a third water treatment plant
Gaba III. The plant is intended to increase water supply by 80,000 m
3
and 180,000m
3
per day in
wet and dry seasons respectively, over the current volume of supplied 150,000 m
3
per day
(Tumwebaze 2006). Reticulated water caters for only 51 percent of the demand of the Ugandan
population (Ndege 1996).  50
The non‐reticulated water withdrawals in the country are hard to monitor. However, considering
the rate growth of riparian region’s population,   one can safely deduce that the non‐reticulated
water withdrawals have also increased significantly.   Consequently, the availability of water per
capita in the country has gradually been decreasing (ibid). This situation has not been helped by
the deterioration of the quantity and quality of Lake Victoria’s water which has raised the cost of
providing water and increased the risk of disease to the users. Both natural (global climatic
conditions) and human activities have interplayed to undermine the quality of the lake’s water.
Figure 3.3: Aerial shot of the Lake Victoria
Source: Wikipedia (2007c)51
In the last decade, the Lake Victoria region experienced the hottest seasons and the most
droughts in 50 years (Ddumba 2006). The geology of the lake makes it very responsive
(vulnerable) to climatic changes. Among the factors that determine the level of vulnerability
of a water body, three are most critical: the size of store (water body), the rate water inflow
and the evapotranspiration rate; (WatBal model in Hassan 2006). Lake Victoria is
disadvantaged on all three fronts: the lake is shallow, has limited river inflow and has a large
surface area relative to its volume (Wikipedia 2007c). Changes in the climatic conditions,
such as the currently rising of global temperatures have negatively impacted on the lake.
However, irresponsible use of the lake on the part of the Ugandan users has aggravated the
negative impacts of the climate, which has seen the level of the lake drop to its lowest in 80
years (Mugabe &Kisambira 2006, Pearce 2006, Wikipedia 2007c, Reynolds 2005). This drop
was assisted by the construction of parallel hydro‐electric generating plants on the lake that
have drawn more water out of it than is supposed to flow out of it (ibid).
At the time of the construction of the Owen Falls power dam, the first dam on the lake, in 1952,
the ‘agreed curve’
9
   was estimated to mimic the flow rate without the dam (natural flow rate).
This estimated flow was between 300–1700 cubic meters of water per second (Pearce 2006,
Wikipedia 2007c). The demand for more HEP prompted the construction of a second dam on the
lake. In 2002, the construction of this dam came to completion. This event coincided with the
first significant lake’s head loss (ibid).   A look at the hydrological data revealed that with the
second dam, the rate of outflow from the lake was averaging between 1 250–55 000 cubic
meters per second. This rate is 55 percentage points above the natural rate of outflow   (Pearce
2006). Since 2003, the lake has lost 75 cubic kilometers of its water, approximately 3 percent of
its volume (ibid). A DWD report on the matter resolved that, the dams accounted for 0.47m
head loss, which is 60 percent of the lake’s total head loss, and the droughts account for 0.31m
head loss which is 40 percent of the total head loss (Mugabe &Kisambira 2006).
The immediate impact has been lower HEP generation at both dams. The machines at the two
dams are operating under stress due to lack of adequate water yielding a total of 195MW
                                                     
9
Formula used to calculate the rate of flow of water  52
relative to their capacity of 380MW (Pearce 2006). Uganda experiences very frequent power
outages daily because of this lower power generated in the face of a heightened demand that
had earlier necessitated the building of a second HEP plant. There is another problem that has
resulted from this reduced water volumes in the lake, one of lower quality levels of the lake’s
water.
Earlier studies had warned that the increased human activity around Lake Victoria and
urbanization had increased the rate of point and non‐point pollution and that the lake’s quality
levels were deteriorating (Klohn &Andjelic 1997). Since the early 2000s, significant deterioration
in the lake’s water quality has been observed (Oyoo 2006, Tumwebaze 2006, Wikipedia 2007c).
One probable explanation, according to Winter (2005), would be that the decreased volume of
the water has increased the concentration of the pathogens in this water.
The most vivid visual evidence of this quality problem is the ever growing algal infestation on the
lake.    According to the senior laboratory technician at Gaba II, Tumwebaze (2006), the algal
count is so high at the Murchison bay where Kampala abstracts her water, that the colour of the
water is tainted green.
Figure 3.4:  Water treatment cost (in shillings) per cubic meter at the Gaba II
Source: Tumwebaze (2006)53
The algae may not be as harmful in itself but, the fact that it can rot presents a risk for the users.
Further evidence of the quality problems are reflected in the ever increasing cost of treatment
of the Lake Victoria water, as can be seen in the figure 3.4 above. It should be noted that the
sharp decline in the treatment cost in the period of 2003 does not signify an improvement in
water quality.
Table 3.1: Raw Water Quality in terms of Turbidity and colour
Year Turbidity, NTU
10
Colour, PTU
11
1992 2.5 28
1993 2.6 30
1994 2.8 40
1995 3.1 46
1996 3 55
1997 3.3 55
1998 3.5 58
1999 4.5 70
2000 4.6 72
2001 6.3 120
2002 5.5 110
2003 9.9 149
2004 9.6 153
Source: Tumwebaze (2006)
                                                     
10
Nephelometric Turbidity Unit
11
True Colour Unit54
As Tumwebaze (2006) explains, often the risks of having inadequate supply of water far
outweigh the risks of having poor quality water. In the 2003 period above, significant quality
deterioration prompted treatment by‐pass in order to meet the high demand. This period
coincided with coming into operation of the second HEP plant on Lake Victoria.
The other parameters of water quality such as turbidity and color have been increasing too as
can be drawn from the table 3.1 above.
3.1.2. Ground water sources
Ground water is subsurface water that accumulates in voids and permeable geological
formations (Kulabako 2005).  The existence of ground water is influenced by the underlying rock
structure and how permeable it is. Most of Uganda (90 percent) is underlain by crystalline
basement complex rocks that are very impermeable. These are rocks such as granites, granitoid
gneisses and gneisses and in some cases migmatised (UN‐HABITAT 2007). This water can only be
found in places with weathered rock beds. Kampala has a hilly topography, characterized by
differentiated thin weathered meta‐sedimentary bedrock types. These hold ground water for a
short residence time discharging this water to springs that are at the bottom of the hills
(ARGOSS 2002). This water, according to Kulabako (2005), supports the peri‐urban settlements
that harbor  60 percent of Kampala’s population.
The ground water, especially around the more urbanized settings of the country, is not potable
as found Kulabako et al (2004). Their survey (ibid) was conducted in Kampala’s peri‐urban
developments. Their work found that Kampala’s water table sits at just between 1‐2 meters
below the surface which makes ground water very prone to pollution in the city. During the
rainy seasons, the risk of the pollution rate more than doubles. Averaging the impurities’ level
over the seasons revealed that there was a   high concentration of coliforms, from 1‐16 x 107
cfu/100ml, and nitrate levels, 0.10‐779mg/l    ((Kulabako et al. 2004), compared to WHO ’s
minimum guideline of 0 cfu/100ml and 50mg/l respectively. These findings are in line with those
of Haruna et al. (2005), who sampled some of Kampala’s protected springs. The bacteriological
quality of the springs and protected waters is poor.55
The chemical parameters for the ground water have also been found to be posing a risk to the
users, if the water were used without treatment, in Kampala. The phosphorous levels in the
ground water averaged at 0.001‐13 mg/l (Kulabako et al. 2004). The chemical has no stipulated
allowable limit according the WHO (2004) guidelines. However, Lehtola (2002) explains that the
chemical, should help in reducing microbial growth in water; hence acting as a purifier. The
nitrate concentration in the water, at 0.10‐779 mg/l, was way above the WHO’s (2004) allowable
limit of  50 mg/l. The ground water should therefore be treated before use.
Figure 3.5:  Typical Kampala Drainage channel
Photo by Wasswa at Nateete
The major source of contamination of the ground water Kulabako et al (op. cit)  identified, were
the numerous pit latrines in the area and the huge amounts of solid waste carelessly disposed.
In the rainy season, the coliforms and chemical impurities percolate though the thin table and
contaminate the ground water (Soutter 2005). The drainage infrastructure in the peri‐urban56
settings is ill‐equipped prevent contamination of the ground water. The figure 3.5 above shows
ill maintained infrastructure of the city that compounds pollution of the open and ground water
sources.
The ground water sources are recharged at a rate of 90 and 220 mm per annum which is about 7
and 20% of the average annual precipitation in Uganda (UN‐HABITAT 2007).
3.2. The Uganda Water management structure
The Uganda water management has, since the mid 1990s, undergone several transformations in
terms of its legislative and the institutional structuring. The discussion that follows provides a
background of the earlier management structure, pointing out the factors that prompted the
transition to the new management. The new management structure is then presented and new
policy and institutional developments described.
3.2.1. The Uganda water sector legal Framework  
Before the 1990s, legislation in the water sector was inadequate (Uganda 1999). After her
independence, Uganda adopted wholly the colonial policies governing environmental resources,
where the government had the leading role (NEMA 1996). Most of the laws were suited for very
specific resources, with little emphasis put on the interrelationships between these. There was
little indication that these laws catered for the environmental economic aspects of these
resources. On top of this, these laws had a week constitutional backing and hence, compliance
was poor (ibid).
The reforms of the 1990s came when the need to manage water in a more sustainable way was
envisaged. This vision was born at the International Conference on Water and the Environment
(ICWE) in January 1992 in Dublin. At the conference, experts contended that the state of water
resources was critical (WMO 1992). Water scarcity was seen as a problem that would have
serious implications on world economies in the near future unless positive action was taken.
Recommendations on such action were summarized in the Dublin Statement and the conference57
report, adopted at the end of the conference.    Such action needed political commitment and
involvement from the highest levels of government to the smallest communities (WMO 1992). In
this vein world leaders convened, in Rio de Janeiro, In June of 1992, at the United Nations
Conference on Environment and Development (UNCED). World leaders were urged to translate
recommendations of the Dublin statement into urgent action programs for water and
sustainable development (WMO ibid).
The main objective was to reduce over consumption, pollution and drought and flood threats.
This action was based on four guiding principles namely:
• “Freshwater is a finite and vulnerable resource, essential to sustain life development and
the environment.
• Water development and management should be based on a participatory approach,
involving users, planners and policy makers at all levels.
• Women play a central part in the provision, management and safeguarding of water.
• Water has an economic value in all its competing uses and should be recognized as an
economic good” (WMO 1992).
The principles were echoed in the Agenda 21’s chapter 18 on freshwater resources (Uganda
1999), and has influenced most countries’ to manage the water resources more responsibly. In
Uganda, The Uganda Water Action Plan (WAP) of 1995 was the first of such attempts. It
recognised water as an economic good with an economic value (Uganda 1999). The plan has
been very influential in shaping subsequent policies and legislation in the country’s Water
Sector.
When poverty was prioritized in the Poverty Eradication Action Plan (PEAP) of 1997, revised in
2000 and 2004, water was recognized as an intervention to improve the people’s    wellbeing
(ADF 2005). Most policies, in water legislation specifically, were reformed to cater for this new
plan of poverty eradication. In 1999, the National Water Policy (NWP) was enacted, a legislation
consolidating all earlier legislations in the country’s water sector. The NWP (1999) was drafted
to be the new integrated approach of managing water resources in ways that were sustainable58
and beneficial to Ugandans, yet recognizing the resources’ social value as well as their economic
value (Uganda 1999).
The NWP (1999) amalgamated all the water policies that existed pre‐1999, namely:
• The Uganda constitution (1995): this recognises the access to safe and clean water as a
right to all the Ugandans. The constitution also vests the duty of management and
conservation of water resources at all levels to government in objectives XIV, XXI and XXVII
(Uganda 1999).
• The Local Government Act (1997): the act stipulates that management of water should
be decentralized. It therefore transfers power from the central government to the local
governments which encourages a more participatory process in decision making and which
tailors the outcomes to local needs (ibid).
• The National Environment Management Policy (1994) and The National Environment
Management Statute (1995): these advocate for water resources to be managed in sustainable
manner that allows for provision of water of acceptable quality for all social and economic
needs (Uganda 1999).
• The Water Statute (1995): this is a stipulation of the mode of operation of the Water and
Sewerage authorities in the country. It assigns the authorities’’ the duty availing water in
ample quantity and good quality for domestic use. Other uses should be catered for while
precluding these uses’ ability to degrade the water resources at large (ibid).
• The National Water and Sewerage Cooperation Statute (1995): this lays out the
operations code for the supplier of water and sewerage services of the NWSC.
• The Ugandan Plan of Action for Children (UNPAC) 1992: this policy document is geared
toward ensuring the provision of basic social services, among which are clean water and basic
sanitation, to as many Ugandans as possible, paying special attention to the needs of children
and women (Uganda 1999).
The NWP of 1999 by and large defines the legal stipulations governing the use of water in the
country. Other legislations that are currently consulted in the sector to complement the NWP
(1999) include the Water Act (1995), the National Gender Policy (1997), The Health Sector59
Strategic Plan (2000) the Kampala Declaration on Sanitation (1997), the Children Statute (1999),
Uganda National Bureau of Standard’s Specification for Drinking (Potable) water and The
National Health Policy (1999).
The Uganda water legislation assigns the roles of allocation and management of the country’s
water resources to the state. Allocation of water is regulated by the administering of permits
which specify the type of uses to be regulated and the abstraction fees charged. The
administering of permits is by the MWEL’s DWD. However this role can be delegated to the
districts after they have assessed the impacts of abstraction in their jurisdictions (Uganda 1999).
The legislation protects the water quality in the country both directly and indirectly. Directly,
maximum allowable concentrations of physical, biological and chemical parameters for the
country’s water are laid out in the county’s water legislation such as the National Environmental
Statute (1995). Such parameters for reticulated water are stipulated in the UNBS’s Standard
Specification for Drinking (Potable) Water (1994). Indirectly, quality is protected by maximum
concentration of the physical, biological and chemical parameters for waste water, sewerage,
and industrial effluent, as this can end up in the open and ground water sources. Such
parameters are stipulated in National Environment Statute (1995) sections 27 and 108.
One drawback of the legislation seems to be that, although all Uganda’s surface water sources
are a part of the transboundary water shed (Uganda 1999) , so far there isn’t a transboundary
legislation in place to govern their use. This absence stifles any efforts to conserve natural
resources quality in the country because the upstream users (all countries in the southern
hemisphere) often influence downstream water quality. The problems currently being
experienced on Lake Victoria are not of Uganda’s making alone. Other users like Tanzania and
Kenya have also had a part in degrading the lake. For instance, population growth in Bukoba
(Tanzania) was reported to be 4 percent (UN‐HABITAT 2007), i.e. also high.
3.2.2. Uganda Water Legislation on Economic Appraisal  
One of the major flaws for public utilities’ provision in the developing countries has been lack of
appraisal ex‐post; it is mainly for this reason that most of them have remained inefficient in60
these countries (Wall 2001). Efforts to avoid such shortfalls should therefore be made before
undertaking these utilities, if the best performance of these is to be achieved.
The NWP (1999), having the WAP (1995) as one of its cornerstones (see section 3.2.1),
recognises water as an economic good. The NWP (1999) Chapter 5 stipulates that reticulated
water should be provided in a way that ensures financial viability of the public utilities used in its
provision. It goes on to say that the provision of water should be governed by a demand driven
approach, in which the users fully contribute to the cost of facilities and services to promote
ownership and sustainability (Uganda 1999).
The Water act Section 54 (1) stipulates that “…an authority may recover the costs or any part of
the costs of acquiring or constructing any works after considering the benefits from the works
and any contribution that the authority may assess to be contribution from the owner of the
land.”  Implicitly the Uganda water legislation allows for economic appraisal to be conducted on
public utility (water) projects. One such project has been the Appraisal of the Rural Water
Supply and Sanitation Program of 2005. However, the evidence of such a process being
followed in the past when undertaking public projects, specifically in water, is lacking.
On the whole the Uganda water legislation is supportive of quality preservation of the natural
resources and quality of reticulated water.   61
Figure 3.6: Structure of Ministry of Water Lands and Environment (MWLE)
Source: MWLE (2004)62
3.2.3. The Uganda Water Institutional Framework:
According to the section 5 of The Uganda Water Act (1995), duties of managing water are
vested in the government; under the Ministry of Water Lands and Environment (MWLE). The
ministry’s structure is laid out in the figure 3.6 above. The leading government agency for the
water and sanitation sector under the ministry is the Directorate of Water Development
(DWD). The department has three sub‐divisions:
• “The rural water supply Department (RWSD): this subdivision is responsible for service
delivery in the rural area. It manages protected springs, bore holes shallow wells, valley
dams and in some areas piped water to ensure adequate provision for the local uses.
• Urban Water and Sewerage Department: this is the subdivision responsible developing
water and sewerage facilities for urban areas. Also this division that oversees all the
regulation and quality assurance.
• The Water resources Management Department: This ensures the sustainable
management of the water resources” (DWD 2007).
The task of water and sewerage services supply is by an agency under the office permanent
secretary called NWSC. The NWSC operates in 15 districts and in the financial year 2004/05
alone sold a volume of 38.2 million cubic meters (NWSC 2004). Kampala is the country’s
largest water user, consuming approximately 70 percent of the total water supplied by the
NWSC (ibid). This water is produced at the Gaba I and II treatment plants in Kampala.
3.2.4. Kampala Water Supply
Kampala is divided into 11 Water distribution branches namely: Branch 1 City Center, Branch 2
Kansanga, branch 3 Nnajjanankumbi, branch 4 Bwaise, branch 5 Ntinda branch 6 Kitintale.
Branch 7 is an abstract branch reserved for the big consumers like industrial users, branch 8 is
Nateete, branch 9 Nansana, branch 10 Mukono and branch 11 Kajjansi (Akol 2006). The
branches are serviced by two treatment plants, Gaba I and Gaba II. Gaba I was commissioned63
in 1930 to cater for a then less crowded Kampala city, and currently produces 15,000m
3
per
day (Tumwebaze 2006). Gaba II was commissioned in 1992 to meet the rising demand of water
for the city. The plant produces a volume of 80,000m
3
per day (Ddumba 2006, Tumwebaze
2006).    Currently construction of a third plant Gaba III is underway to meet the growing
demand for water in the country (ibid).
The water however, is not potable, as highlighted earlier. The water authorities at the NWCS
maintain that the corporation produces potable water and deterioration occurs at the
consumer level (Oyo 2006, Oburu 2006). Oyo (2006) explains that storage tanks have to
regularly be cleaned, which is not done by most Ugandan consumers and this leads to growth
of disease causing germs in the tanks. The users therefore create their own problem of
infection. However, according to the survey conducted in the country in 2006 March involving
204 randomly selected respondents in the city, the majority of the users contested the water
authorities’ view. When asked to grade the quality of the water in terms of microbial safety,
colour taste and odour either as excellent very good, good, fair, poor or very poor, a majority
of 92 percent of the respondents contended that the water needed treatment prior to
drinking. A more elaborate discussion of the results is given in the chapter six.
In agreement with survey results, (To‐Opoya 2006) of Century Bottling Company limited, which
manufactures Coca‐Cola products in the country, notes that the NWSC’s water is of poor
quality, especially from microbiological aspect. It is standard for beverage plants to have a
treatment facility on site to ensure use of high quality water in production. In the Ugandan
case, To‐Opoya (2006) explains that the percentage cost of treating water as an aggregate of
the total production costs of producing the beverages at the plant, is high relative to other
East African countries and Mozambique.
3.3. Kampala Water treatment: The Gaba II plant
Gaba II is one of two plants currently supplying the city with water; the plant was chosen for
the study as it has the latest technology in the country for water treatment.  64
Raw water for the Kampala city is abstracted from Lake Victoria at the Murchison bay. The
transmission of water from the source to the treatment plant can be done by use of pumps or
gravitational flow, depending either distance between source and plant topology. Gaba II uses
gravitational flow. Water in its raw state contains impurities from heavy particles to the finest
microorganisms that can cause disease. The treatment process involves both chemical and
physical processes. It isolates these and removes them to make the water safe for
consumption. It is these processes, this study next turns to. The description of the processes
was done based on an interview with the shift overseer, Mugume (2006)  
On its way into the sump, water from the lake is screened for heavy particles that could
otherwise damage the impellers. The screening is done by two fine screens and a coarse
screen. Water in the sump, now free of heavy impurities is conveyed to the mixing chamber by
3 pumps for the second stage which is the first chemical treatment (Mugume 2006).
In the mixing chamber, the chemical process of coagulation and flocculation are initiated to
enable the isolation fine impurities in the water. Doses of Aluminum Sulphate (AlSo4) are
introduced to the water in the chamber. Special pumps called positive displacement pumps
send the same amount of AlSo4 solution every transmission from the chemical house to
enable a proper reaction, as this is per given volume of water (ibid). The dose rate is 50mg/l.
Gaba II on average uses 7.2 tonnes of AlSo4 daily. A kilo of the chemical costs Shillings 803/=
(Tumwebaze 2006). The reaction of the chemical with the water causes the fine impurities in
the water to coagulate forming lumps of various sizes (Mugume 2006). These lumps can then
be separated physically from the water. This separation only happens in the clarification
process, which is the next stage of treatment.
Water is conveyed to the clarifier via an inlet channel, such as shown in the figure 3.7 below.
Clarification is a physical process that entails the separation of finer impurities from the water.
The end product should be colourless yet odourless water. It is referred to as Supernatant
Water (Mugume 2006). The water from the mixing chamber (the previous treatment stage) is
conveyed to the clarifiers by an inlet channel lined with special pipes standing vertically, called65
tridents. The tridents convey the floc‐laden water into the clarifier from its bottom and the
separation of the impurities occurs as the water rises in the clarifier (ibid).
Figure 3.7: The clarifier and inlet channel (Gaba II)  
Photos by Wasswa
The heaviest flocs settle at the bottom of the clarifier forming a mixture referred to as sludge.
This is removed form the clarifier via the dislodging valve, routed to a nearby swamp that does
the natural purification before it gets to the lake (Victoria). The lighter flocs remain suspended
in the clarifier as the water rises forming a layer relatively closely compacted flocs referred to
as the floc blanket. The floc blanket helps in the purification process in its ability to retain
sizable flocs as the water passes through it (Mugume 2006).
Special pockets lining the walls of the clarifier run horizontally to collect the excess floc blanket
as it rises. The finest particles that could not be retained by the floc blanket are often very light
and so float on the water in the clarifier. This is skimmed off as scum. The top of the clarifier is
lined with special pipes called collets. The collets deliver the water into the inlet channel which
conveys it to the next stage of treatment, which is filtration.
Scum
Inlet channel with Supernatant Water
Clarifier  66
Figure 3.8: The Filter bed (Gaba II)  
Photos by Wasswa
Filtration is a physical process that further purifies the water. At the Gaba II plant this is done
by rapid gravity sand filtration (Mugume 2006). The sand media is laid in a three strata
arrangement. The top‐most stratum is made of the finest sand, the mid‐stratum made
relatively larger sand called coarse sand and the bottom stratum made of gravels (ibid).  As the
water percolates through the sand media, any impurities that might still be left in the water
are isolated by the filtration process, leaving clear water. At the bottom of the filter bed are
nozzles that deliver water to the filtrate channel which conveys water to the next treatment
stage of chlorination. However, before we describe the chlorination process we need to look
at the backwash process, a process that determines the efficiency of the filtration process.
The sand media in its functioning could get impaired. This could be due to too much dirt
clinging to the sand particles hence clogging it to impermeability. Also the weight of the water
could compact the sand media restricting percolation. Often, this is identified by a rise in the
water head in the filter bed (Mugume 2006). Rectifying this impairment is the work of the
backwash process.
For the backwash process to begin the inlet channel is closed to cut of any supply of water into
the filter bed. Next a valve called the hydraulic valve (see figure 3.8 above) is powered which67
keeps the filtrate channel open for the water head to recede. On attainment of the desired
head, the valve is de‐energised now to cut off any supply from the filter bed to the filter
channel (Mugume 2006). Gaba II has three filter beds only one bed can be cleaned at a time.
Pressurised air is blown into the filter bed from the bottom which is known as air scouring. The
air scouring process stirs up the water in the bed as well as displacing the sand media. In the
process, sand particles rub against each other detaching the dirt (this is a 3 to 4 minute
process). The up‐wash process aerates the sand media allowing for easier permeability. Water
from the service tank is then introduced with pressure from the bottom of the bed to displace
the dirty mixture from the backwash process. The filter bed has a section closed off by a weir
in which the water pours. This stands along the bed with a height rising three quarter way to
the top.    The dirty water is pushed out when the filter bed‐water head is midway. The
pressurised water is washed into the tank until the water clears. Then the water head is
allowed to rise by closing of the outlet on the other side of the weir. This process takes close to
2 hours (Mugume 2006).
Chlorination is a disinfection process that uses Chlorine (Cl2) chemical to kill bacteria and
disease causing organisms. The dosage of chlorine depends on the quality of the water but its
dosage rate is 2 ‐ 4mg per liter. The chemical is administered as 99 percent gas into the water
by the facility shown in figure 3.9 below. Gaba II uses an average 550Kg of Chlorine gas daily
(Tumwebaze 2006). Uganda’s current source of chlorine is South Africa. A kilo of the gas being
Uganda Shillings 3500/= (ibid).   The gas is mixed with water from the service tank; water only
acting as a conduit for the gas and connected to the filtrate cannel. As the water mixes a
chemical reaction that forms acidic compounds that attack and destroy the enzyme of the
bacteria (Mugume 2006).68
Figure 3.9: The chlorination room
Photos by Wasswa
During the chemical processes a lot of acidic compound are formed by the reaction processes.
If left in the water, the acid could corrode pipes and concrete and could be harmful to the
water users.
Figure 3.10: Chlorine storage room and Treated water tank
Photos by Wasswa
Chlorine Storage
Tanks  Neutralisation process69
So the last treatment on the water is neturalisation, by use of Sodium Carbonate (Na2Co3).
The dose rate of the chemical is 26mg per liter of water and Gaba II uses a daily average of
4800Kg of the chemical.  A kilo costs Uganda Shillings 707/= (Tumwebaze ibid).
The water is then drawn from the treated water tank to be pumped to the main Reservoir at
Muyenga which is less than 10 Kilometers away. On its way the highly pressurized water
passes the surge vessel, containing half water and half compressed air (see figure 3.11 below).
The container helps to cushion the effects of an unprecedented backflow which can result
from a power outage or malfunctioning of the pumps. A backflow of such water could be so
powerful that it raptures the concrete and pipes at the pumping point. The surge vessel
normalizes the pressure of the water at all times, even in case of a backflow (Mugume 2006).    
Figure 3.11: The Pump House and Surge vessel
Photos by Wasswa
3.4. Conclusion    
This chapter described the current status of the Uganda water sector. It was established that
the current water usage has been irresponsible and has left the water resources in the country
in a poor state. Lake Victoria’s degraded state serves as a testament to this irresponsible use.
Service pumps  Surge vessel 70
Also, much as the legislation exists that caters for water quality in the country, little been done
to enforce it, resulting in improper management of the resources. Internationally the water
legislation is limited in its provisions. No clear transboundary water policy exists to protect
Uganda’s water bodies in spite of most of these water bodies, transcending boarders (Uganda
1999). As a result, the water bodies are increasingly being degraded, hence compounding the
social cost on the economy both directly and indirectly.
The water treatment process for the Kampala district’s water appears to generate potable
water at plant site (Oyoo 2006, Tumwebaze 2006). By final delivery however, this quality is
lost. This suggests that the chlorine concentration at the plant level should be maintained at
higher levels or that more monitoring is required of the water as it passes through the
reticulation phases.  71
CHAPTER FOUR:
COST­BENEFIT ANALYSIS
THEORY72
4.0 Introduction
The methodology employed in assessing the feasibility of improving Kampala’s water to a
drinkable standard was Cost‐Benefit Analysis (CBA). CBA is an appraisal tool in economics that
informs decisions between alternatives (of projects or policy) by allowing for inputs and
outputs to be measured in comparable units. Money inputs form the cost stream of a project
and the outputs, also measured in monetary terms, form the project’s benefits (Brent 2003).
First mention of the CBA dates as far back as the Nineteenth Century in work of Jules Dupuit
(1844) on the utility of public works. Dupuit (ibid) was concerned with decisions to invest in
goods that had no commercial returns like public goods (Brouwer &Pearce 2005). He (Dupuit
1844) established the concepts of consumer and producer surplus. The latter is the exces
return a producer recieves to supply a good. Consumer surplus on the other hand, is the net
benefit from consuming a good or service measured by the excess willingness to pay over the
cost of acquiring the good. He (ibid) resolved that the change in consumer surplus was
responsible for measuring the benefit of providing more of either market or public goods.
The CBA model was developed and made more usable by Alfred Marshall (1898). However, it
was not until the 1930s that CBA became widely applied. The methodology was employed in
the formulation of the Federal Navigation Act of 1936 also referred to as the Flood Control Act
of 1936. Under the Act, US engineers were supposed to carry out waterway systems  ‐
improving projects contingent on whether the projects yielded benefits in excess of the costs
involved in setting them up. The engineers needed a way of quantifying and comparing these
benefits and costs. Thus, the cost benefit analysis methodology of today was born (Hanley
&Spash 1993, Watkins 2003). Since then, the method has been extended to analyses involving
public goods like improving wildlife habitats, air quality, human health and aesthetic appeal of
landscapes.
The extension of CBA into the valuation of public goods has required the challenge of valuation
of intangible benefits and passive use values (PUVs) to be overcome. This challenge is one73
posed by the inexistence of markets for the exchange of such benefits and hence the
inexistence of prices and behavioral trends from which to infer their value.
There are many valuation methods employed to measure the value of goods and services
whose markets are non‐existent. These are called non market valuation techniques. The value
these techniques measure, makes up the benefit stream of a project. The CBA’s biggest
attribute is that in one analysis it can compare both costs and benefits of a project. In this
analysis, one allows for the time value of money. Employing the discount rate, benefits and
costs occurring over time are made comparable. The total discounted costs are subtracted
from the total discounted benefits to determine the over all desirability of the project. The
CBA framework requires the quantification of benefits and costs and the choice of the discount
rate. Emphasis will be given to these in the discussion of the application of CBA.
4.1. Application of the CBA
A typical CBA follows five major stages; these are discussed in detail below:
4.1.1. Defining the Scope of the Project
The first stage of the CBA is defining the objectives of the proposed project (Hanley &Spash
1993). This implies defining targets the proposed project seeks to achieve (benefits), who the
beneficiaries are and the means with which these targets are to be achieved (costs). The
researcher then decides on the mode of valuing these targets/benefits and how to measure
the means employed in achieving these targets/costs. The anticipated challenges in achieving
the targets have to be laid out. Also, there might be some targets that are immeasurable. The
researcher must decide, at this stage, how important these are to the whole analysis and so
resolve the implications thereof. It is good practice to have a basis for reference to determine
the project’s societal welfare implications. Normally, a base case – a scenario where the
project is not in place (also the same as the   status quo) is looked at in reference to the74
scenario when the project is in place; the project case scenario (Commonwealth‐of‐Australia
2006).
4.1.2. Quantifying the project benefits and costs:
The second stage of the CBA application is the quantification of the parameters to be
compared, the benefits and costs. The quantification of the project’s inputs (costs) is normally
conducted by making reference to the data on capital costs incurred in setting up of the
project and the operation and maintenance costs anticipated over the project’s useful life
(Commonwealth‐of‐Australia 2006).
The quantification of the costs can be done by one of three approaches. The first is the survey
approach, which involves the researcher asking the bearers of the costs what the estimate of
these would be over the project’s useful life. The second approach is the engineering
approach, which employs general engineering information to estimate a project’s inputs over
its useful life. The third approach combines the engineering and survey approaches to
estimate the project’s costs (Tietenberg 2000). The estimated costs should be validated by
how much it reflects the producers’ and consumers’ behaviour (Watkins 2003).
The benefits are sometimes harder to determine and there are often issues surrounding their
estimation. A project’s benefits should reflect the consumers’ behaviour (Watkins 2003); but,
data on consumer behaviour is not always available as in the case of PUVs. Consumers
increase consumption to a point where marginal benefits are equal to the marginal costs
(Watkins 2003). Therefore, an increase in consumption can only occur if prices are lowered. By
implication, the price of a good or service reflects the net benefit a consumer gets from a
good. This relationship can be studied by making reference to the marginal benefit curve,
which is the same as the demand curve (ibid). By ploting the demand curve of goods and
services, whether private or public, the total benefits can be determined (predicted) by
estimating the total consumer willingness to pay (WTP).  75
Benefits of water uses, the good being valued in this research, have both characteristics of a
private good and a public good. In non consumptive use it has the latter characteristic.
The other important issue is the project’s useful life. The useful life of a project is measured in
years and is determined by the nature of the project (UCL 2005). For projects involving health
gains, the life expectancy is often selected as the time horizon for the project.
4.1.3. Discounting the Benefits and Costs
An analysis of costs and benefits occurring over time is achieved through discounting. Money
income has a time value, and the discount rate reflects the rate at which the individuals would
be willing to sacrifice current consumption for returns in the future,   in other words, the rate
of time preference (NOAA‐Costal‐Services‐Center 2006). The choice of such a rate however,
has been, and still, is highly contentious. One could argue that the controversy over such a
rate, in part, hails from the fact that its choice rests on the economist’s discretion.
Rational allocation of a consumer’s resources would preclude decisions to forego current
consumption to invest in projects whose full returns materialize only after the contributors are
dead. Why then would individuals be asked to make sacrifices for more durable projects
whose full benefits they would not enjoy? The answer lies in ethical foundations and not
economic justifications (Marglin 2002). Future generations are likely to be harmed by
irresponsible resource usage on the part of the current generation. It is therefore, a moral
obligation of the current generation to protect future ones (Hanley &Spash 1993). The
protection of the future generation’s interests is achieved through low discount rates, but the
same instrument is required to also manage several other inter‐temporal trade‐offs.
The discount rate could be any number between zero and one. A zero discount rate would
equitably weight any benefits accruing to the future generation. However, the existence of
time preference and opportunity cost of capital suggests that the discount rate should be
greater than zero (Hanley &Spash 1993). The onus remains on us to determine what the
appropriate rate for a particular project would be.  The literature surrounding the choice of the76
discount rate advises on three approaches when choosing an appropriate rate for a particular
project.
4.1.3.1. Time preference rate of consumption
When analysing the feasibility of a project proposal whose costs and benefits are consumption
related, it is recommended that discounting be done by use of the time rate of consumption
(UCL 2005). This could be either the social rate of time preference or the private rate of time
preference. The social rate of time preference expresses the rate at which the society chooses
to forego current consumption. The private marginal rate of time preference on the other
hand expresses the rate at which individuals are prepared to sacrifice their current
consumption in favour of returns in the future (ibid).
The two rates are different from each other because decisions taken as a group are different
from those people make alone (Hanley &Spash 1993, Marglin 2002). Individuals may prefer set
high discount rates so as to enjoy current consumption, but moral obligations may induce
them to set lower rates. In the case of public project an authority such as the government has
the responsibility of setting the rate.
Empirical evidence offered by Marglin (2002) points to the use of the social rate of time
preference rather than the private rate of preference in guiding what rate to set.
4.1.3.2. Social opportunity cost of Capital
When analysing projects undertaken in one sector of the economy and not in another, the
discounting of the benefits and costs is best done by use of the social opportunity cost of
capital (UCL 2005). An investment undertaken in any one sector of the economy is done at the
expense of another project that could have been undertaken in another sector of that
economy. The discount rate based on social opportunity cost reflects the rate of return on an
investment foregone elsewhere in the economy by implementing one proposal (ibid).  77
4.1.3.3. Weighted discount rates
The last category of rates is ones set endogenously by the planners; for instance rates set for
projects with specific cost of capital (UCL 2005). Planners could also set this rate by weighting
the consumption rate of interest and rate of return on capital (Hanley & Spash 1993).   More
generally, Drummond et al. (1987) advise checking any rate adopted against economic theory.
Further, they argue that rates recommended by the government should be given preference in
discounting most public projects as these reflect the cost to government (ibid).
4.1.4. Decision criteria:
The crux of CBA is to inform on the feasibility of a policy or project. Various criteria are
employed in qualifying the feasibility of a project. The three most common decision criteria
employed are: the Net Present Value (NPV), Benefit Cost Ratio (BCR) and the Internal Rate of
Return (IRR).
The NPV is the most desirable criterion for decision determining the feasibility of a project
because of its simplicity. It is the amount by which the discounted benefits exceed the
discounted costs. It follows that such a figure has to be positive for a project to be feasible.
NPV is computed by subtracting the discounted costs from the discounted benefits, as
expressed below:
NPV = Present Value Benefits (PVB) – Present Value Costs (PVC)
Also expressed as:
     
( )
∑
=
+
=
N
t 0
t
t t
1 i
B ‐ C
NPV78
Present Value Costs
Present ValueBenefits
BCR =
The BCR expresses the rate of return per unit cost invested. A result greater than one implies
that the benefits exceed the costs and so the project is worth undertaking. The ratio is
calculated according to the expression below:
Lastly, the IRR criterion is the rate of interest that makes the NPV of all cash flows equal to
zero. This value should be higher than the cost of capital, also called the hurdle rate, for a
project to be deemed feasible. It is computed according to the formula below.
Where: i is the required value of the IRR and the numerator being the net benefit stream over
the given time t.
There might be multiple IRRs for a single project. This happens in case of multiple sign changes
in the series of the cash flows (Tietenberg 2000) and this makes the decision rule based on the
IRR harder to implement.
4.1.5. Sensitivity analysis:
A sensitivity analysis is the final stage of the CBA. The purpose of sensitivity analysis is to check
the reliability of the CBA results. Many assumptions are made in the CBA process. A CBA ought
to test for the impact on criteria if these were changed. This is called a sensitivity analysis. For
instance, one could check reliability by testing a pessimistic benefit scenario against a more
optimistic scenario of the project case (UCL 2005). Further, one could adjust the discount rate
to see the implications of an alternate rates.
( )
∑
=
+
=
N
t 0
t
t t
1 i
B ‐ C
079
4.2. The case for CBA
The CBA methodology was chosen for evaluating the efficiency case for the Kampala water
projects because it is appropriate. The discussion below presents these advantages (of CBA)
and the shortfalls of the other methods.
Efficiency in welfare economics is implied in the Paretian sense, which is an allocation that can
make no one better off while making the other worse off. In reality, this kind of allocation is
unlikely. This is where the compensatory role of CBA plays a major role. A Pareto efficient
allocation would be an allocation that compensates all losers for any losses they may suffer as
a result of the implementing a project (ibid). This is exactly what the decision criteria seen
before were trying to establish. If the analysis reveals benefits in excess of costs it means there
are enough gains to compensate all the losers, and CBA identifies what is efficient.
The CBA methodology also, ensures proper distributional impacts to the society(UCL 2005).
The model assigns equal weights to both the losers and gainers the decision criteria therefore,
reflect the true impact of the proposal to the society.
Cost effective analysis (CEA), as an alternative, compares costs and effects of a project. A
project with the lowest cost effect ratio is always preferred. In order to compare projects all
the effects and consequences of the projects or policies under review have to exactly be the
same (Brent 2003). A second alternative can be Cost Utility analysis (CUA). This is a method
that looks at the satisfaction gained form project’s implementation and the cost of attaining
this satisfaction. CUA is most suited to projects that have a fixed budget constraint. Which is to
say that CUA is best suited for choosing between one or more projects are socially desirable
whose funds are already allocated (Brent 2003).  
4.3. Conclusion
The merits the CBA methodology possesses over the alternative feasibility assessment
methods like the CEA and CUA make it best suited for the appraisal of the efficiency case of
the Kampala Water quality improving project.  80
The major issues surrounding the application of the CBA are the estimation of the project’s
benefits and costs and the choice of the discount rate. As resolved in the discussion of the
application of the CBA, a discount rate should be chosen with reference to the type of project
being appraised. For projects with consumption‐related outcomes, the time rate of preference
would discount the benefits and cost, whereas projects undertaken in a particular sector of an
economy, the social opportunity cost of capital would be the most suited discount rate to
employ (Hanley &Spash 1993, UCL 2005).  
As regards the second issue, that of benefits’ estimation, a mode of estimation is required that
can estimate the consumer surplus. Chapter five presents the theory of an approach of
estimation of projects’ benefits that elicits consumer surplus/WTP of using a good or enjoying
a service, the CVM.  81
CHAPTER FIVE:
CONTINGENT VALUATI ON METHOD
THEORY82
5.0. Introduction:
The benefits of the Kampala water quality improving project were quantified by use of the
CVM. The CVM is a survey based technique for estimating the value of non‐market resources
like environmental amenities. The CVM technique estimates value by measuring the
individuals’ willingness to pay for an improved resource scenario, or their willingness to accept
compensation for reduced resource service (Arrow et al. 2002, Brent 2003).
Proper estimation of benefits, be it for environmental amenities or water resources, should
account for both the use and non‐use values. The use values (UV) are those benefits users
derive from direct and indirect contact with the amenity in question (O'Doherty 1993). Market
values for UVs exist, and so their worth to the users can be easily estimated by referring to
such market values. These are values like water tariffs, levies to access a recreational site,
fishing permits and boat licenses to mention but a few. The Passive Use Values (PUV) on the
other hand are benefits such as existence, bequest and option values. The existence values are
gains individuals place on the resource‘s existence. This value could be for scenic beauty.
Bequest values are the benefits the individual perceives to accrue to the future generations.
Option values are the alternative uses the resource could avail in the future to the individuals
(Tietenberg 2000).
The nature of the PUVs suggests that there are neither markets nor behavioral patterns from
which their value to users could be obtained. The public nature of such benefits, implying they
possess the qualities of non exclusion and non‐rivalry, precludes the existence of markets for
their exchange. The properties bring about externalities in the enjoyment of these benefits. By
implication therefore, the market transactions cannot capture users’ preferences for these
goods (Hanemann 2002). The abstract nature and the temporal placing (in the future) of these
values also makes behavioral patterns of users difficult to ascertain. For these reasons the
estimation of the value of PUVs is difficult, but excluding them would result in an under
estimation of the total economic value. This argument strengthens the case for assigning a
monetary value to the PUVs and this may be done through use of the CVM. The technique83
predicts monetary value through models based on elicitation of the minimum amount a
person is willing to pay to enjoy an improved scenario or to prevent a decreased scenario
(Arrow et al. 2002, Haab &McConnell 2004). An alternative is to elicit the maximum value a
person is willing to accept for a decrease in the amenity, or the maximum value a person is
willing to accept to forego an improvement in the scenario (ibid).  For either moral or personal
reasons, people are willing to pay for the PUVs and The only way for the researcher to learn
about this value is by asking the users about it.      
The CVM model exploits collective choice to elicit from individuals how they would value
welfare improvements brought about by an improved scenario vis‐à‐vis the status quo. In the
case of this research, the welfare level of the Kampala households shall be assessed after the
city’s water quality is improved to a drinkable standard. The valuation of the projects’ benefits
as seen in the previous chapter is best done with the approximation of the area under the
demand curve. This area defines the consumer surplus or the consumers’ willingness to pay
that the CVM model seeks to predict (Arrow et al. 2002). However, an interesting question is
how valid the estimates generated by the CVM model are.
The question sits at the heart of the controversy that surrounds CVM model application. It is
not one that can be answered with precision but with reference to guidelines governing the
conducting of the CVM. The widely referred to set of guidelines are those prescribed by the
NOAA panel on contingent valuation. The panel resolved that if the guidelines are followed in
conducting a CVM, the estimates predicted offer a reliable measure of value (Brent 2003).
This chapter is arranged as follows: the background of the CVM is given, the issues surrounding
the method outlined and the CVM model as guided by the NOAA prescriptions is presented.
Finally, the stages of applying the CVM are described.
5.1. The CVM Model: Rationale and Critique
The CVM model is a utility maximization based approach (Haab &McConnell 2004). It predicts
marginal benefits of resources involving uses and user effects. Under the utility maximization84
theory, consumers maximize utility when they equate marginal utility per unit price of all
goods and services they consume. By implication, the price a consumer is willing to pay for a
good equals the additional satisfaction they derive from that good or service. There cannot be
a positive WTP unless consumers expect a positive net benefit from such an allocation (ibid).
Critics argue that the model cannot reliably predict the value of benefits, as the basis for this
valuation is hypothetical. In making valid scientific conclusions a means of disproof has to be
employed. The nature PUVs precludes any sort of disproof, by making reference to users’
behavior or past market trends (Arrow et al. 2002). The CVM therefore, could predict results
no one would be in a position to validate or invalidate. How can one ascertain the reliability of
such values? These pundits argue that results of CVM surveys in some instances are
inconsistent with rational choice (Arrow et al. 2002). For instance, in the work of Kahneman
(1986), it was found that the WTP estimate of cleaning up of all lakes in Ontario was only
slightly more than WTP to clean up lakes in one region. By implication therefore, WTP does not
increase with the quantity of the good.
The critics also call into question the plausibility of respondents’ WTP responses, as normally
these aggregate to very high figures, when the Total WTP is postulated from predicted WTP
(ibid) . This postulation is done by multiplying predicted WTP by the relevant population.
Further, they argue that the very hypothetical nature of the CVM valuation survey question
implies that the respondents have no meaningful budget constraints when stating their WTP,
making their responses all the more implausible (Arrow et al. 2002).
They also argue that a lot of new information is disseminated during the interview process and
it is unlikely that the respondents have enough time to fully process all this information and
take a reliable decision based on this (Arrow et al. 2002).  Even if the respondents were to fully
process this information, the pundits argue that the extent of the market for the PUVs might
not always be clearly defined. Therefore, there is a chance that the relevant parts of the
population might be left out in the CVM survey sampling (ibid).
Finally, there is the argument that the respondents may respond to a WTP question out of a
sense of duty, to appear environmentally responsible, also referred to as the ‘warm glow85
effect’ (Arrow et al. 2002). WTP predictions based on such impulses would not be a reliable
measure of value.
Such was the nature of criticisms raised in the case of Ohio Vs Department of the interior in
which the D.C court of appeal ruled in favour of inclusion of PUVs in assessment of resources
damage, if these could be reliably measured. If the CVM had to be used in valuation exercises,
the criticisms above had to be addressed. In this vein, the NOAA appointed a blue ribbon panel
headed by Kenneth Arrow including other prominent economists like Solow, Leamer, Radner,
Shuman and Portney, to ascertain whether the CVM could reliably predict the value of PUVs.
The panel resolved that if the design of a CVM survey was good, the results of this exercise
would be reliable (Arrow et al. 2002). The panel stipulated guidelines to be followed if a good
CVM were to be conducted and it is to these that this analysis next turns.
5.2. NOAA CVM design Guidelines
The NOAA panel CVM guidelines stipulate the procedure to follow in the design and
application of the CVM model if reliable estimates of value are to be attained.
5.2.1. Elicitation with the Referendum format
The presentation of the valuation question will in most, if not all, cases influence the
respondents WTP (Arrow et al. 2002) because CVM surveys rely on the respondents’ attitudes.
For this reason caution must be taken if reliable WTP estimates are to be attained.
The NOAA panel recommended that the valuation that yields more conservative estimates
would be more reliable than one that yields excessively large values. To this end WTP and not
WTA should be employed. In the same vein, the panel argued that the   valuation question be
put in the referendum style format, as the format   normally yields conservative estimates of
WTP compared to other stylised formats (Arrow et al. 2002). These other formats are
discussed in more detail later in this chapter. This stage in the CVM is where the respondents
are asked to vote for or against a proposal to increase deductions from their income to pay for86
an improved scenario or to prevent a decreased scenario.   The format, they argue, is realistic
and is not uncommon in real life. The respondents therefore, are bound to take the survey
seriously when the referendum format is used (ibid).
Much of the controversy, seen earlier, exists because of the impossibility to validate CVM
results (Arrow et al. 2002). The use of the referendum format provides a comparable case for
CVM results with real life referenda. Consider the changes in consumption of public goods,
such as reticulated water, or trips to a park after the tariffs of using these are raised. The
trends in consumption behaviour of the users, as a result of change in the status‐quo of the
public goods, can be employed as a validation check for the CVM when the referendum format
is used (Arrow et al. 2002).
However, the referendum format of elicitation is not free from flaws. The NOAA panel warned
that the respondents may misrepresent their actual WTP value for a number of reasons
including feeling pressed into accepting a value offered by the interviewer, poor information
on the project is given the respondents, if the respondents feel unfair demands are being
made for the programme in question, or if they are protesting about the questionnaire format.
Another reason for misrepresenting their real WTP could be that the respondents feel that
their vote may not have an impact on the decision concerning the project. In all the cases
above the   WTP may reflect a lack of thought in the answer given and hence would be an
unreliable measure of value (Arrow et al. 2002).
5.2.2. Addressing the Embedding problem
Where different, but similar samples of respondents were asked about the WTP for prevention
of identical welfare decreasing scenarios of different scale,  the WTP estimates have not been
found to be consistent with the scale of damage. This problem is known as the ‘embedding
phenomenon’. Normally respondents’ WTP would be independent of the size of welfare
decreasing scenario being averted (damage) but when it is not the WTP response may be
inaccurate (Arrow et al. 2002). Under such circumstances WTP would not be a measure of the87
money value of the utility of the welfare scenario preserved, but rather a measure of the value
of respondents would be attach to appearing morally and/or ethically right (warm glow) (ibid).
The panel recommended that the context of the valuation question be carefully looked at in
the design of the survey instrument to reflect a realistic scenario and take into account budget
constraints (Arrow et al. 2002). In their recommendations, the panel argued that embedding
type results could reflect diminishing marginal utility of the resource scenario in question
(ibid).
5.2.3. Time considerations and CVM surveys
There are two issues to consider in the time considerations of the CVM survey. The first issue
regards the resource amenity changes occurring over time. Where a resource has interim and
steady state PUVs, a valuation question that does not distinguish between these would predict
the wrong estimate of the resource’s value from the respondents’ WTP. Also, a valuation
question that does not make aware to the respondents that full restoration of the resource is a
reality would not yield reliable results. The predicted WTP would be based on wrong
information to the respondents and hence the results would be unreliable. The panel
recommended that adequate time should be allowed to lapse after the event for respondents
to regard full restoration of the resource a reality (Arrow et al. 2002). Experts should do the
calculation of the value of restoration over time, stipulating clearly what welfare level is being
valued, rather than leaving this to the respondents (ibid).    The second issue concerns WTP
values collected over time. If there is substantial time difference between the responses,
reliability of the results would be called into question. The panel recommended that WTP
response results should be averaged across independently drawn samples over time (ibid).
5.2.4. Sample size and type
The relevant sample population should be identified with the help of focus groups and the
technique to use in sampling this be selected. The panel recommended that as the choice of88
the sample size and sampling technique might be a challenging exercise, it should be done
with the guidance of statisticians (Arrow et al. 2002). Also, stratification and clustering should
be considered when face‐to‐face interviews are used. It is important to use random sub‐
samples, especially if the referendum format is being used (ibid).
5.2.5. Comprehensive survey instrument
The panel recommended personal interviews to any other. This method has to be supported
by a properly designed survey instrument. A properly designed survey instrument, according
to the panel, would adequately describe the welfare improving scenario being valued. It
should do this by use of photographic images of different resource scenarios. It should remind
the respondent of their undamaged substitutes for the resource, if any. The valuation question
should have a ‘No answer option’ to give the respondents full autonomy in their decision
making. It should provide for follow‐up questions to ascertain whether such a response was a
protest vote or otherwise.    The instrument should have a section to pre‐test there
respondents’ effects‐yes/no follow‐up questions, so as to determine of the respondents fully
understood and accepted survey mode and questions. The survey instrument should have a
section that captures the respondents’ household characteristics like income, education level
and age, to mention but a few, that would help in cross referencing their responses. Finally,
the survey instrument should be carefully pre‐tested before it is administered (Arrow et al.
2002).
5.2.6. Reporting
The panel recommended that every CVM report should clearly indicate the population
sampled, the sampling frame used and the sample size. It should indicate the non‐response
rate on all important questions, although these should be minimised in the process of
conducting the interview. The report should reproduce the exact wording and sequence of the
survey instrument. The data should be made available to interested parties (Ecosystem‐
Valuation 2007).89
5.3. Application of the CVM model
The CVM application should draw from the guidelines stipulated above. A properly conducted
CVM would involve six main steps as outlined below.
5.3.1. Defining the valuation problem
The first stage of the CVM is where the researcher determines exactly what services are being
valued and what welfare change implications are in question (Ecosystem Valuation 2007). It is
the stage where the relevant population is identified. For the Kampala water improving
project, the services to be valued would be the value of the health gains or the additional
Quality Adjusted Life Years (QALYs) and the value of the increased income due to averted
expenditure. Since this research defines water as an economic good rather than a right, the
relevant population for the Kampala survey would be all the users serviced by the NWSC.
5.3.2. Preliminary decisions about the survey
The second stage of the CVM model is where the researcher makes preliminary decisions
about the survey. These are decisions such as the sample size, the mode of conducting the
survey and the size of the budget (UDHHS 2007). The panel recommended that choice of a
sample size should be done in liaison with statisticians (see section 5.2 above).
The statistical guidelines to sample size selection prescribe the use of either probability or non
probability designs. The probability design relies on the laws of chance to estimate the sample
size while the non‐probability design relies on value judgment (2001). Since the use of value
judgments always invites controversy, an addition of which we would not need for the
maligned CVM model, a probabilistic approach to calculation of the sample size was opted for.
One probability design method that has gained popularity in sample size estimation is one that
relies on the tolerable error. The method was used for instance in the B‐Rao   (2002), Gope90
(2003), Mirabella (2000) as was applied to the CVM survey sample size selection by Vaughan &
Darling (2000).
The sample size is defined by the formula 5.1 below
                                N = [ ]
2
E
σ
Z
2
α
    ………………………. (5.1)
Where:
N ‐ Desired Sample size  
Z ‐ The 95% confidence interval Statistic
σ ‐ The Standard deviation of Income (also of WTP)
E ‐ The acceptable error in the sample estimate of the population WTP
The standard deviation and the variance in the expression 5.1 may not be known at this stage
of the survey, which may present a challenge to the estimation of the sample size. However,
Vose (1996) argues that by a priori predicting a simple triangular distribution for the WTP
responses, a standard deviation and mean of the WTP can be estimated.
If we predicted the maximum, modal and minimum WTP to be values a, b and c respectively,
the mean WTP can be calculated by the formula 5.2 below (Vose 1996):
Mean WTP= (a + b + c)/3 …………… (5.2)
The standard deviation (SD) would be calculated by the formula 5.3 below:
SD = Variance   = (a b c ‐ ab ‐ac ‐bc)/18
2 2 2
+ + …… (5.3).
Substituting the mean and standard deviation into the expression 5.1 would yield the required
sample size.91
5.3.3. Survey Design:
The third stage of the CVM involves designing of the survey instrument (Haab &McConnell
2004). This stage involves consulting focus groups and liaison with experts in the relevant field
of study. The mode of elicitation of the users’ WTP would be decided on at this stage. The
preliminary versions of the survey instrument have to be pre‐tested among randomly selected
households. The results from the pre‐testing exercise have to be assessed to help in
developing the final survey instrument. The other issue in design is identifying the ‘protest
bids’. The NOAA panel recommended that these should be eliminated from the analysis of
data. Therefore, clear means of identifying the ‘protest bids’ have to be included in the survey
instrument.
The referendum format of elicitation of WTP was opted for in the Kampala survey because of
the advantages presented earlier (see section 5.2 above). The other ways of posing the
valuation question would be:
The iterative bidding approach: This presents the respondent with an initial bid and based on
their response, the figure is revised upwards (downward) till the respondent declines (or
accepts) to pay, the final figure agreed to by the respondent is deemed their WTP (FAO 2002).
This method has the advantage of mimicking a true bargaining scenario, which may serve to
tease out the true value consumers are willing to pay. However, it carries the flaw of a starting
point bias  ‐ an undesirable tendency in surveys for the initial value leading the respondent to
believe this is the ‘right’ value. The repeated questioning also has the likelihood of irritating
the respondents, causing them to respond for the sake of getting the interview over with
(Haab &McConnell 2004).  
The payment card system: in this approach the respondents are presented with a list of
predetermined ranges of values of WTP on a card and asked to choose the range of their WTP
(Arrow et al. 2002) . Statistical models are employed to ascertain the distribution of the WTP
and predict the respondents’ expected WTP. The payment card method has the flaw of
restricting the respondents’ WTP only to the listed options. Proponents of the method, such as92
Breffle, Schulze, Rowe & Chesnut (1996) suggest that this is not a cause for concern if the
ranges of values are comprehensive.
Finally, the open‐ ended approach is where the respondents are asked to state their WTP in
the absence of any option.   The NOAA panel warned that this approach lacks realism, as the
consumers are rarely consulted in the allocation of public goods. Also this format would invite
strategic behaviour, under estimation of their WTP and over estimation of WTA (FAO 2002).
The discussion of the other modes of elicitation of the WTP estimate leaves the referendum
mode of elicitation as the more suited mode to asking the valuation question, if reliable
responses are to be obtained. The major issue concerning the use of the referendum approach
is the selection of bid values to use in elicitation of the WTP. The range of values of the bids
used in any dichotomous choice CVM, also referred to as the bid vector, has a big influence on
the outcome of the CVM survey. The bid vector can reduce the size of the sample needed to
attain a desired level of efficiency (Alberini & Carson 1993, Cameron & Huppert 1991). It is for
this reason that bid selection has received increased attention in the last two decades.
However, there is little agreement on what the preferred range of bid values should be.
There are two competing views concerning the selection of the range of bid values to use in
the elicitation of WTP. On the one hand a bid vector could be selected in such a way to have
closely spaced values (Ruud &Kooten 1998). On the other hand, if the researcher can make
logical predictions at what the expected WTP for the program would be, a bid vector allowing
for as little variation in both directions of this value    can be employed (Alberini 1995).
Maximum efficiency in predicting WTP with dichotomous choice CVM when only two bid
values are used, if the WTP is assumed to follow a normal distribution, (Alberini 1995).
However, the use of more than two bid values allows the researcher to test the distributional
assumptions of WTP. Also expanding the bid vector would increase the researcher’s likelihood
of having actual bid values falling in the region of the optimal bid value (ibid). For sample sizes
between 240 and 480 the bid vector should consist of either 6 or 8 bid values, and for sample
sizes up to 1200 and beyond the bid vector could consist of 6 up to a maximum of 12 bid93
values. There is minimal gain in efficiency by expanding the bid vector beyond 12 bid values for
any sample size (Vaughan &Darling 2000).
The selection of the range of the bid values entirely depends on the researcher themselves.
Implicit in Alberini’s (ibid) findings is the fact that the bid vector should consist of at least 6 bid
values, especially those conducted on sample sizes below 240 respondents. However, it is
important to consult the relevant experts such as the engineers and financial managers on
what the minimum value would be for the restoration as this determines the lower bound of
the bid vector. In doing so, the researcher ensures they predict a meaningful WTP estimate
and generate a reliable measure of value for the project.
5.3.4. Administering the survey
The fourth stage of the CVM and is where the survey instrument is administered to the
respondents. The personal interview process, recommended by the NOAA panel, was opted
for as this would ensure that the respondents fully understand the services being valued.
Randomly selected respondents of the relevant population were to be interviewed.
5.3.5. Processing of the data
The fifth stage of the CVM deals with analysis of data (Haab &McConnell 2004). The first
exercise involved in data analysis would be ‘cleaning the data’. This would involve identifying
and elimination of protest bids, removal of unrealistic bid values and identifying any other
unusable data. After this, the descriptive statistics are calculated and econometric models
appropriate to the type of valuation question are fitted to the data to predict the WTP.  
The basic model for analyzing the dichotomous choice responses is the Random Utility Model
(RUM) (Haab &McConnell 2004).
In statistical analysis of the dichotomous choice, two important assumptions are made. The
first is that there are only two scenarios to be analysed and the second is that the94
respondent’s WTP is unobservable to the researcher. The two scenarios could be such as the
status quo (without/base case) (0) and the improved scenario (with/project) (1). The indirect
utility of the respondent j (Uij
) can be defined as:
where the individual’s utility depends on:    their income yj
, an array of household
characteristics and attributes of choice (which may include questionnaire variation) denoted
by an m‐dimensional vector Zj
, and a set of preferences known to the respondent and yet
unobservable to the researcher, namely the error term (
ij
ε ) (Haab &McConnell 1997). The
subscript i denotes the scenario prevailing at the time (0 or 1)
In terms of the second mentioned assumption; if the respondent agrees to the bid presented
to them, their actual WTP is predicted to be above that bid value. If the respondent declines
the bid their actual WTP lies below that bid value (ibid). This random part of respondent’s
preferences, we can only make postulates about. The respondent, ceteris paribus will accept a
bid offered if they believe that they will be better off in the new scenario even with the
deduction to pay for the proposed project they make. This deduction we denoted by the letter
t. Therefore, the probability for a respondent saying yes is:
     
In order to proceed the functional form of ui
(yj
,zj
,εij
) has to be determined and the distribution
of  εij be specified (Haab &McConnell 2004). If we allow the utility function to be additively
separable in the deterministic and the stochastic preference, the utility function will be of a
form such as 5.6 below:
 
     
This redefines the probability function 5.5 above to become:
………………………. (5.4) u u (y ,z ,ε )
ij
= i j j ij
P{ } Yes = P{u1
(y
j
− t,z,ε1j
) > u0
(y
j
,z
j
,ε0j
)} …………. (5.5)
( )
i j j ij i j j ij
u (y ,z ,ε ) = v y ,z + ε …………. (5.6)95
P{ } Yes = P{v1
(y
j
− t,z
j
) + ε1j
> v0
(y
j
,z
j
) + ε0j
}
P{ } Yes = P{βln(y
j
− t) + α1
z
j
+ ε1j
> βln(y
j
) + α0
z
j
) + ε0j
}
P{ } Yes = P{βln(y
j
− t) − ln(y
j
) + (α1
− α0
)z
j
+ (ε1j
− ε0j
) > 0}
The general utility function in the expressions 5.6 and 5.7 does not characterize the marginal
utility of income of the respondents. A realistic scenario is that the marginal utility of income
decreases. If we consider a plausible utility function, that implies non‐constant marginal utility
of income, such as a log‐linear distribution of the income, the utility function 5.6 may be
redefined to become the function 5.8 below (Haab &McConnell 2004).
                       
                     
ij
v (y z ) ε βln(y ) α z ε
i j j
+ ij
= j
+ i j
+
 
Allowing the deterministic components of the probability function 5.8 to be additively
separable (ibid):  
Letting
( ) α1
− α0
= α and ( ) 1j 0j j
ε − ε = ε yields expression 5.11 below.
…………. (5.7)
……… (5.8)
P{ } βln( ) y
αz
j
ε
j
0
y t
= + + >
−
…….. (5.10)
……… (5.9)
…… … (5.11)96
If we assume εj
to be normally distributed with a zero mean and variance σ
2
we can specify the
parametric form of the equation as a standard normal probability as below (Haab &McConnell
2004):
                                                 
( ( ))
⎥
⎦
⎤
⎢
⎣
⎡ +
=
−
σ
αz βln
P[Yes ] Φ
y
y t
j
j
After the model has been estimated the parametric estimates can be employed to calculate
the mean and median WTP according to the functions below (ibid):
The mean
( ) ( 2
)
2
β
σ
2
1
β j
α
j j j j
E WTP α,β,z , y = y − y exp − z +
The median
⎟
⎠
⎞
⎜
⎝
⎛
⎟ = − −
⎠
⎞
⎜
⎝
⎛
_
β j
α
_
j
_
j
_
j
_
j
Md WTP α,β,z , y y y exp z
The use of the referendum approach has proved problematic when it comes to calculating
welfare measures from the estimated coefficients (Haab &McConnell 2004). The estimates can
generate negative WTP or overly high ones. In the event of such an occurrence, the researcher
could use a non‐parametric estimate to estimate the welfare measures. One such estimate
that has gained popularity is the Turnbull estimator (ibid).
The Turnbull estimator computes the mean and median WTP for the referendum format CVM
by assuming away any kind distribution for the data. Its algorithm is as follows (Haab
&McConnell 2004):
ƒ The lowest bid value is used to calculate the proportion of refusals to pay Fj
(for j=1,…,
k).
ƒ The next lowest bid value is used to calculate the proportion of refusals Fj+1 (for j=1,…,
k).
…… … (5.12)
…… … (5.13)
…… … (5.14)97
ƒ Fj and Fj+1 are compared. The Fj
is assumed to be monotonically increasing through to
Fk. If Fj
is greater than Fj+1, proceed with calculation of Fj+2; otherwise pool the cells j
and j+1 into one cell with bid boundary levels tj
, tj+2 and calculate Fj as the total number
of refusals in the two cells over the total responses in the two cells.  
ƒ Continue with the pooling process until a monotonically increasing Cumulative Density
Function (CDF) is achieved.
ƒ Set FK+1 to 1.
ƒ Calculate the Probability Density Function (PDF) fj as the step difference in the final CDF
(i.e. Fj
 ‐ Fj‐1).
The lower bound of the mean WTP can be estimated by calculating the sum‐product of the bid
values and the probability density function
                                      ∑
+
=
= +
k 1
j 0
LB j j 1
E t f …………. (5.15)
The median using the Turnbull estimator is not a discrete value but a range. The Turnbull
method provides an estimate of the range in which the median WTP lies. The lower bound of
the median is the bid for which the distribution function just passes 0.5 and the next highest
price identifies the upper bound (Hanley &Spash 1993).
5.3.6. Test for Reliability and Validity of Results
The final stage in the CVM application and is where the researcher checks the plausibility of
their results against a set of objective criteria to establish a basis for validation of the results of
the CVM. Three major criteria are recommended, namely repeatability, validity and esteem,
against which the results could be checked (Hanley &Spash 1993)
The repeatability test is a check employed to test whether the survey replicated under the
same conditions would yield the same results. Where the test has been employed, such as in
the work of Mitchell & Carson (1989) on benefits of improved freshwater and in Heberlien98
(1986) on the WTP of deer hunting permits, the CVM  has yielded stable and consistent results
over time (Hanley &Spash 1993). Often however, due to lack of resources and time, the
repeatability test is not performed.
There are three facets of validity that can be looked into to asses the credibility of a CVM
namely: construct, theoretical and convergent validity.
Construct validity is a test that checks if the design of the survey was according to the
generally accepted procedure. Checks are made concerning the questionnaire design and the
sample design.
Theoretical validity checks if the results generated are in line with economic theory (Hassan
2006), such as respondents exhibiting diminishing marginal utility in their valuing of resources.
The yielded results are checked against objective criteria such as expected signs of parameters.
For instance, income and WTP should have a positive relationship. The other criterion is the
explanatory power of the model fitted to the CVM data depicted in the R‐squared. Hanley &
Spash (Hanley &Spash 1993), recommend such value to be 15 percent at the minimum. Finally,
convergent validity is a test that checks whether the CVM results of a survey share some sort
of commonality with results where different valuation methods are employed. For lack of
monetary resources and time this test is often not performed.
The third check is that of esteem. It relates to how a particular survey’s results are
authenticated when compared to those of studies done before it. The researcher looks beyond
convergence of the CVM valuation techniques in the model’s application to establish how
believable a predicted WTP estimate is in reference to earlier studies’ findings. The results
analysed in comparison to those of other studies would bolsters the researchers position and
hence the credibility of the research.
5.4. Conclusion
While helpful in the quantification of both the UVs and PUVs, the CVM model has not been
free from controversy. In the CVM’s quantifying of PUVs lies the model’s most supportive and99
discrediting points. The nature of PUVs precludes any sort of meaningful validation. While the
NOAA panel guidelines of the CVM may not have furnished a scientific mode of validating the
CVM results, they helped harmonise application of the model in the world over. The panel
resolved that if adhered to, the guidelines would yield a reliable measure of the value of non‐
market goods. This conclusion would suggest that validation of CVM results would be possible
after all. By implication, one would check any CVM results against the NOAA panel guidelines
to determine if these were reliable or not.
The Kampala CVM was designed in line with the NOAA guidelines in mind (at least) to try and
ensure reliable estimates of the users’ WTP. The next chapter shows how the CVM framework
was applied to the Kampala case.  100
CHAPTER SIX:
KAMPALA CVM SURVEY101
6.1. Introduction:
In the preceding chapter, it was shown that a reliable CVM model is that designed according to
the prescribing of NOAA panel. CBA Studies from around the world using the CVM in informing
about policy changes in the environment and water, have mainly been developed under these
guidelines (NWSC 2005). The CVM is employed to quantify the benefits (see section 4.12,
chapter 4). To this end, the Kampala CVM survey was developed and conducted with the
NOAA guidelines in mind. This chapter shall focus on the survey itself, highlighting how the
guidelines were followed to arrive at the results. The chapter is arranged as follows: the
process of developing of the Kampala CVM survey is described, the results and statistical
analysis are presented, the reliability tests are reported and finally the conclusions are drawn.
6.2. Conducting the Kampala CVM
6.2.1. Defining the Kampala CVM scenario: Questionnaire design
The questionnaire used in the Kampala CVM survey was designed in consultation with focus
groups consisting of engineers, micro‐biologists and corporate planners in the Uganda water
Sector. Similar focus groups were consulted in South Africa from Nelson Mandela Metropolitan
Municipality water distribution section, Port Elizabeth.
From the consultation exercise, it was established that the major benefits of the project would
be health gains and cost mitigation. It was deemed important to understand the water quality
perceptions of the respondents, how they collected and stored their water and whether the
waterborne diseases were a common experience. The questionnaire was therefore drawn up
with these aspects in mind.
To this end the Kampala water quality‐improving project CVM scenario took an ex‐ante
perspective where the respondent was asked to value a project that provided them class 0
quality water either in‐house or in a vicinity less than a kilometer from their home. The quality
level specified (class 0) ensures that the water at the user level is water that is free from all102
disease causing microbes, chemical and physical impurities and so not deserving any
treatment before use. Their contribution toward the project would be through higher water
tariffs to the provider of water in the country, NWSC.
The questionnaire was divided into five sections; a copy of the questionnaire administered is
to be found in the appendix A. The first section covered bio data section. This section sought to
collect information on the respondents’ household characteristics. The second section sought
to collect information on where the respondents collected their water, how much it cost them,
how they stored it, any waterborne diseases suffered due to use of the water and their quality
perceptions about the NWSC water quality, i.e.  their water for current consumption. The third
section sought to collect data on how the respondents treat their water to make it safe.
Section four contained information about the nature of the project proposed and the benefits
such a project would bring. Respondents were informed that the project would be financed by
the financial contribution they were willing to make for such a project. The last section was the
valuation section. This section sought to learn about the respondents’ WTP response. The
section had follow up questions to establish the reliability of such a response.
6.2.2. Details of the survey
The details of the quality improving project should be found identified (Scenario being valued)
6.2.2.1. Sample size
Prior to conducting the survey the sample size was determined by use of a probabilistic
approach. This approach estimates the sample size by use of the expression 5.1 restated below
                                N = [ ]
2
E
σ
Z
2
α
    ………………………. (5.1)
Where:
N ‐ Desired Sample size   103
Z ‐ The 95% confidence interval Statistic
σ ‐ The Standard deviation of Income (also of WTP)
E ‐ The acceptable error in the sample estimate of the population WTP
The standard deviation of WTP and the tolerable error (in expression 5.1) are unknown. These
measures have to be established before the sample size can be determined. The standard
deviation was calculated in line with Vose’s (1996) argument of predicting a priori a simple
triangular distribution of WTP, viz, this is a distribution with an upper limit denoted by a, a
mode we could denote by b and a lower limit denoted by c.
The upper limit, a, was set by the maximum WTP permissible of 5 percent of one’s income as
suggested in the CVM guidelines (see FAO 2002). According to recent data from the Ugandan
central bank, the average household annual income was Ushs 565,908/= (BOU 2005). The
NWSC on average produces 112,862 m
3
per day in Kampala (NWSC 2005), which is 41,194,630
m
3
annually. The user population served by the NWSC in the district is 872,433 people (Lane
2005). Therefore, per capita water use is 47.2 m
3
. If we allow maximum willingness to pay to
be 5 percent of ones income, then the upper limit a is Ushs 599.2/= per m
3
(5% of
565908/47.2). The modal WTP b was predicted to be 2 percent of ones income, which is what
users currently devote to water. The modal WTP was therefore calculated as Ushs 240.8/= per
m
3
(2% of 565908/47.2). The lower limit was predicted to be 0, hence c is 0.
For a triangular distribution, the mean and SD can be calculated as follows:
Substituting these values for a, b and c into the expression 5.2 for the mean we have that:
Mean WTP= (a + b + c)/3 …………… (5.2)
Mean WTP = (599.2 + 240.8 + 0)/3 = Ushs 280/=
Substituting these values into the 5.3 for the standard deviation we have that:104
SD =   Variance = (a b c ‐ ab ‐ ac ‐bc)/ 18
2 2 2
+ + …… (5.3).
SD = Variance = (599.2 240.8 0 ‐144287.36)/18
2 2
+ +
SD = Ushs 15117.42
Therefore:
SD= Ushs 122.95/=
In determining the permitted/tolerable error, the budgetary provisions for the survey have to
be taken into account. The lower the error, the higher the sample size and the cost of the
survey, and vice versa.  The financial provisions for the Kampala survey permitted an error of 6
percent and higher. For this reason a tolerable error of 6 percent was allowed for, i.e. within
6% of the mean, hence E= Ushs16.8/=.
Substituting all the values into the expression 5.1 we have that:
N = [1.96 x 122.95)/ 16.8]
2
= 205.755
Based on Vose’s (1996) method, the sample size for the Kampala survey adopted was 206,
which constituted 0.2 percent of the total consumer population (see NWSC 2005).
6.2.2.2. Representativity of the sample
Representativity of the sample size, using the probabilistic approach is dependent on the level
of error permitted in the calculations. Normally, the level of error should not exceed 10
percent as employed in Vaughan & Darling (2000). The 6 percent level of error employed for
the Kampala survey sample size selection is 4 percentage points below the maximum and so
deeming the sample size representative.   105
Representativity of a sample depends on how well the sampling frame is defined – this must
be representative of the target population being studied. Also, the sampling technique to learn
about this sample should control for sampling bias. The respondents in the Kampala survey
were randomly selected reticulated water users from the 10 Kampala water branches; this
gave all respondents an equal probability of being selected (obviating biasness). The sampling
frame included all reticulated water users. Based on the unbiased sampling technique and the
representative sampling frame, the Kampala CVM survey sample was deemed to be
representative.  
6.2.2.3. Elicitation of WTP
The referendum style format for CVM surveys was adopted in designing the questionnaire; this
being one of the formats accepted by the NOAA panel for the CVM. The elicitation question
was put in such a way that let the respondents envisage the scenario as though they were
voting to tax themselves for the water improvement project (see fig 6.1 below). The payment
vehicle, higher monthly water bills, was selected after careful pretesting.
Figure 6.1: The elicitation
Suppose the NWSC undertakes a project to improve Kampala’s water to a safe drinkable
standard. To finance this project, would you be prepared to pay an extra Ushs………. /= per
m
3
? This extra charge will be collected by the NWSC in the normal monthly way.
                   Yes                          No             No Answer  
6.2.2.4. Bid selection
The referendum approach is suited to a water quality improvement project as there is a large
degree of ‘publicness’ in the provision of potable water to urban communities. With such an106
approach, the tendency to respond strategically is reduced as the more truthful answer
favours the respondent. As a result the reliability of the CVM survey is improved (Arrow et al
2002).
The referendum format must attend to the bid vector as this influences the predicted WTP.
Drawing from the earlier discussion (see chapter 5), a bid vector that allowed for bid values to
be equally spaced from each other was adopted. It provided for this had six bid values.
Planners were consulted in the process to preclude selection of a bid vector that would not
cover the project costs. Based on their estimates, the cost of improving water quality was
estimated to be Ushs 1451/= per cubic meter (Kahwa 2006). The highest bid offered to the
respondents was calculated as the difference between this value and what the Kampala water
users currently pay (Ushs 806/= per cubic meter) to give a value of Ushs 645/=. The
subsequent five bid values were values equally separated by 16 percentage points from the
other (100/n, where n=6). Hence the bid values were 542, 456, 384, 322 and 270.
6.3. The Results
6.3.1. Social economic characteristics of respondents
The CVM survey was conducted in the Kampala water jurisdiction area. This area is sub‐divided
into eleven branches. The city’s map, showing the water zones, is presented in the appendix
(see appendix B). The branch 7 is not demarcated like the others. It is a branch reserved for
servicing the big consumers of the city, like industries and government institutions. It is thus
not a physical location that could be surveyed. For this reason only ten branches could be
surveyed. Twenty questionnaires were administered to randomly selected residents in each of
the zones.107
Figure 6.2.: Gender Participation in survey
Prior to conducting the data analysis, the data was screened to identify all the ‘protest zeros’,
incomplete questionnaires and all invalid responses. After the screening exercise 156
questionnaires were found to be usable.This data revealed the gender participation rate to be
60.9 percent for males and 39.1 percent for females (see fig 6.2).
The average age of the respondents was 33 (see in table 6.1 below). The average annual
income was 1,667,629 shillings. The distribution of age was skewed. As a result, the median
income of 1250001 shillings was deduced to be the more plausible measure of central
tendency. The median is not affected by extreme values (Raymer 2005, Winter 2005).
Table 6.1: Descriptive statistics from the Kampala Survey
Parameter N Minimum Maximum Mean Median Std.
Deviation
AGE (In Years) 156 23 68 35.88 33 11
HOUSEHOLD NUMBER 156 1 20 5.33 4.5 3.68
INCOME (In Shillings) 156 50000.5 4000001 1667629 1250001 1373068
MITIGATION EXPENDITURE (In
Shillings)
156 0 302835 31137.08 2835 56580.07
COST OF WATER   156 0 145000 29246.76 15000 31899.79
(In Shillings)108
The average household number was 5 persons. The mean cost of water was of 29246.76
shillings. This cost was an aggregate of what the respondents spent on their water bills
monthly, or what they spent monthly collecting water from a different source, either at a
nearby tap, well, or open source. For those that did not spend in monetary terms, a time cost
was approximated, based on the average population earnings.  
The majority of the respondents (54 percent) had more than thirteen years of school (table 6.2
below). Most respondents understood the implications of the proposed project.
Table 6.2: Education level of those surveyed.
   Frequency Percent Cumulative
Percent
No formal Education 4 2.6 2.6
Primary Leaving Exam Certificate and Below 12 7.7 10.3
O‐Levels 30 19.2 29.5
A‐Levels 29 18.6 48.1
National Diploma 38 24.4 72.4
University Graduate 39 25 97.4
Post Graduate 4 2.6 100
Total 156 100  
Most of the respondents (52 percent) collected their domestic water from a NWSC tap less
than a kilometre away. About 35 percent of the respondents had an in‐house water
connection. Protected wells were utilised by 7 percent of the respondents, while close to 2
percent of the respondents used storm water.  The role of open water sources (river and lake)
and other sources (including unprotected wells and boreholes) was equally important, catering
for 1.5 percent of the respondents (see figure 6.3).109
Figure 6.3: Respondents’ Water Source
There is no single measure that captures water quality. The latter is based on attributes of
water such as colour, taste turbidity, conductivity and coliform count (Tumwebaze 2006,
Winter 2005). The respondents were asked to rank amenities of taste, colour, odour and
microbial safety on a 6‐point scale as either:   excellent, very good, good, fair, poor, or very
poor.
Over 53 percent of the respondents perceived the NWCS reticulated water as fairly good.
Close to 30 percent perceived the water as good. The percentage of the respondents that
perceived the water as very good was equal to the percentage that perceived the water as
poor, viz 7.7 percent.   Only 0.6 percent thought that the water was excellent and 1.3 percent
thought the water was very poor (see figure 6.4 below).
It is noteworthy however, that all the respondents either boiled their water or treated it with
chemicals before drinking it.110
Figure 6.4: Quality Perceptions
Experience with waterborne diseases was common in the respondents’ households. Typhoid
fever accounted for most sick days (see Figure 6.5 below). About 30 percent of the
respondents had suffered from the disease. Acute diarrhorea and dysentery followed afflicting
10 and 6 percent respectively, of the respondents’ health. Cholera was less commonly
experienced, only affecting 3 percent of the respondents. Other ailments, such as skin rush
and irritation were captured under the ‘other disease’ category. These only affected 1.9
percent of the respondents (see fig 6.5 below).  111
Figure 6.5: Kampala Waterborne Disease Burden
The valuation question sought to learn about how the respondents valued the benefits of the
water project. The results in the table 6.3 below summarise the responses for every offered
bid.  
Table 6.3: Dichotomous Valuation Response Summary  
BID VALUE
(Ushs)
   WTP
Total
NO YES PROTEST
270 13 22 7 42
322 13 18 3 34
384 15 18 2 35
456 14 14 2 30
542 14 16 1 31
645 6 11 3 20
Total 75 99 18 192112
This analysis was conducted before the data was screened for validity. Non Willingness to Pay
increased with the bid amounts, except for the highest bid of Ushs 645/=, which had a low
response rate of just 20 responses, compared with the bid of Ushs 270/=, that had 42
responses. WTP decreased with the bid amounts (see table 6.3).
6.3.2. Parametric Results
The goal of parametric modelling is twofold: first to establish a form of correlation between
the respondents’ WTP and their household characteristics and secondly, to predict the WTP
for the service being valued.
The dependent variable (WTP) is measured using the dichotomous style format of elicitation. It
is a binary response variable that is analysed by use of qualitative response model (Garson
2006). A multinomial logistic model was fitted to the Kampala CVM survey data using the
statistical package SPSS 10.1. This type of process can accommodate both continuous and
design (factors) variables. The method of estimation of the logit coefficients is Maximum
Likelihood Estimation (MLE). The MLE procedure seeks to maximise the log likelihood (LL). The
log likelihood (LL) reflects how likely it is that the observed values of the dependent may be
predicted from the observed values of the explanatory variables (Greene 2000). WTP was
related to the respondents’ household characteristics of: age, gender, education level, income
and number of household, the respondents’ perceptions of the current water quality, the cost
of water, and the mitigation expenditure.
To interpret the coefficients of the logistic model, they are converted to an odds ratio using
exponential function (exp(x)). The odds ratio is a measure of effect size
12
. It defines the impact
changes in the independent variable have on the probability of predicting the dependent
variable (Schwab 2007). The SPSS 10.1 MLE procedure generates these values in its output.
                                                     
12
The measure of the strength of the relationship between two variables113
When interpreted in terms of the parametric form 5.12, the odds ratio describes the impact
the parameter has on the likelihood of the respondent giving a Yes response to the bid offered
them.
The description and anticipated signs of the estimates for the nominal logistic model fitted to
the Kampala CVM survey data are summarized in the table 6.4 below.
ƒ AGE: The coefficient of the age variable was anticipated to be positively related to the
dependent variable because older people were expected to be more likely than
younger people to appreciate benefits of improved water quality, such as health gains.
ƒ EDUC: The coefficient of the education variable was expected to be positive on the
grounds that more educated respondents would be more knowledgeable about the
dangers of having poor quality water, and for this reason support the proposed project.
The MLE procedure permits the use of categorical variables and so education was
regressed as a categorical variable, with cases as seen in the table 6.4 below. The SPSS
procedure converts the categorical variables to dummies automatically leaving out the
last category. This last category is set to the base level, the reference category.
ƒ QUALITY: The quality variable was also regressed as a categorical variable measuring
the impact one’s perceptions about the water’s quality had on the odds of accepting a
bid offered them. The coefficient was expected to be negative as the lower the
respondents perceive the quality of the water to be the more they would be expected
to be willing to pay for a project that improves upon it.
ƒ GENDER: The gender coefficient could be either positively or negatively related to the
dependent variable as both males and females are expected to be equally receptive to
benefits of improved water quality. This variable was included to control for the water
quality perception, differences between the genders.
ƒ LNINCOME: this variable measured the impact the portion out of the respondent’s
income devoted to the project had on the odds of their being WTP. The coefficient was
expected to be negative because the more the water improving project costs the
respondents, the less inclined, it was anticipated, they would be willing to pay for it.    114
Variable
Description
Expected
Sign
Dependent variable
 
WTP Binary response variable with values  
1 = Yes response to a bid offered    
0 = No response to offered bid  
Independent Variables
AGE Respondents age +
GENDER
1= If respondent is Male
+/‐
0= If respondent is Female
LNINCOME
13
Respondent’s income. This variable is estimated as WTP as a portion of
one’s income.
‐
LNCOSTO W Amount respondent spends to have water for Domestic use
‐
LNDEFEX
Amount the respondent commits to making the water safe to drink. It
also included the average amount committed to treating waterborne
diseases.
+
HHNUMBER The count of the respondents’ household members.    +
  EDUC
0  = No formal Education
+
7  = Primary Leaving Exams Certificate and below
13 = A‐Level
14 = Diploma
16 = Graduate
20 = Post Graduate
1 = If respondent perceives quality of water as Very Poor
‐
   2 = If  respondent perceived quality of water as Poor
3 = If  respondent perceived quality of water as Fair
QUALITY 4 = If  respondent perceived quality of water as Good
   5 = If  respondent perceived quality of  water as Very Good
   6 = If  respondent perceived quality of water as Excellent
                                                     
13
 
⎠
⎞
⎝
⎛ −
y
y t
ln Where: y is the respondent’s income and t is the bid value.
Table 6.4: The variable description and anticipated coefficient signs115
ƒ LNCOSTOW: This variable measured the impact the cost of abstracting water had on
the respondent’s acceptance of the bid offered them. The estimate of this was
expected to be negative because the higher the cost of abstracting water, the less
inclined one would be to adding to it.
ƒ LNDEFEX: This variable measured the effect costs incurred in making the water
drinkable (mitigation costs) had on the odds of WTP. The coefficient of this variable
was anticipated to be positive because the more monetary resources the respondent
commits to making the water drinkable, the more willing, they would be to support a
project aimed at improving the quality of the water.
ƒ HHNUMBER: the variable measured the number of members in the household. Its
estimate measures the impact the number of members in the respondent’s household
had on the odds of their WTP. The expected sign of the estimate was positive. This is
because the more members a household has, the higher the likelihood of spread of
waterborne diseases, and also the higher the cost of the burden of disease (as more
that one member may be affected). Therefore, respondents with more household
members are likely to value more the benefits of a water quality improving project and
hence, are likely to have higher WTP.
6.3.3. Discussion of Results
The complete model had eight variables (see table 6.4); of which three were significant at
α=0.05. These were the variables controlling for quality perceptions of the respondent
(QUALITY), income devoted to the project (LNINCOME), and education level (EDUC).
Since MLE seeks to maximise the Log Likelihood (LL), reliability testing of the model is based on
the Log‐Likelihood. The overall fit of the model; that is the existence of a relationship between
the dependent variable and the independents, is tested by the statistical significance of the
final model. The test employed in this vein is called the Log Likelihood ratio test, also referred
to as the G‐test. The log likelihood ratio test is simply the difference between the null model116
.663 .774 .733 1 .392
-169.325 70.236 5.812 1 .016 2.905E-74 4.767-134 1.771E-14
-.188 6352.471 .000 1 1.000 .829 .000 .
a
-19.666 .991 393.795 1 .000 2.878E-09 4.126E-10 2.007E-08
-18.746 .773 588.698 1 .000 7.222E-09 1.589E-09 3.283E-08
-18.849 .734 658.915 1 .000 6.518E-09 1.545E-09 2.749E-08
-19.542 .000 . 1 . 3.259E-09 3.259E-09 3.259E-09
0
b
. . 0 . . . .
18.822 1.245 228.483 1 .000 1.5E+08 13016401.53 1715369946
18.300 .694 694.784 1 .000 8.9E+07 22725189.14 345469078.9
17.526 .531 1087.811 1 .000 4.1E+07 14419457.01 115757808.9
17.414 .566 945.753 1 .000 3.7E+07 12040333.53 110816472.6
18.098 .484 1396.818 1 .000 7.2E+07 28045701.06 187179902.4
17.531 .000 . 1 . 4.1E+07 41088305.08 41088305.08
0
b
. . 0 . . . .
Intercept
LOGINCOM
[QUALITY=1]
[QUALITY=2]
[QUALITY=3]
[QUALITY=4]
[QUALITY=5]
[QUALITY=6]
[EDUC=0]
[EDUC=7]
[EDUC=11]
[EDUC=13]
[EDUC=14]
[EDUC=16]
[EDUC=20]
WTP
0  No
B Std. Error Wald df Sig. Exp(B) Lower Bound Upper Bound
95% Confidence Interval for
Exp(B)
a. Floating point overflow occurred while computing this statistic. Its value is therefore set to system missing.
b. This parameter is set to zero because it is redundant.
(the model with only the constant) and the model with the parameters. The difference in the
likelihood follows a Chi‐square distribution and is referred to as the model Chi‐square‐
2 ( 2 )
2
G = χ = − LLnull
− − LLK
. Applying the G‐test to the complete model fitted to the Kampala
survey data produced a model chi‐square value of 24.44 at 17 degrees of freedom. The model
chi‐square had a probability value of 0.085 that exceeded the level of significance of 0.05
suggesting non significance of the model. By implication, the null hypothesis was accepted that
there is no difference between the model with the independent variables and the baseline
model with only the constant.
Table 6.5: Parameter Estimates
Estimation of the model was redone excluding the non‐significant variables, and this improved
model’s predictive capacity. Results of this estimated model are presented in the table 6.5
above. The reduced model, controlling for three variables, namely, the proportion of income
devoted to the project (LNINCOME), the education level (EDUC) and the quality perceptions of
the respondents (QUALITY), was very significant with a p‐value of 0.017 at 12 degrees of117
freedom. By implication the null hypothesis was rejected that all the coefficients were zero
except the constant.
Success of a logistic model, that is, the errors or accuracy associated with the model in its
predictions, can be tested by use of either the Hosmer‐Lemshow (H‐L) test of goodness of fit or
the classification table. However, in studies where the number of observations is small (<400),
as was the case with the observed data (Kampala survey data), the H‐L test is not
recommended (Hosmer & Lemshow 2000). With the classification table, utility of a
Multinomial Logistic model is determined by assessing how the classification accuracy rate
compares with the by chance accuracy criteria. The ‘by chance’ accuracy ratio is computed by
squaring and summing the proportion the cases in each group defined by the dependent
variable, as shown in table 6.6 below
Table 6.6: The case process summary output for the Kampala CVM model
Case Processing Summary  
 
   N Percentage
WTP 0  No 68 43.59
   1  Yes 88 56.41
Total
   156 100
The by chance accuracy rate for the estimated reduced model was 50.8 percent (0.4359
2
+
0.5641
2
).
A useful Logistic regression model is any that provides a 25 percent improvement over the rate
of accuracy achievable by chance alone (Newsom 2005), i.e. , the proportional by chance
accuracy criterion is that the by chance accuracy rate be enhanced by 25 or more percentage
points. It was calculated as 63.53 percent (1.25 x 0.508). The classification accuracy rate,
drawn from the SPSS‐generated output was 64.7 percent
14
    (see table 6.7 below), which
                                                     
14
156
[31 + 70 ]118
exceeded the by chance accuracy criteria. It was accordingly deduced that the criterion for
classification accuracy was satisfied, implying that the model was useful.
Table 6.7: The Classification Table
Observed
Predicted
0  No 1  Yes Percent Correct
0  No 31 37 45.6
1  Yes 18 70 79.5
Overall Percentage 31.4 68.6 64.7
The classification table above suggests the absence of homoscedasticity, an underlying
principle of MLE. If homoscedasticity existed, the percent correct would be the same for both
(0 No and 1 Yes) rows, but compare 45.6 percent with 79.5 percent.
In MLE, success of the model in explaining the variations in the data is assessed by correlation
measures called Psuedo R
2
. These estimate the strength of the relationship between the
dependent variable and the independents. Though analogy to the OLS’s R‐squared can be
made, the Psuedo R
2
cannot be interpreted as a measure of goodness of fit, i.e. as measures of
the percent of variance explained by the independents. The reason is that the variance of the
dichotomous dependent variable depends on the frequency distribution of that variable.   In
logistic regression, no true R
2
value exists (Newsom 2005, O'Connell 2006). The strength of
association can be measured by either the Cox and Snell R‐squared, the Nagelkerke R‐squared
or the McFadden R‐squared. The Psuedo R‐squared assess predictive power of the model
based on the LL ratio achieving a measure of between 0 and 1. The maximum of the Cox &
Snell cannot reach 1. The Nagelkerke R‐square is the correction that allows for the Cox and
Snell R‐square to have a possible maximum value of 1.   For the reduced model, the Psuedo R‐
square for McFadden was 0.115, for Cox & Snell was 0.146 and for Nagelkerke was 0.195 – all
suggestive of low but acceptable predictive power of the Logistic model. According to Hanley
& Spash (2004), a minimum R‐square of 15 percent can be considered acceptable in these
types of analyses.  119
The MLE algorithm predicts the direction and change in the logit coefficients that will increase
the LL. The estimates of the control variables therefore predict the magnitude and direction of
the odds of being WTP. A 10 percent increase in the proportion of income devoted to the
project (LNINCOME) reduced the odds of being WTP by an infinitesimal margin of 2.905E‐74
(approximately 0 percent), holding other variables constant. However, care should be taken
when interpreting the impact this variable had on WTP, as it exhibited signs of multicollinearity
inherent in the high standard error associated with its parameter (see table 6.5 above).
The coefficients for the quality variable were negative, implying that respondents who
perceived water quality as having a high quality were less inclined to pay for the project. The
quality level 6 (excellent) was set as the base category. The comparison of the respondents
that perceived the water quality as very good [QUALITY 5] compared to those that perceived it
as good [QUALITY 4] yielded an odds ratio of 0.5
15
for a yes response to WTP question, which
implies that the difference in perception about the quality from very good to good lowered the
odds of WTP by 50 percent (0.5 – 1 =  ‐ 0.5). The difference in perception from good [QUALITY
4] to fair [QUALITY 3] lowered the odds of WTP by 9.8 percent. The difference in perception of
the water quality from fair [QUALITY 3] to poor [QUALITY 2] increased the odds of WTP by
more than double, that is, and increase of 150 percent in the odds of WTP.
The coefficients of the Education level categorical variable were positive, indicating that the
more educated people were, the more willing they were to pay for the project. For instance,
presenting the valuation question to a national diploma holder [EDUC 14] rather than a
University graduate [EDUC 16], reduced the odds of being WTP by 43.3 percent. However,
oddly, the presentation of the questionnaire to an A‐level graduate rather than a diploma
holder increased the odds of being WTP by 98 percent. The difference in level between the A‐
level graduate [EDUC 13] and an O‐level graduate [EDUC 11] reduces the odds of WTP by 10.6
percent. If one presented the valuation question to a PLE certificate holder [EDUC 7] rather
than an O‐level graduate [EDUC 11] it reduced the odds of WTP by 53.9 percent. Presenting
the valuation question to an illiterate person [EDUC 0] rather than to a PLE certificate holder
                                                     
15
  Exp(‐19.542+18.849)120
[EDUC 7] reduced the odds of being WTP by 40.6 percent. Being a graduate [EDUC 16] or
postgraduate [EDUC 20] had no major influence (≈ 0 percent) on the odds of being WTP.
6.3.4. Reliability Checks
As noted in the previous chapter the results ought to be checked against objective criteria to
establish their reliability (see section 5.4). Due to constraints of time and money, only
construct and theoretical validity checks could be carried out. The signs of the parameters
were in line with the presupposed hypotheses. In regard to the predictive power of the model,
the three Psuedo R‐squared values averaged (15.2), exceeding the minimum ‘threshold’ R‐
squared value of 15 percent. In relation to other dichotomous CVM studies, the Kampala CVM
results compared well with those of many studies ‐  see Haab & McConnell (2004), Nam & Son
(1995), Alberini (2000), Vaughan & Darling (1998) and Ruud & Kooten (1997), to mention but a
few.
6.5. Estimating WTP.
The WTP is predicted from the model represented by equation 5.13 (see chapter 5). The
predicted WTP failed the reasonability test, as it was negative. Such occurrences are not
uncommon in the calculation of WTP from estimated coefficients, as pointed out by Haab &
McConnell (Arrow et al. 2002). In the event of having a negative WTP Haab & McConnell (ibid)
recommend the use of the Turnbull estimator to predict WTP (see equation 5.15,chapter 5) as
shown in the table 6.8 below.  121
Table 6.8: The Turnbull Estimation of WTP
BID (Ushs)
WTP
Total
Unrestricted Turnbull
1  YES 2  NO Prop No F (CDF) f (PDF)
270 22 10 32 0.313 0.313 0.313
322 14 12 26 0.462 0.462 0.149
384 15 13 28 0.464 0.464 0.003
456 13 14 27 0.519 0.471 0.007
542 14 13 27 0.481 Pooled Back Pooled Back
645 10 6 16 0.375 Pooled Back Pooled Back
645+  ‐   ‐   ‐  1 1 0.529
Total 88 68 156      
The lower bound of the Mean WTP was estimated by the expression 5.15 restated below:
∑
+
=
= +
k 1
j 0
LB j j 1
E t f  
Substituting the Probability Density function (f) into the expression we have that:
ELB= Ushs 385.07/=
The median WTP is in the range of 456 and 542 Ushs.
The benefits of the Kampala water quality improving project were therefore deduced to be
385.07 Ushs per cubic meter of water, the WTP predicted for the median respondent.
6.6. Conclusion:
The CVM was adopted to estimate the benefits of the Kampala Water Quality Improving
project as part of the CBA methodology. In the Chapter Five, the theory of applying the CVM122
was presented. The general conclusion was that if the CVM was conducted in line with the
NOAA guidelines, the results would reliably predict a projects benefits (Oburu 2006, Oyoo
2006).
This chapter (six) was a presentation of the CVM conducted for Kampala. The survey was
conducted with the intention of adhering to as many NOAA guidelines as possible, and the
results were used to predict the project’s benefits. The descriptive statistics showed that over
85 percent of the respondents used NWSC water. Over 60 percent of the respondents
perceived the water to be of good quality, even though they all treated it before use for
drinking. Waterborne diseases were a common experience among the respondents, especially
typhoid. From the results it is evident that there are water quality problems in the Kampala
district even among the NWSC water consumers, but perhaps that the users have got so used
to the inconsistent quality that they now accept it as good – not believing that there really is a
better alternative.
Parametric analysis of the survey results of the survey results was done by fitting a logistic
model to the household characteristics. The parametric results showed that the more
educated a respondent was, the more they were WTP. The more it cost the respondent, the
less inclined one was to pay for the project, and the better one perceived the quality of their
water, the less inclined they were to pay for the Water quality improving project. The results
were tested for credibility – most of the glitches identified were those experienced by
researchers using the CVM, i.e. were human researcher error types.
The crux of the CVM survey, and hence this chapter, was to predict user WTP for a Kampala
water quality improving project. The benefits were estimated by the WTP of the median
respondent by a non‐parametric estimator, the Turnbull Estimator, to be Ushs 385.07/= per
cubic meter of water. In the next chapter, this value shall be compared to the anticipated
project costs per cubic meter of water.    123
CHAPTER SEVEN:
KAMPALA COST­BENEFIT ANALYSIS124
7.1. Introduction
The benefits and costs of the Kampala water quality‐improving project compared, inform us
about such a project’s feasibility; this is done in this chapter. The chapter (seven) draws mainly
from the Chapters four and six. In the chapter four, theory and application of the CBA was
presented indicating how the various variables required for the methodology can be predicted
for a project being assessed. The quantification of benefits normally raises issues and for this
reason, the chapter six was dedicated to this exercise.
In this chapter (seven), all the elements of the Kampala Water quality‐improving project CBA
are combined in the same analysis to determine the project’s feasibility. The chapter is
arranged as follows:   the elements of the Kampala Water‐quality improving project CBA are
presented, the costs and benefits are presented and compared then, the decision criteria
reported. A sensitivity analysis is then conducted followed by an investigation of studies
conducted in the developing countries about feasibility of improving water quality for a
comparative assessment and conclusions drawn.
7.2. The Kampala Water Quality­Improving Project costs:
The Kampala water quality improving project costs were arrived at through interviews
conducted with focus groups consisting of engineers both at the Kampala water treatment
plant (the Gaba II plant) and from private consulting and planning managers of firms handling
water in Uganda. Also, reference to NWSC cost data was made.
It was the general consensus among the NWSC engineers that the water treatment processes
were not to blame for the water quality problems at the end‐user level in the district (1999).
The same conclusion was reached by Howard & Luyima (Ddumba 2006, Howard &Luyima 1999)
in their surveillance of urban and peri‐urban water supply in the country. They remarked that
the water treatment standards at the treatment‐plant level were in compliance with125
international portability standards. For this reason, the costs of improving the treatment
processes were deemed immaterial for the CBA.
The water quality problems of the country, and of Kampala in particular are due to
transmission and distributional flaws of the reticulation system (Kahwa 2006). For this reason
the costs of the project were deemed to be the costs anticipated in the upgrade of the
transmission and distribution system by which water is reticulated to the residents of Kampala
district.
For this purpose, the estimates of the corporate planning manager for the NWSC were used.
As such a project was currently not being considered, his estimates were approximate. He
estimated the expected cost of upgrading water systems in the country, and in Kampala
specifically, to be Ushs 1451/= per cubic meter (NWSC 2005). In the subsequent years, the cost
stream of the project would only comprise of the additional operational and maintenance
costs. An average of the operating expenses, as laid out in NWSC‐Kampala water branch’s
income statement for the year ended 30
th
June 2005, was used to approximate the project’s
operating and maintenance (OM) costs. These operating costs are shown in table 7.1 below.
Table 7.1: Operating Costs for Kampala for the years 2004 and 2005
Operating Expense 2005 Ushs ' 000 2004 Ushs ' 000
Staff Costs 5,840,757 3,464,451
Premises Maintenance 552,625 202,902
Static Plant and Pipe Network Maintenance 4,065,651 3,053,483
Transport and Mobile plant costs 429,452 158,485
Supplies and Services 1,507,370 929,429
Administrative Costs 521, 018 2,728,928
Financing Costs 10,071,567 7,676,688
Total 22,988,440 18,214,366
 
 Source: (NWSC 2005)
The OM costs for the project were estimated as a per unit cost. The NWSC branch responsible
for water supply and management for the Kampala district, produces an average of 112,862126
cubic meters of water per day (NWSC 2005). Annually, this averages to 41,646,078 cubic
meters of water. The operation and maintenance costs of the Kampala water quality
improving project was deduced to be Ushs 591.77/= per cubic meter of water.  
7.3. The Kampala Water Quality­Improving Project benefits.
The most significant benefits of improving Kampala’s water quality are related to health gains
and cost mitigation. These benefit values were estimated using the CVM. Using the
dichotomous choice framework, the questionnaire described the nature of the project and its
associated benefits. The respondents (in Kampala) were asked to vote for or against a proposal
to increase their water bills in order for them to enjoy such benefits.
The resulting benefits’ estimate was WTP for stated improvements in the quality of the water.
This benefit estimate, predicted by the WTP for the median respondent, was Ushs 385.07/=
per cubic meter of water (see chapter 6).
7.4. Discounting the Kampala Water Project
In the chapter four it was shown that the discount rate could either be the rate of time
preference in consumption, the social opportunity cost of capital, or a weighted rate. As the
water project would be a commitment of capital in one sector of the economy, it was deduced
that the social opportunity cost of capital was the more suited rate for discounting such a
project’s benefits and costs.
According to the African Development Bank (ADB) the social opportunity cost of capital in
Uganda was 12 percent in 2004 (IndexMundi 2005). Based on the above estimate, the discount
rate for the Kampala water quality‐improving project was deduced to be 12 percent.
7.5. Project life span
The benefits of the water improvement project accrue over an individual’s life span. For this
reason, the time horizon of the Kampala water quality improving project was deduced to be127
the life expectancy level of a Ugandan. The life expectancy in Uganda was estimated to be 52
years. This estimate was predicted from 2004 demographic data; according to this data, the
life expectancy was 51.67 for a male and 53.69 for a female (IndexMundi 2005).
In the analysis below the life‐span of the project is not a critical factor; but, for the purpose of
comparison the users’ life expectancy was adopted as the water quality improving project’s
useful life. The alternative way of assessing the feasibility of such a project can be the life‐cycle
assessment of the benefits of the project to the user. This method would compare all benefits
from the water quality with the total costs involved over the user’s lifespan. Therefore, the
results of the Kampala CBA whose benefits were estimated by use of the CVM would be
comparable with those of the life‐cycle assessment as the same time horizon is involved.
7.6. Decision Criteria:
All the variables required for conducting the Kampala CBA, drawn from the discussion above,
are summarized in the table 7.2 below. The decision criteria for CBA exercise were predicted
by use of using MS Excel 2007.
Table 7.2: Kampala CBA Components
Element  Estimate  Accrued at
Annual Benefits (Ushs per m3
)  385.07  End of year 1, …, 52
Once off Investment Costs (Ushs per m3
)  1451  End of year 0
Annual Operating and Maintenance costs (Ushs per m3
)  591.77  End of year 1, …, 52
Discount Rate  12%  End of year 0, …, 52
Project Life (in years)  52  -
The complete CBA outlay for the Kampala water improving project is presented in the
appendix (see appendix c).
The first decision criterion employed was the Benefit cost ratio:  128
BCR=
(PresentValueCost)
(PresentValueBenefits)
BCR=
6368.81
3200.07
BCR= 0.5
The second criterion of NPV was calculated as follows:
NPV=
( )
∑ ⎥
⎦
⎤
⎢
⎣
⎡
+
t
0
t
1 r
Present ValueBenefits ‐Present Value Cost
NPV= Ushs ‐ 3168.75/=
The third criterion‐IRR‐is calculated by the expression below:
For the Kampala Water quality improving project, the IRR was non‐positive.
The decision criteria do not support the undertaking of the Kampala water quality improving
project. From the table 7.2 it can be seen that the operational costs exceed the WTP (benefits).
By implication therefore, no matter what discount rate one employed, the project would not
be feasible. The BCR is 0.5, implying that a shilling invested in the project would yield only half
of its value in returns. The NPV is Ushs – 3168.75/= implying that investing in the project would
yield negative benefits. The IRR is non‐positive.
( )
∑
=
+
=
N
t 0
t
t t
1 IRR
B ‐ C
0129
7.7. Sensitivity Analysis
In conducting of the Kampala water quality‐improving project’s CBA it was acknowledged that
some benefits may not be fully captured in the scenario valued, for instance, benefits accruing
to non‐household users like tourists and industries (see section 1.3 chapter one). To reliably
inform decisions therefore, we must consider possible adjustments to the assumptions made
in the conducting of the Kampala CBA‐ this is the sensitivity analysis.
The first adjustment made is a 50 percent increase in the project’s benefits to account for
these omissions. At the 12 percent discount rate, the decision criteria do not support the
undertaking of the project as can be seen in table 7.3 below. The second adjustment was to
double the benefits. After this adjustment, the BCR criterion of 1 suggests that the project
would breakeven, the NPV of Ushs 31.32/= is positive but small, suggestive of meager benefits.
The IRR of 0 percent is below the hurdle rate of 12 percent, indicating that the project would
not be feasible (see table 7.3 below).  
Table 7.3: Sensitivity analysis decision criteria
Criterion
50% Benefit’s
enhancement at 12%
discount rate
100% Benefit’s enhancement at 12%
discount rate
BCR  0.75  1.00
NPV  -1568.72 31.32
 IRR  -  0.00
A secondary sensitivity analysis was made concerning the discounting of the project’s benefits
and costs. The social opportunity cost of capital rate, of 12 percent, was the rate opted for in
the discounting of the Kampala project’s benefits and costs. In this secondary sensitivity
analysis such a rate is challenged. While such a rate might be well suited for other
investments, it might not be the best rate for discounting the health benefits and costs. Such a
rate may be too high for health investment decisions. For instance, it is unlikely that
individuals would value the benefits of health 12 percent less in the subsequent year, as
implied by the Uganda social opportunity cost of capital rate of 12 percent.   130
In the new alternate project case therefore, a lower discount rate, lower than 12%, is opted
for, to make room for the possibility that users would have a futuristic preference for health
benefits. This is because there is a higher likelihood of one’s health stock declining as one ages
as was shown in the Grossman (1972) model (see chapter 2) – consequently, one desires more
health, and hence any health related benefits, as one gets older.
Furthermore, from a fiscal stand point, modern models of discounted utility theory like the
life‐cycle model and the permanent‐income hypothesis suggest that individuals would opt for
health benefits that accrue in the future. The life‐cycle model, by Modigliani et al (1954),
suggests that individuals maintain the highest smooth consumption path they can get. The
permanent‐income hypothesis by Friedman (1957), which is complimentary to the life‐cycle
model, suggests that individuals’ consumption depends on what they expect to earn over their
lives. People save during times of high earnings and dissave during times of low earnings to
smoothen out their consumption path. With the possibility of lower health stock after one has
aged (normally implying lower earnings), a futuristic preference for health benefits may be the
more realistic scenario. When this sort of intertemporal allocation is allowed for, a lower
discount rate may be used to define users’ preferences; this implies the use of a discount rate
between 0 and 5 percent.
The CBA now compares the doubled benefits (catering for possible omissions) to the costs
discounted at lower rates of 0 to 12 percent. The decision criteria of the alternate Kampala
water quality scenario redefined as above, support the undertaking of the project (see table
7.4 below).
Table 7.4: Secondary sensitivity analysis decision criteria
Criterion
                                                        Discount Rate
0% 1% 2% 3% 4% 5%
BCR 1.24 1.23 1.21 1.19 1.17 1.15
NPV  (in Ushs) 7824.24 5754.11 4282.69 3216.27 2428.12 1834.23
IRR (%) 0.12 0.11 0.10 0.09 0.08 0.07131
The BCR of the alternate scenarios is greater than one for all the six discount rates tested,
implying that the benefits of such a project would exceed its costs. Also, the NPV is positive for
all the six discount rates in the sensitivity analysis implying that the project’s cash flows would
yield positive net benefits.  The IRR is greater than the hurdle rate in all the six cases tested
implying that the project is efficient and so would be worth undertaking. Best scenario for the
water quality‐improving project is one where the discount rate is zero as this has the highest
BCR, NPV and IRR.
7.8. Feasibility studies for improving drinking water quality in Lower income
countries
Another check for the reliability of the Kampala CBA results would be to check how these
compare with those of other studies conducted in the country or in countries of similar
circumstances. Most feasibility studies conducted in LICs and East and Central Africa for that
matter are concerned with increasing water supply in these countries. The benefits and costs
of such projects are not comparable with those of a water quality improving project as the
scenarios differ.
The studies more relevant to the Kampala project case are those conducted by the UN and
WHO (2004) in assessing the achievability of the Millennium Development Goals (MDGs) and
another conducted in Uganda to assess the importance of improved water supply
infrastructure. The findings of these studies are consistent with those found in Kampala water
quality‐improving CBA.
For instance, the study investigating the achievability of the MDGs’ water target of a 50
percent reduction in the number of number of people without access to safe water and
sanitation by Markandya (2004). The study compares costs of availing safe water, with the
benefits of reduced waterborne disease burden. The benefits are measured by use of the
Disability Adjusted Life Years (DALY)  ‐ this being the total life years lost due to morbidity and
premature mortality as a result of the waterborne disease.   It takes into account the different132
probabilities of one catching a disease, the severity of that particular disease
16
and the total
sick days of the disease (Havelaar et al. 2000). Comparing the costs of providing safe water
with the benefits of reduced DALYs indicated that a 50 percent reduction in the number of
people without access to safe water and sanitation is marginally justifiable globally
(Markandya 2004). With specific reference to Africa and South Asia, provision of safe water
and sanitation is not feasible as the costs of doing so are above the upper bound of benefits of
reduced DALY (ibid). The benefits of providing just safe water reveals that the benefits would
just equal costs, if the mid point between the upper and lower bound of benefits is taken to
estimate the total benefits of reduced DALYs. However, there are no guarantees that the
benefits would not be below this midpoint. These findings are in line with those of Gadgil &
Derby (2003) who deduce that for safe water to be provided in the LICs a private‐public
partnership needs to be entered into in order to distribute the huge capital cost burden.
More relevant to Uganda, a study to investigate the importance of improved water supply
infrastructure revealed that the benefits of such a project are limited in the country. The study
was conducted on micro and small enterprises in two towns ‐ Wobulenzi, with newly improved
water supply infrastructure installed and another – Lugazi, without an improved water system.
Comparing the costs of installing new water infrastructure with benefits estimated by the
CVM, reported benefits from better water services and estimating mitigation costs of water
source switching it was found that the users were not wiling to pay for the improved services
(Davis et al. 2001).
The findings Kampala CBA are consistent with other studies conducted about the improving
water quality in developing countries
7.9. Conclusion
The methodology employed to assess the feasibility of the Kampala water quality improving
project did this by comparing the projects benefits and costs on the same annual footing, over
the project’s deemed useful life.
                                                     
16
The severity of the various diseases are given in Havelaar et al. (2000)133
The benefits were quantified through a survey based approach that approximated the users’
WTP for the project. The costs were estimated through a combination of the engineering
approach and approximating the Kampala service plant OM costs from recent data.    The
project’s useful life was the life expectancy of a Ugandan. The discount rate was the social
opportunity cost of capital in the country, as investing the water project meant foregoing an
investment in another sector of the economy.
According to the decision criteria of the CBA, the Kampala water quality improving project is
not efficient as its returns would be lower than the costs involved in its undertaking.
It is an interesting finding. The results furnish an economic reason as to why such a project has
not been undertaken in Uganda as yet  ‐ the paying population in Kampala is not willing to
support the undertaking of such a project. This is not to suggest that Kampala residents do not
recognize the benefits from improved water quality, only that their incomes are tied up in
other commitments – Uganda is one of the poorest countries in the would with a per capita
income of $300 (World Bank 2007). Some, such as UNDP (2007), argue that the poverty
measures should investigate beyond income earned and so proposing use of more
comprehensive measures like the Human Development Index (HDI) – Uganda has a low HDI of
0.505 (UNDP 2007). This ranks the country 154
th
out of the 174 poorest countries (ibid). For
this reason (insufficient household income), the users of the NWSC water are prepared to
tolerate the poor quality of the water.
The sensitivity analysis conducted enhanced the project’s benefits by 50 and 100 percent at 12
percent discount rate to account for possible omissions in the benefits’ estimation by the
scenario presented to the respondents in the actual survey. The 50 percent enhancement
made no difference; the project was still not feasible. The project scenario with doubled
benefits was weakly supported by the BCR and NPV criteria, and not supported by the IRR.
A secondary sensitivity analysis tested the doubled benefits’ scenario at a lower discount rate
to make room for the possibility that users could have futuristic preference for health benefits.
Although the results of this analysis indicated feasibility of such a project, they revealed that
the benefits of the project were marginal (with the BCR barely exceeding 1).134
Studies from developing countries and Uganda concerning the feasibility of investing in
projects to improve water quality have results consistent with the Kampala water quality‐
improving project’s CBA  ‐ these revealed that the benefits of such investments fall below the
costs involved in their undertaking. The optimistic benefits’ cases revealed that the projects
would only be marginally justifiable.  
There is a need however, to improve the country’s water quality, as was established in
introducing the problem of this research. According to the findings of this research, the only
hope such a project will be undertaken is if it is externally financed; which implies an amplified
need for NGOs, grants and donations and    aid participation in the provision of safe water
provision of the country.         135
CHAPTER EIGHT:
CONCLUSION AND
RECOMMENDATIONS136
8.1. Conclusion:
Safe water provision has lately, and indeed rightly, been brought to the forefront of the
developing countries’ development agenda. On the global scale, challenges presented by
water are those linked with the increased strain of the planet’s water resources, due to
overexploitation. The consequences of the strained water resources among many, are the
problems associated with water quality, such as diseases and high costs of abstraction
associated with basic needs’ use.
There is considerable consensus among scientists that the water challenges are exacerbated
by the changing climatic conditions. These consequences negatively impact on economies. For
instance, Barlow (2002), Bakker (2005), Hulme, Githeko & Matthies (2004) and Hunt (Uganda
1999), argue that the water challenges carry a huge economic cost, and so hamper growth
prospects of the economies concerned. The burden of the economic cost posed by the water
challenges is born particularly heavy in developing countries. Interventions in these countries
are needed to help mitigate the impact of climate change and over‐exploitation. Two of these
interventions are the improvement in the quality of the water and increase in access to (safe)
water. There is international consensus on these needs  ‐  water‐related goals have been
prioritized in international fora agendas and models of growth in developing countries have
been redefined to include safe water provision. In achieving these goals, privatisation of the
water systems may not be too attractive a solution for the developing countries as the
experience from where it has been used shows that it breeds problems of its own especially
those related to increased costs of abstraction and use of the water (Budds &McGranahan
2003). Also, privatization is likely to exclude the very poor who are in need of improved
services (ibid).
The conventional paradigm in its provision is that prescribed in the WMO (1992) Dublin
statement. The statement summarized the recommendations to the developing countries,
regarding intervention in water, in four principles known as the Dublin four principles. The
principles mainly articulated the guidelines to follow in the proper management of water. The137
fourth principle recommended that water intervention plans/projects should be economically
efficient. Such efficiency necessitates comparison of benefits with the costs of the
intervention; this defined the problem of this research.
The Ugandan water supply situation, investigated in this research, revealed that access to safe
water is still a big challenge in the country. Waterborne diseases claim a considerable number
of the labour force ‐  440 lives a week (Uganda 1999). On the other hand, the water treatment
costs keep rising (Tumwebaze 2006). In spite of the rising costs, the end product is still not safe
for consumption (Howard &Pedley 2004, Nasinyama et al. 2001, To‐Opoya 2006). Efforts, on a
small and unsustainable level, have been taken to make the water safe. For instance, the
distribution of the PuR treatment tab needs to be expanded. The majority of the users of the
NWSC water boil their water to make it safe for basic needs (Ddumba 2006, Howard &Pedley
2004, Nasinyama et al. 2001, To‐Opoya 2006). There is therefore ample evidence of the need
for safe water in the country. Why then is safe water not provided? This question motivated
the undertaking of this research.
The methodology employed to investigate the problem was CBA. It compared the anticipated
project benefits with anticipated costs. The variables required for conducting the CBA were
defined after structured and unstructured interviews had been conducted as well as reference
made to the literature on the application of CBA. The costs were obtained by consulting the
planers and engineers in the Uganda water systems and through reference to NWSC data. The
benefits were harder to estimate as they had few behavioral traits for reference to be made.
The technique selected to do this job was the CVM. The CVM was conducted in Kampala, in
accordance with as many NOAA panel guidelines as possible.
Preliminary findings did not support the undertaking of the Kampala water quality improving
project. The operating costs were above the users’ WTP for the project. At any discount rate
employed, the project would not be economically and financially feasible. The BCR (0.5) and
NPV (Ushs – 3168.75/=) were below the minimum thresholds of 1 and 0 respectively. Based on
the results of the decision criteria, it was concluded that the project would not be feasible.138
The above findings provide an indication why an intervention in safe water provision has not
been undertaken in Kampala ‐ the paying population (NWSC water users) is unable to support
the undertaking of such a project. They recognize the merits of improved water quality, as can
be deduced from their treatment of the water before using it for basic needs, but their
incomes are insufficient. According to the World Bank (2007), Uganda is one of the poorest
countries in the world with a per capita income of $ 300 and one of the lowest HDI (UNDP
2007). For this reason, they are obliged to tolerate the poor quality water. It is deduced that
financing such a project should be according to a model that does not lay the full burden of the
project’s costs on the users, i.e. that external support be elicited.    
Alternatively, when we double the benefits to account for possible omissions by the Kampala
project scenario presented to the respondents, these are such as benefits that would accrue to
non‐household users, and lower the discount rate to allow for a possible futuristic preference
for health benefits by the water users we have positive net benefits. Even though these
benefits are marginal, they would justify the undertaking of such a water quality‐improving
project.
8.2. Recommendations        
a) One of the reliability tests identified in the valuation of the projects benefits was
convergent validity. Owing to limited resources and time this test could not be
conducted. To this end, this study recommends that other valuation techniques such as
the choice modeling or hedonic pricing be conducted in the estimation of the project’s
benefits to establish reliability of this research’s estimates.
b) Establishing of the quality problem in the Kampala water supply system was done
solely based on the advice of the water suppliers (NWSC). This study recommends that
independent scientists be employed to asses the country’s water quality problems to
avoid any tinge of biasness in the projects valuation.
c) A snow type analysis would be important in identifying what the true cause of the
quality deterioration is.139
d) Since the country as a whole is experiencing water quality problems, the feasibility
analysis for improving water quality on a national scale should be conducted.
e) The long term solution for the country’s water quality problems lie in prevention rather
than fixing the current problems. While emphasizing the need to fix current water
problems in the country more work has to be done in trying to establish proper waste
disposal and proper sanitation services in the country.
f) In the same breath, work has to be done in trying design a policy that prevents further
pollution of the main source for the country’s water Lake Victoria. Otherwise, the
benefits forecasted in this research could be erased by the higher costs of treatment.  140
REFERENCES
ABAYAWARDANA, S. & HUSSAIN, I. (2003) Water, health and poverty linkages: A case study from Sri
Lanka. Dhaka, Senegal  Paper Presented at Asian Development Bank Regional Consultaion Workshop on
Water and Poverty, 22‐26
th
September, [Online] Available from:
http://www.adb.org/Water/Actions/SRI/SRI_water_health_poverty.pdf, [Accessed 15
th
June 2006].
ABRAMS, L. (2001) Water Scarcity, [Online] Available from:
http://www.africanwater.org/drought_water_scarcity.htm, [Accessed 20
th
November 2007].
ADF (2005) Appraisal Report Rural Water Supply And Sanitation Program Uganda, [Online] Available
from: www.afdb.org/pls/portal/url/ ITEM/084B449D5E817267E040C00A0C3D4328, [Accessed 5
th
June
2006].
AKOL, G. (2006). Personal Communication. Kampala: GIS officer, Kampala Water NWSC.
ALBERINI, A. (1995) Testing Willingness‐to‐Pay Models pf Discrete Choice Contingent Valuation Survey
Data. Land Economics, Vol 71, no, Pg 83‐95.
ALLGOOD, G., COWAL, S. & CARRIER, L. (2005) Home water Treatment Launched in Uganda, [Online]
Available from: http://www.pghsi.com/releases/home%20water%20treatment%20uganda1.pdf,
[Accessed 14
th
March 2005].
AQUASTAT (2005) FAO's information system on water and agriculture ‐ Uganda., [Online] Available
from: http://www.fao.org/ag/ag.index.stm, [Accessed 4
th
May 2006].
ARGOSS (2002) Assessing Risk to Groundwater from On‐site Sanitation: Scientific Review and Case
Studies. British Geological Survey Commissioned Report, CR/02/079N. London, National Environmental
Research Council.141
ARROW, K., SOLOW, R., PORTNEY, P., LEAMER, E., RADNER, R. & SHUMAN, H. (2002) Appendix I: Report
of the NOAA Panel on Contingent Valuation. IN HARBERGER, A. C. J., G. P (Ed.) Cost‐Benefit Analysis.
Cheltenham, Edward Elgar.
B‐RAO, C. (2001) Sample Size Considerations in Genetic Polymorphism Studies. International Journal of
Human and Medical Genetics, Vol 52, no, Pg 191‐200.
BAKKER, K. (2002) Water Management Issues from Commons to Commodity? Privatising and
Commercialising water Supply, [Online] Available from: www.sfu.ca/cstudies/science/water/pdf/water‐
Ch12, [Accessed 24
th
September 2005].
BARLOW, M. (2001) Blue Gold: The global water crisis and commodification of the world's water supply,
[Online] Available from: http://www.thirdworldtraveler.com/water/Water/Blue_Gold.html, [Accessed
23
rd
March 2006].
BARUNGI, A. (2003) Contracts and Commerce in Water Services in Uganda: Impact of the Private Sector
Participation on the rural poor in Uganda, [Online] Available from: www.wateraid.org, [Accessed 20
th
April 2005].
BGS (2001) Groundwater Quality: Uganda., [Online] Available from: http://www.wateraid.org> other/
startdownload.asp?openType=forced&documentID=199, [Accessed 12
th
March 2005].
BINGHAM, P., VERLANDER, N. Q. & CHEAL, M. J. (2004) John Snow, William Farr and the 1849 outbreak
of cholera that affected London: a reworking of the data highlights the importance of the water supply.
Public Health, Vol 118, no, Pg 387‐394.
BIRD, A. (2005). Personal Communication. Port Elizabeth: Microbiologist  NMMM.
BLACKMAN, R. (2001) Technology, Drinking water and poverty, [Online] Available from:
http://www.sussex.ac.uk/units/PRU/notes4.pdf, [Accessed 27
th
September 2007].
BRENT, R. J. (2003) Cost‐Benefit Analysis and Health Care Evaluations, Cheltenham, Edward Elgar.142
BROUWER, R. & PEARCE, D. W. (2005) Introduction. IN BROWER, R. & PEARCE, D. W. (Eds.) Cost‐Benefit
Analysis and Water Resources Management. Cheltenham, Edward Elgar.
BUDDS, J. & MCGRANAHAN, G. (2003) Are the Debates on Water Privatization Missing the Point?
Experiences from Africa, Asia, Latin America. Environment and Urbanisation, Vol 15, no, Pg 87‐113.
BYAMUKAMA, D., KANSIIME, F., MACH, R. L. & FARNLEITNER, A. H. (2000) Determination of Eschericha
Coli Contamination with Chromocult Coliform Agar Showed High Level of Discrimination Efficiency for
Differing Fecal Pollution levels in Tropical Waters of Kampala, Uganda., [Online] Available from:
http://aem.asm.org/cgi/content/full/66/2/864, [Accessed 18
th
March 2005].
COMMONWEALTH‐OF‐AUSTRALIA (2006) Introduction to Cost‐Benefit Analysis and Alternative
Evaluation Methodologies: Financial Management Reference Material No. 5., [Online] Available from:
www.ag.gov.au/cca, [Accessed 1
st
July 2006].
CONSTANTINIDES, G. (2000) Cost‐Benefit Analysis case studies in Eastern Africa for the GPA strategic
action plan on sewage., [Online] Available from: www.gpa.unep.org, [Accessed 12
th
June 2005].
CRC (2003) All About Australia's Drinking Water., [Online] Available from: www.waterquality.crc.org.au,
[Accessed 4
th
July 2006].
DARDANONI, V. (1986) A note on a simple Model of Health Investment. Bulletin of Economic Research,
Vol 38, no, Pg 97‐100.
DAVIS, J., KANG, A., VINCENT, J. & WHITTINGTON, D. (2001) How important is Improved water
infrastructure to Microenterprises? Evidence from Uganda. World Development, Vol 29, no, Pg 1753‐
1767.
DDUMBA, F. (2006). Personal Communication. Kampala: Aqua Consult, Civil Engineer; In Charge of Civil
works Gaba III Treatment Works.143
DOLAN, P. (2003) Grossman’s theory of the demand for health care, [Online] Available from:
http://www.oekonomi.uio.no/info/EMNER/h03/grossman.ppt#256,1,Grossman’s theory of the demand
for health care, [Accessed 17
th
November 2007].
ECOSYSTEM‐VALUATION (2007) Contingent Valuation Method, [Online] Available from:
http://www.ecosystemvaluation.org/contingent_valuation.htm, [Accessed 18
th
June 2007].
FAO (2002) Applications of The Contigent Valuation method in developing Countries, [Online] Available
from: http://www.fao.org/docrep/003/X8955E/X8955e00.HTM, [Accessed 04
th
July 2005].
FARR, W. (1852) Influence of Evelation on the fatality of Cholera. Journal of the Statistical Society of
London, Vol 15, no, Pg 155‐183.
FOLLAND, S., GOODMAN, A. C. & STANO, M. (1997) The Economics of Health and Health care, Upper
Saddle River, N.J, Prentince Hall.
GADGIL, A. J. & DERBY, E. A. (2003) Providing safe drinking water to the 1.2 Billion Unserved People,
[Online] Available from: http://eetd.lbl.gov/IE/pdf/LBNL‐52374.pdf, [Accessed 20
th
November 2007].
GARSON, D. G. (2006) Logistic Regression, [Online] Available from:
http://www.statisticssolutions.com/Logistic_Regression.htm, [Accessed 9
th
Febuary 2007].
GHASSEMI, F., JAKEMAN, A. J. & NIX, H. A. (1995) Salinisation of Land and Water Resources: Human
Causes, Extent, Management, CAB International.
GOPE, P. C. (2002) Determination of Minimum Number of Specimens in S‐N Testing. Journal of
Engineering Material and Technology, Vol 124, no, Pg 421‐427.
GREENE, W. H. (2000) Econometric Analysis, New Jersey, Prentice Hall.
GROSSMAN, M. (1972) On the Concept of Health Capital and the Demand for Health. The Journal of
Political Economy, Vol 80, no, Pg 223‐255.144
HAAB, T. C. & MCCONNELL, K. E. (1997) Referundum Models and Negative Willingness to Pay:
Alternative Solutions. Journal of Environmental Economics and Management, Vol 32, no, Pg 251‐270.
HAAB, T. C. & MCCONNELL, K. E. (2004) Valuing Environmental and Natural Resources: The Econometrics
of Non‐Market Valuation, Cheltenham, Edward Elgar.
HAMOUDI, A. A. & SACHS, J. D. (1999) Economic Consequences of Health Status: A Review of the
Evidence, [Online] Available from: http://www2.cid.havard.edu/cidwp/030.pdf, [Accessed 25
th
August
2006].
HANEMANN, M. W. (2002) Valuing The Environment Through Contingent Valuation. IN HARBERGER, A.
C. & JENKINS, G. P. (Eds.) Cost‐Benefit Analysis. Cheltenham, Edward Elgar.
HANLEY, N. & SPASH, C. (1993) Cost‐Benefit Analysis and the Environment, Aldershot, Edward Elgar.
HARUNA, R., EJOBI, F. & KABAGAMBE, E. K. (2005) The quality of water from protected springs in Katwe
and Kisenyi parishes, Kampala city, Uganda. African Health Sciences, Vol 5, no, Pg 14‐20.
HASSAN, R. (2006) Climate change and African agriculture. Policy Note No.30, August 2006 CEEPA.
HAVELAAR, A. H., DE‐WIT, M., VAN‐KONINGSVELD, R. & VAN‐KEMPEN, E. (2000) Health burden in the
Netherlands due to infection with thermophilic Campylobacter spp. Epidemiol Infect, Vol 125, no, Pg
505‐522.
HOWARD, G. & LUYIMA, P. (1999) Urban and Peri Urban Water Supply Surveillence Programme: Report
on water supply surveillence in Ten selected urban areas of Uganda., [Online] Available from:
www.lboro.ac.uk/watermark/reports/ uganda‐national‐report‐english.pdf, [Accessed 6
th
February 2005].
HOWARD, G. & PEDLEY, S. (2004) Quantitative Microbial Assessment Evaluation of WSPS: Qualitative
Risk Assesment for Piped Water Supplies in Uganda., [Online] Available from: www.lboro.ac.uk,
[Accessed 17
TH
March 2005].145
HULME, M., GITHEKO, A. & MATTHIES, F. (2005) Climate change and infectious diseases in Africa. IN
OFFICE, O. S. A. T. (Ed.) In Infectious diseases in Africa: using science to fight the evolving threat.
Department of Trade and Industry (UK).
HUNT, C. E. (2004) Thirsty Planet: Strategies for Sustainable Water Management, New York, Zed Books.
INDEXMUNDI (2005) Uganda Life Expectancy at Birth, [Online] Available from:
http:/www.indexmundi.com/uganda/life_expectancy_atbirth.html, [Accessed 23
th
July 2006].
KAHWA, D. I. (2006). Personal Communication. Kampala: Manager Corporate Planning, NWSC Corporate
Planning Department.
KIRBY, A. (2004) Water Scarcity; A looming Crisis? [Online] Available from: http://www.bbc.co.uk,
[Accessed 19
th
October 2004].
KLOHN, W. & ANDJELIC, M. (1997) Lake Victoria: A Case in International Cooperation., [Online] Available
from: www.fao.org/ag/AGLW/webpub/lakevic/LAKEVIC4.htm, [Accessed 2
nd
June 2006].
KRAVCIK, M., HRONSKY, J., TESLIAR, J. & ZVARA, R. (2002) The New Theory of The Global Warming,
[Online] Available from: http://www.changenet.sk/peopleandwater/spava.stm?x=1964, [Accessed 12
th
October 2005].
KULABAKO, R. (2005) Analysis of the impact of anthropogenic pollution on shallow groundwater in peri‐
urban Kampala. Stockholm: Land and Water Resource Engineering.
KULABAKO, R., NALUBEGA, M. & THUNVIK, R. (2004) Charaterization of Peri‐ Urban Anthropogenic
Pollution in Kampala, Uganda. Vientiane, Laos  Paper Presented at 30th WEDC International Conference,
[Online] Available from: http://wedc.lboro.ac.uk/conferences/pdfs/30/Kulabako.pdf, [Accessed 24
th
May 2007].
LANE, D. (2005) Introduction to Central Tendency, [Online] Available from:
http://cnx.org/content/m10942/latest/, [Accessed 20th August 2007].146
LEHTOLA, M. (2002) Microbially available phosphorus in drinking water, Mannerheimintie, National
Public Health Institute.
LULE, J. R., MERMIN, J., EKWARU, J. P., MALAMBA, S., DOWNING, R., RANSOM, R., NAKANJAKO, D.,
WAFULA, W., HUGHES, P., BUNNELL, R., KAHARUZA, F., COUTINHO, A., KIGOZI, A. & QUICK, R. (2005)
Effect of home‐based water chlorination and safe storage on diarrhea among persons with Human
Immunodeficiency Virus in Uganda. The American Society of Tropical Medicine and Hygiene, Vol 73, no,
Pg 926‐933.
LWASA, S. (2005) Urban Expansion Process of Kampala In Uganda: Perspectives on Contrasts with Cities
of Developed Countries, [Online] Available from:
http://www.populationenvironmentresearch.org/seminars112004.jsp, [Accessed 5
th
October 2007].
MARGLIN, S. A. (2002) The Social Rate of Discount and the Optimal Rate of Investment. IN HARBERGER,
A. & JENKINS, G. P. (Eds.) Cost‐Benefit Analysis. Cheltenham, Edward Elgar.
MARKANDYA, A. (2004) Water quality Issues in Developing Countries, [Online] Available from:
www.dse.unibo.it/seminari/markandia.pdf, [Accessed 20
th
November 2007].
MARSDEN, R. (1995) John Snow and the Broad Street Cholera Outbreak of 1854, [Online] Available from:
http://www.winwaed.com/sci/cholera/john_snow.shtml, [Accessed 26
th
Febuary 2005].
MIRABELLA, J. (2003) Sampling for Success, [Online] Available from:
http;//www.drjimmirabella.com/conferences/GSA‐Sampling.pdf, [Accessed 11
th
Febuary 2007].
MOH (1999) Health Statistics, [Online] Available from: http://www.heath.go.ug/summary_chart.htm,
[Accessed 28
th
Feb 2007].
MUGABE, D. & KISAMBIRA, E. (2006) Lake Victoria Levels at Jinja Dam raise Eyebrows., [The East African
Business Week. Monday 16
th
January 2006
Online] Available from:
http://www.businessweek.com/index.php?option=com_content&task=view&id=921&itemid=9,
[Accessed 24
th
May 2007].147
MUGUME, R. (2006). Personal Communication. Kampala: NWSC, Cheif Overseer Gaba II Water
Treatment Plant.
MURRAY, C. J. & LOPEZ, A. D. (1996) Global Burden of Disease. Geneva, World Health Organisation.
MWLE (2004) The Water and Sanitation Sector Performance Roport 2004. Kampala, Ministry of Water
Lands and Environment (Uganda).
NAM, P. K. & SON, T. V. H. (2004) Household Demand for improved water services in Ho Chi Minh City: A
comparison of contingent Valuation and Choice modeling Estimates, [Online] Available from:
http://een.anu.edu.au/e05prpap/pham.doc, [Accessed 3
rd
May 2005].
NASINYAMA, WILSON, TOEWS, GYLES & ASIBO (2001) Risk factors for Acute Diarrhoea Among
inhabitants of Kampala district Uganda. South African Medical Journal, Vol 90, no, Pg 891‐898.
NDEGE, M. (1996) Strain, Water Demand, and Supply Directions in the Most Stressed Water Systems of
Eastern Africa. IN EGLAL RACHED, E. R. B., D. B. (Ed.) WATER MANAGEMENT IN AFRICA AND THE MIDDLE
EAST: Challenges and Opportunities.
NEMA (1996) State of the Environment Report for Uganda 1996, [Online] Available from:
http://easd.org.za/Soe/Uganda/augindx.htm, [Accessed 28
th
May 2007].
NEMA (1998) District State of Environment Report 1997, [Online] Available from:
http://www.nemaug.org, [Accessed 16
th
March 2005].
NEMA (2000) State of the Environment Report for Uganda, 2000/2001., [Online] Available from:
http://www.nemaug.org/uploads/soe/soe2000.pdf, [Accessed 10
th
June 2006].
NEWSOM (2005) Data Analysis II: Logistic Regression. Published Class Notes, [Online] Available from:
http://www.cas.lancs.ac.uk/short_courses/notes/inter_spss/session6.pdf, [Accessed 20
th
August 2007].148
NOAA‐COSTAL‐SERVICES‐CENTER (2006) Restoration Economics: Discounting and Time Preference,
[Online] Available from: http://www.csc.noaa.gov/coastal/economics/discounting.htm, [Accessed
8
th
September 2006].
NTAMBIRWEKI, J. & DRIBIDU, E. (1999) The Legal and Policy instruments for Sustainable Management of
Natural Resources in Uganda. The Case of Implementation of Uganda Water Action Plan., [Online]
Available from: http://www.cdr.dk/sscafrica> /nta &drib.htm, [Accessed 14
th
February 2005].
NWSC (2004) National Water and Sewerage Corporation Annual Report 2003/2004. Kampala, National
Water and Sewerage Corporation.
NWSC (2005) National Water and Sewerage Corporation Annual Report 2004/2005. Kampala, National
Water and Sewerage Corporation.
NYAKAANA, J. B., SENGENDO, H. & LWASA, S. (2007) Population Urban Development and the
Environment in Uganda: The Case of Kampala city and its Environs, [Online] Available from:
http://pripode.cicred.org/IMG/pdf_Popul‐urb‐dev‐Nyak‐Mar‐2007‐UG.ppt.pdf, [Accessed 5
th
October
2007].
O'CONNELL, A. A. (2006) Logistic Regression Analysis: Model fit, [Online] Available from:
http://fuculty.education.uconn.edu/epsy/aoconnell/webpage%20440%20files/2b.Logreg.modelfit.ppt,
[Accessed 9
th
Febuary 2007].
O'DOHERTY, R. (1993) The Contingent Valuation Method. CSERGE Working Paper PA 93‐01.
OBURU, P. (2006). Personal Communication. Kampala: Network Engineer, NWSC Kampala Water.
OYOO, R. (2006). Personal Communication. Kampala: Ag. Quality Assurance Manager, NWSC Kampala
Water.
PAGE, J., BIRDSALL, N., CAMPOS, E., CORDEN, W. M., KIM, C.‐S., PACK, H., SABOT, R., STIGLITZ, J. & UY,
M. (2004) The making of the East Asia Miracle. World Bank Policy Research Bulletin, Vol 4, no, Pg 1‐5.149
PEARCE, F. (2006) Uganda Pulls Plug on Lake Victoria, [New Scientist Online] Available from:
http://www.newscientist.com/article.ns?id=mg18925384.100, [Accessed 17
th
April 2007].
POSTEL, S. (2002) Water and Sustainability: Dimentions of Global Challenge, [Online] Available from:
www.sfu.ca/cstudies/science/water/pdf/water‐Ch01.pdf, [Accessed 28
th
September 2005].
RAYMER, D. (2005). Personal Communication. Port Elizabeth: Assistant Manager Water Management of
Bulk Water; NMMM.
REES‐GILDEA, P. & GELETA, B. (1999) Uganda: Cholera, [Online] Available from:
http://www.ifrc.org/docs/appeals/rpts99/ug001.pdf, [Accessed 12
th
October 2005].
REYNOLDS, C. (2005) Low Water Levels observed on Lake Victoria, [Online] Available from:
http//www.fas.usda.gov/pecad/highlights/2005/09/uganda_26sep2005, [Accessed 17
th
April 2007].
RUUD, M. & KOOTEN, G. C. V. (1998) How Accurate are Dichotomous Choice Contingent Valuation
Welfare Measures when Agents have Heterogeneous Preferences? [Online] Available from: http://sfm‐
1.biology.ualberta.ca/english/pubs/PDF/WP_1998‐15.pdf, [Accessed 12
th
June 2005].
SAMSON, P. & CHARRIER, B. (1997) International Freshwater Conflict: Issues and Prevention Strategies.
Geneva, Green cross international.
SCHWAB, J. (2007) Multinomial Logistic Regression Basic Relationships. Published class notes, [Online]
Available from:
http://www.utexas.edu/courses/schwab/sw388r7/SolvingProblems/MultinomialLogisticRegression_BasicRelations
hips.ppt, [Accessed 23
rd
August 2007].
SELVARANTNAM, R. R., DESILVA, L. D., PATHMESWARAN, A. & DESILVA, N. R. (2003) Nutritional Status
and Productivity of Sri Lankan tea pluckers. The Ceylon Medical Journal, Vol 48, no, Pg 114‐118.
SHIMADA, J., TANG, C., TANAKA, T., YANG, Y., SAKURA, Y., SONG, X. & LUI, C. (2000) Irrigation caused
ground water drawdown beneath the North Chaina Plain, [Online] Available from: http://dbx.cr.chiba‐
u.jp/NCP/papers/shimada_et_al2000.pdf, [Accessed 27
th
March 2007].150
SNOW, J. (1855) On the Mode of Communication of Cholera, London, John Churchill, New Burlington
Street.
SUTHERLAND, J. (2006) Uganda Team Manual, [Online] Available from: www.rmni.org/1/team‐
manual.html ‐ 99k, [Accessed 30
th
October 2007].
TIETENBERG, T. (2000) Environmental and Natural Resource Economics, New York, Addison Wesley.
TO‐OPOYA, H. (2006). Personal Communication. Kampala: Production Services Manager, Coca‐Cola
Century Bottling Company Limited.
TUMWEBAZE, S. (2006). Personal Communication. Kampala: NWSC, Senior Laboratory technician Gaba
II.
UCL (2005) Techniques of Environmental Valuation: Lecture 4, Contingent Valuation, [Online] Available
from: http://www.homepages.ucl.ac.uk/~uctpb20/TEV_Lecture4.htm, [Accessed 19
th
July 2006].
UDHHS (2007) Statistical Sampling: Techniques for effective management of Programs, [Online]
Available from: http://www.acf.hhs.gov/programs/cse/pubs/2002/tempo/statistical_sampling.html,
[Accessed 3
rd
Febuary 2007].
UGANDA (1999) A National Water Policy, Ministry of Water Lands and Environment, Kampala.
UN‐HABITAT (2007) Bukoba Fact Sheet, [Online] Available from:
http://www.unhabitat.org/content.asp?cid=3556&catid=462&typeid=24&subMenuId=0, [Accessed 22
nd
May 2007].
UN (2002) Report of the World Summit on Sustainable Development. New York, UN.
UN (2006) Water: A Shared Responsibility ‐ The United Nations World Water Development Report 2.,
[Online] Available from: unesdoc.unesco.org/images/0014/001431/143120E.pdf, [Accessed 3
rd
June
2006].151
UNDP (2007) Uganda: The Human Development Index ‐ going beyond income, [Online] Available from:
http://hdrstats.undp.org/countries/country_fact_sheets/cty_fs_UGA.html, [Accessed 2
nd
December
2007].
UNESCO (2003) International Year of Freshwater, [Online] Available from:
http://www.wateryear2003.org, [Accessed 13
th
May 2005].
UNICEF (1999) State of the World's Children. New York, United Children's Fund (UNICEF).
USAID (2005) USAID Budget, [Online] Available from: http://www.usaid.gov/policy/budget/cbj
2005/afr/ug.html, [Accessed 12
th
March 2005.].
UTB (2007) Travel Tips, [Online] Available from: http://www.visituganda.com/tips.html, [Accessed 20
th
November 2007].
VAUGHAN, W. J. & DARLING, A. H. (2000) The Optimal Sample Size for Contingent Valuation Surveys:
Applications to Project Analysis, [Online] Available from: http://www.mdb‐egp.net/sds/doc/ENV‐136‐
WVaugman‐E.pdf, [Accessed 15
th
October 2005].
WALL, K. (2000) A Résumé of World Bank water and Sanitation Experience of Value to South Africa.
Pretoria, Water Research Commission.
WATKINS, T. (2003) An Introduction To Cost Benefit Analysis, [Online] Available from:
www.sjsu.edu/faculty/watkins/cba.htm, [Accessed 6
th
September 2006].
WHO (2001) IDS/Health Information Bulletin No.1, [Online] Available from:
http://www.cdc.gov/idsr/files/CSR_bulletin_June_01/CSR_bulletin_June_01.pdf, [Accessed 8
th
October
2007].
WHO (2004) Guidelines for Drinking ‐Water Quality. Vol 1, 3rd Ed. Geneva, World Health Organisation.
WHO/UNICEF (2005) Water for Life: Making it Happen, [Online] Available from:
http://www.who.int/water_sanitation_health/waterforlife.pdf, [Accessed 20
th
July 2006].152
WIKIPEDIA (2007a) Economic Growth, [Online] Available from:
http://en.wikipedia.org/wiki/Econmic_growth, [Accessed 5
th
March 2007].
WIKIPEDIA (2007b) Intertropical Convergence Zone, [Online] Available from:
http://en.wikipedia.org/wiki/Intertropical_Convergence_Zone, [Accessed 24
th
March 2007].
WIKIPEDIA (2007c) Lake Victoria, [Online] Available from: http://en.wikipedia.org/wiki/Lake_Victoria,
[Accessed 17
th
April 2007].
WIKIPEDIA (2007d) Uganda, [Online] Available from: http://en.wikipedia.org/wiki/Uganda, [Accessed
24
th
March 2007].
WIKIPEDIA (2007e) Water Privatisation, [Online] Available from:
http://en.wikipedia.org/wiki/Water_privatization, [Accessed 24
th
March 2007].
WINTER, P. (2005). Personal Communication. Port Elizabeth: Senoir Lecturer at the Department of
Zoology NMMU.
WMO (1992) The Dublin Statement on Water and Sustainable Development., [Online] Available from:
http://www.wmo.ch/web/homs/documents/english/icwedece.html, [Accessed 5
th
June 2006].
WORLD‐BANK (2007) Country Brief, [Online] Available from:
http://web.worldbank.org/WBSITE/EXTERNAL/COUNTRIES/AFRICAEXT/UGANDAEXTN/0,menuPK:37494
7~pagePK:141132~piPK:141107~theSitePK:374864,00.html, [Accessed 2
nd
December 2007].
WSSCC (2000) Vision 21: A shared vision of hygiene sanitation and water supply and a framework for
action. Geneva, Water Supply and Sanitation Collaborative Council (WSSCC).
WWAP (2006) National Water Development Report: Uganda. Geneva, United Nations‐Water.
WWC (2005) Water at a Glance, [Online] Available from:
http://www.worldwatercouncil.org/index.php/id=5, [Accessed 26
th
October 2005].153
YAHYA, A. (2004) Klaten Farmers Say No to AQUA‐Danone Bottled Water Company, [Online] Available
from: http://www.eng.walhi.or.id/kampanye/air/privatisas/klaten_aqua/, [Accessed 14
th
September
2005].
ZEHNDER, A. J. B., YANG, H. & SCHERTENLEIB, R. (2003) Water issues: the need for action at different
levels. Aquatic Science, Vol 65, no, Pg 1‐20.
ZON, A. & MUYSKEN, J. (2003) Health as A Principle Determinant of Economic Growth, [Online] Available
from: http://edata.ub.unimaas.nl/www‐edocs/loader/file.asp?id=825, [Accessed 13
th
June 2005].
  154
APPENDICES155
Date of interview
Appendix A: The Survey Instrument
Contingent Valuation Questionnaire for A Water Quality Improvement Project for the
Kampala District of Uganda
This is a survey conducted in order to collect data for my research project as a requirement for
the completion of my Mcom (Economics) at the Nelson Mandela Metropolitan University. My
research deals with assessing the feasibility of improving Kampala’s water quality. This survey is
conducted to determine how much money people in Kampala would be willing to pay for a
project to improve the quality and safety of water reticulated to them by the National Water
and Sewerage Corporation (NWSC). Your response will be treated with the strictest of
confidence. The results of this survey will be presented in a way that precludes your identity
from being connected with the survey.
a). Name of interviewer. ______________________
b). Locality of the Interview. ___________________________
c). Respondent’s gender
Male                      Female
Bio Data Section:
1. How old are you? ______Years (Tick against the list)
1.1 20 ‐ 25
1.2 26 ‐ 30
1.3 31 ‐ 35
1.4 36 ‐ 40
1.5 41‐ 45
1.6 46 ‐ 50
1.7 51 ‐ 55
1.8 56 ‐ 60
1.9 61 ‐ 65
1.10 Above 65
No:156
2. What is the highest qualification you have attained?
2.1 No formal education
2.2 Primary Leaving Certificate (PLE)
2.3 O‐Levels Certificate (UCE)
2.4 A‐Level Certificate (UACE)
2.5 National Diploma
2.6 University Degree
2.7 Post Graduate Degree
3. What was your Pre‐tax income last year? (Include all forms of remuneration from the
government and from other institutions)
          (In Ugandan Shillings)
3.1 0
3.2 1‐100,000
3.3 100,001‐500,000
3.4 500,001‐1,000,000
3.5 1,000,001‐1,500,000
3.6 1,500,001‐2,000,000
3.7 2,000,001‐2,500,000
3.8 2,500,001‐3,000,000
3.9 3,000,001‐3,500,000
3.10 3,500,001‐4,000,000
3.11 Above  4,000,001157
4. How many people make up your household? ____________
5. What is Your Occupation?
5.1 Civil servant in Government
5.2 Employee in Non‐Governmental Franchise
5.3 Private entrepreneur, Businessman
5.4 Agriculturalist (Crop and animal husbandry, fisherman)
5.5 Hired manual Labourer
5.6 Other Specify
6. In the last 5 years, which of the following water related diseases affected any
member of your family?(Tick as many against the list)
6.1 Cholera
6.2 Dysentery
6.3 Typhoid and Para Typhoid Fever
6.4 Acute and persistent diarrhoea
6.5 Other (Specify)158
7. How many people of this household did the diseases above affect and for how long?
_________ People, ill for ________days.
8. Approximate how much it cost your household to treat the diseases for all the
members in the following years?
Year Cost (In Ugandan Shillings
8.1 2005  
8.2 2004
8.3 2003
8.4 2002
8.5 2001
9. What is your primary source of water for domestic use?
9.1 NWSC water in house connection
9.2 NWSC tap water at <1km away
9.3 Protected well
9.4 Storm/Rain water
9.5
Open water sources (Streams, the lake and unprotected
wells)
9.6 Other Specify159
10. Which of these applies to you? (Question applies if answer is any in 9 above except
9.1).
Storage Capacity (in litres) Cost (in Shillings)
10.1 Store water in an Overhead Tank or
underground    
10.2 Store water in a Drum    
10.3 Store water in a Collection vessels
(Jerry‐cans, Buckets)    
10.4 I use Bottled water    
10.5 Others(Specify)    
11. What was your average monthly water bill in the last 6 months?
Sh_____________/= (Question applies if respondent selected 9.1)
12. How much did it cost you in the last 6 months to collect water from a different
source? Sh.____________./=
(Question 12 and 13 apply to all who use other water sources but have to pay for this
water, even NWSC clients who supplement their water usage from other sources; include
cost such as fuel costs the water vendor’s charge?).
13. How long in hours, does it take you to collect water from other sources?_______ Hrs
a day
14. How would you rate the quality of NWSC reticulated water in terms of these
aspects? (Please tick one provision for every amenity)160
Amenity Excellent Very good Good Fair Poor Very Poor
14.1 Taste
14.2 Colour
14.3 Odour
14.4 Microbial Safety
Averted Costs Section:
15. Do you use bottled water?
      Yes                                No
16. If your answer is yes in 15 above, what quantity of bottled water do you use monthly
and how much does it cost you? ________________________
17. If your answer is no in 15 above, how do you treat your water to make it drinkable?
17.1 Boiling
17.2
Add Chemicals (E.g. PuR Purifier)
17.3 Other (Specify)
Proceed to question 18 if above answer is 17.1
18. How long does it normally take you to boil water and what heating system do you
most frequently use? ______ Hrs, with: (tick next to appropriate box)  161
Firewood            Charcoal           Electricity            Gas
Proceed to question 19 if above answer is 17.2
19. What chemicals do you use every month in the treatment of your drinking water and
in what quantities? _________________
Proceed to question 20 if above answer is 17.3
20. What processes and costs are involved in the other method?
_____________________________________________________________
Contingent Valuation Section:
As you may well be aware, the NWSC reticulated water in Kampala is not of a safe
drinkable standard. It is for this reason that people treat the water by boiling, adding
chemicals, using bottled water, etc. For a long time, this has been the state of affairs
in Kampala.   The danger of letting this state continue is that people’s health suffers
and hence their productivity. According to the Uganda Ministry of Health, diarrhoeal
diseases and other water borne diseases, such as cholera, acute and persistent
diarrhoea, dysentery, typhoid and Para typhoid fever among others, account for 75%
of life years lost to premature death in the country (WHO 2004). The majority of the
people who contract these diseases do not lose their lives but suffer debilitating ill
health. In this respect, water borne diseases impose a financial burden, in medical
expenses, that weighs heaviest on the low income earners who constitute majority of
the Ugandans.
The diseases in question are spread by pathogens such as: vibrio cholerea, typhoid,
Shigella, Rota Virus, to mention but a few, which are transmitted through drinking
water.   These pathogens find there way into the reticulation system from the water162
source. If the treatment process is not thorough enough to kill these coliforms,
consumers of the water are at the risk of suffering from one of the water borne
diseases. Also, if the reticulation system is improperly functioning consumers are at
risk.
This research proposes a project to upgrade Uganda’s tap water to class 0 quality, i.e.
water drinkable straight from the tap and posing no risk of any disease or mineral
deficiency to you the consumer. This upgrade will involve increasing effectiveness of
water treatment and replacing faulty joints and pipes in the system. The benefits for
you are lower risk of disease associated with water like Cholera, Typhoid etc, and
avoided expenditure that would otherwise be incurred to make the water drinkable.
They are costs such as involved in boiling water, buying bottled water, or buying
chemicals to treat your water. The other form of avoided expenditure is in terms of
medical expenses in treating water‐borne diseases.
In addition, being the first country in East and central Africa to upgrade its water to
drinkable levels, Uganda’s tourism industry could be promoted owing to the
increased water health safety for tourists. To be able to implement a project that
upgrades and maintains the reticulated water in your area to a drinkable standard,
funds have to be raised from the Users and collected by the NWSC. This survey seeks
to determine how you would value such a project as reflected in the payment you
would be prepared to make toward this water quality improvement scheme.
Valuation Question:
21. In answering the valuation question it is important to that you consider all your
existing financial obligations before making your decision, and that you answer as
truthfully as possible.    Suppose the NWSC undertakes the project to improve
Kampala’s water to a safe drinkable standard. To finance this project, would you
be prepared to pay an extra Sh………. /= per m
3
? This extra charge will be collected
by the NWSC in the normal monthly way. In answering you must take into163
consideration your average monthly water consumption and the Current NWSC
water tariffs (refer to working box below) and what you can afford.
                           
Yes                      No             No Answer
22. If you selected the No option, why so?
 
22.1 I cannot afford
22.2
I already pay enough for the water
service
22.3
Someone else has got to pay for
Such projects
22.4
I am against every project
proposed by the government
22.5 Other (Specify)
NWSC Tariffs as of 2005: Not Reticulated:
Cost per Cubic metre of water 806/=
Connection Fees 58,500/= if main NWSC
distribution pipe is between
1-50 metres Away
Service Charge 1500/=
Beyond 50 metres An extra 3000/= per metre
for 1 inch pipe
VAT 18%
An Extra 2000/= per meter
for a 3/4 of an inch pipe 164
23. If you   chose the No answer Option, Why so?
23..1
I am unable to make a decision without
more time
23.2
I am bored by your survey and I am
anxious to get it over with
23.3
I do not concur with the method
employed ascertain my decision
23.4 Other (Specify)
24. Do you have any comments or additions in relation to the project as proposed?
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
This Questionnaire was designed by Wasswa Francis Department of Economics NMMU Contact:
wasachief@myself.com. Alternatively Prof. S.G Hosking and Noluntu Dhyubele may be contacted in
connection with the study at: Stephen.Hosking@nmmu.ac.za and Dhyubele.Noluntu@nmmu.ac.za
Thank you165
K A  N  S A  N  G A
B R  A  N  C H   2 :
B R  A  N  C H   5 :
N T  I N D A
C I T  Y   C E  N  T  R  E
N A  T  E E T  E
B R  A  N  C H   8 :
N A  N S A  N A
K  A  J J A  N  S I
Z O N  E   -   A
M U K O N O
Z O N  E   -   B
N A  J J A  N A  N K U M  B I
B R  A  N  C H   3
B R  A  N  C H   4 :
B W  A  I S E
B R A  N  C H   1 :
C I T  Y   C E  N  T  R  E
B R A  N  C H   6 :
K I T  I N T  A  L E
B A  K  U  L I
K I R E K A
H  O   T  A N K /   R  E  S ER  VO IR
N BP
I N T  E R  C O  N  N E C T IO N
I N T E R C O  N N E C T IO N
K  A  M  P A L A   W  A  T E R   -   E X I S T I N G   F L O  W   Z O  N IN G
B M  2
B M   31
B M  3 0
B M   35
B M  3 6
B M  1 9
B M   32
B M  3 4
B M  3 9
B M  2 6
B M  4 8
B M   13
B M  1 4
B M   11
B M   38
B M   45
B M  4 6
B M  0 6
B M  4 3
B M   41
B M  0 9 B M   22
B M   21
B M  xx
B M   xx
B M  xx
B M   25
B M   xx
B M   xx
B M  2 9
B M   10
B M   23
B M  1 6
B M   15
B M   44
B M  8
B M   02
B M   01
B M  4 9
B M  2 7
B M  2 8
B M  0 4
B M  1 8
B M  1 7
B M   05
B M   07
B M  3 3
B M  6 7
Appendix B: Kampala water­Existing flow Zoning166
Appendix C: Kampala Water Quality­Improving project Costs and
Benefits Outlay over 52 years, Adjusted at 12% discount rate
Year
Cost (Ushs/m3
)  Benefit (Ushs/m3
)
Discount
Factor for
12%
PVC  PVB
0 1451 0 1 1451 0
1 591.77 385.07 0.8929 528.37 343.81
2 591.77 385.07 0.7972 471.76 306.98
3 591.77 385.07 0.7118 421.21 274.09
4 591.77 385.07 0.6355 376.08 244.72
5 591.77 385.07 0.5674 335.79 218.50
6 591.77 385.07 0.5066 299.81 195.09
7 591.77 385.07 0.4523 267.69 174.19
8 591.77 385.07 0.4039 239.01 155.52
9 591.77 385.07 0.3606 213.40 138.86
10 591.77 385.07 0.3220 190.53 123.98
11 591.77 385.07 0.2875 170.12 110.70
12 591.77 385.07 0.2567 151.89 98.84
13 591.77 385.07 0.2292 135.62 88.25
14 591.77 385.07 0.2046 121.09 78.79
15 591.77 385.07 0.1827 108.11 70.35
16 591.77 385.07 0.1631 96.53 62.81
17 591.77 385.07 0.1456 86.19 56.08
18 591.77 385.07 0.1300 76.95 50.07
19 591.77 385.07 0.1161 68.71 44.71
20 591.77 385.07 0.1037 61.35 39.92
21 591.77 385.07 0.0926 54.77 35.64
22 591.77 385.07 0.0826 48.91 31.82
23 591.77 385.07 0.0738 43.67 28.41
24 591.77 385.07 0.0659 38.99 25.37
25 591.77 385.07 0.0588 34.81 22.65
26 591.77 385.07 0.0525 31.08 20.22
27 591.77 385.07 0.0469 27.75 18.06
28 591.77 385.07 0.0419 24.78 16.12
29 591.77 385.07 0.0374 22.12 14.40
30 591.77 385.07 0.0334 19.75 12.85 167
Year  Cost (Ushs/m3)  Benefit (Ushs/m3)
Discount
Factor for
12%  PVC  PVB
31 591.77 385.07 0.0298 17.64 11.48
32 591.77 385.07 0.0266 15.75 10.25
33 591.77 385.07 0.0238 14.06 9.15
34 591.77 385.07 0.0212 12.55 8.17
35 591.77 385.07 0.0189 11.21 7.29
36 591.77 385.07 0.0169 10.01 6.51
37 591.77 385.07 0.0151 8.93 5.81
38 591.77 385.07 0.0135 7.98 5.19
39 591.77 385.07 0.0120 7.12 4.63
40 591.77 385.07 0.0107 6.36 4.14
41 591.77 385.07 0.0096 5.68 3.69
42 591.77 385.07 0.0086 5.07 3.3
43 591.77 385.07 0.0076 4.53 2.95
44 591.77 385.07 0.0068 4.04 2.63
45 591.77 385.07 0.0061 3.61 2.35
46 591.77 385.07 0.0054 3.22 2.1
47 591.77 385.07 0.0049 2.88 1.87
48 591.77 385.07 0.0043 2.57 1.67
49 591.77 385.07 0.0039 2.29 1.49
50 591.77 385.07 0.0035 2.05 1.33
51 591.77 385.07 0.0031 1.83 1.19
52 591.77 385.07 0.0028 1.63 1.06 168
APPENDIX C: Complete Kampala CVM Model: The MLE SPSS­output
Nominal Regression
Case Processing Summary
63
85
2
12
40
81
12
1
59
89
2
12
29
26
37
38
4
148
8
156
0  No
1  Yes
WTP
1  Very Poor
2  Poor
3  Fair
4  Good
5  Very Good
6  Excellent
QUALITY
OF WATER
0  Females
1  Males
GENDER
0  No Formal Education
7  Primary Leaving Exam
Certificate and Below
11  O-Levels
13  A-Levels
14  National Diploma
16  University Graduate
20  Post Graduate
EDUC (In
Years)
Valid
Missing
Total
N
Model Fitting Information
201.889
176.446 25.444 17 .085
Model
Intercept Only
Final
-2 Log
Likelihood Chi-Square df Sig.169
Goodness-of-Fit
142.188 130 .219
176.446 130 .004
Pearson
Deviance
Chi-Square df Sig.
Pseudo R-Square
.158
.212
.126
Cox and Snell
Nagelkerke
McFadden
Parameter Estimates
2.175 2.115 1.058 1 .304
-140.848 70.910 3.945 1 .047 6.769E-62 2.965-122 .155
-2.97E-02 .021 1.942 1 .163 .971 .931 1.012
2.622E-02 .056 .216 1 .642 1.027 .919 1.147
3.089E-02 .187 .027 1 .869 1.031 .714 1.489
-.153 .115 1.776 1 .183 .858 .685 1.075
.234 6356.311 .000 1 1.000 1.264 .000 .
a
-19.272 1.019 357.373 1 .000 4.268E-09 5.788E-10 3.148E-08
-18.503 .809 523.726 1 .000 9.206E-09 1.887E-09 4.490E-08
-18.200 .751 587.308 1 .000 1.247E-08 2.863E-09 5.436E-08
-18.944 .000 . 1 . 5.925E-09 5.925E-09 5.925E-09
0
b
. . 0 . . . .
-4.48E-02 .392 .013 1 .909 .956 .444 2.060
0
b
. . 0 . . . .
18.118 1.568 133.537 1 .000 7.4E+07 3418700.268 1595704210
18.265 .740 608.856 1 .000 8.6E+07 20064799.19 365269850.0
17.588 .570 952.488 1 .000 4.3E+07 14232335.64 132876341.3
17.450 .625 780.687 1 .000 3.8E+07 11139661.67 128851323.9
17.996 .530 1155.076 1 .000 6.5E+07 23173802.68 184691865.0
17.581 .000 . 1 . 4.3E+07 43178578.23 43178578.23
0
b
. . 0 . . . .
Intercept
LOGINCOM
AGE
HHNUMBER
LNCOSTOW
LNDEFEX
[QUALITY=1]
[QUALITY=2]
[QUALITY=3]
[QUALITY=4]
[QUALITY=5]
[QUALITY=6]
[GENDER=0]
[GENDER=1]
[EDUC=0]
[EDUC=7]
[EDUC=11]
[EDUC=13]
[EDUC=14]
[EDUC=16]
[EDUC=20]
WTP
0  No
B Std. Error Wald df Sig. Exp(B) Lower Bound Upper Bound
95% Confidence Interval for
Exp(B)
a. Floating point overflow occurred while computing this statistic. Its value is therefore set to system missing.
b. This parameter is set to zero because it is redundant.170
Classification
30 33 47.6%
18 67 78.8%
32.4% 67.6% 65.5%
Observed
0  No
1  Yes
Overall Percentage
0  No 1  Yes
Percent
Correct
Predicted