.. . . ..
3. ;2_ ,
in. in.

 

      

v7. VI». ,
,f... .123

 

     

 

w.
. , .o

 

 

      

     

 

 

y..‘.... . ..
5......1. t..V..v.....

     

       

i.

 

    

      

._. :33... ,5:

 

.u...

. .
. 2:: z... ..

 

     

      

r. .1 4
5:13.?

                           

: .53 .:

nu“ ,. 1 mun-Au- -

1.

“4 ‘—

        

....: .
.1 7...!
..§

 

   

Ir;
.1.

 

 

1.1.1“.

  
   
  
 

 
 
 
 

 
 
 
 
 

 
 

 
 
 
 
 

  

 

as? .r 9,. 3 ..__ .5

.Zii 1.1.

  

L . V n. :1, .3 4 N. ... 2x
,4?W. rmfikiwifu, mTJ

  
      

Ill
lil'll!

‘

THESIS
.7

‘r‘

2(‘0!

This is to certify that the

dissertation entitled

Three Essays on Water Service:
A Case Study of the Residential
Sector of Cairo

presented by
Hiba Ahmed

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in Agicultural Economics

Wflfl

MM rprofessor

 

Date 8/18/00

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

LIBRARY
Michigan State
University

 

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

DEC 13 2007
Flt-£154?

 

N

 

 

 

 

 

 

 

 

 

 

 

 

 

1100 W-14

THREE ESSAYS ON WATER SERVICE: A CASE STUDY OF THE
RESIDENTIAL SECTOR OF CAIRO

By

Hiba Ahmed

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements for the degree of
DOCTOR OF PHILOSOPHY

Department of Agricultural Economics

2000

ABSTRACT

THREE ESSAYS ON WATER SERVICE: A CASE STUDY OF THE
RESIDENTIAL SECTOR OF CAIRO

By
Hiba Ahmed

The first essay reports on research that found that households in Cairo are willing to
pay to avoid unreliable water service' with those who are hurt the most willing to pay the
highest amount. Further, willingness to pay consists of a risk premium (fixed across
households) paid to entirely eliminate unreliability and an amount that varies by household
paid to reduce unreliability. Conservation efforts in Cairo provide a limited amount of water
such that eliminating the unreliability problem entirely is infeasible. Accordingly, the
magnitude of the total benefits of conservation efforts, as measured by the summation of
households’ willingness to pay, depends on the manner in which the conserved water is
utilized. Total benefits are underestimated when the conserved water is allocated equally
among households suffering from unreliable service but maximized when the limited
conserved water is allocated starting with those having the least unreliability problem. The
latter allocation allows more households to eliminate unreliability and captures the fixed risk
premium component of willingness to pay.

The second essay reports on research that found that when water service is unreliable,
Cairo households invest in water-improving technologies (WIT)- such as storage tanks and
electrical pressure boosting pumps. Households’ willingness to pay for water improvement
programs is higher the higher the ongoing costs of defensive technologies and lower the

higher the technologies’ ability to mitigate the risk of inadequate water service. Households

with both a pump and a tank are willing to pay LE 5.052 ($1.53) while those using only a
pump are willing to pay LE 3.18 ($0.96) per week for a water reliability improvement
program. These amounts are 2.4 and 1.5 percent of average household monthly income.
These findings indicate that the upward effect on willingness to pay of hi gher operating costs
associated with two technologies is higher than the downward effect of more risk mitigation
using them. The essay also finds that the current household weekly cost of operating and
maintaining a pump of LB 1.04 ($0.32) and both a pump and a tank of LB 1.27 ($0.38)
constitute a lower bound on willingness to pay implying that the defensive technologies are
not successful in eliminating the risk of inadequate water service.

The third essay reports on research that found that females are willing to pay less than
males for residential water connections. This is despite the fact that such connections
provide females with substantial benefits in the form of time and effort savings as well as
improved health status. The data did not support the hypothesis that the lower female
willingness to pay results from controlling only a fraction of the total household income.

Notes
1. Characterized by frequent water cutoffs and/or incidents of low water pressure such
that the household’s water needs cannot be met from the tap at home.
2. LE denotes Egyptian Pounds. The 1995 exchange rate was 1$=3.3 LE. The 1995

average household yearly income was LE 10447.

Covyright by

HIBA AHMED
2000

DEDICATION

To Abdalla, my husband and my best friend, for all his support and to my family for
helping me achieve my goals.

iv

ACKNOWLEDGEMENTS

I would first and foremost like to express my gratitude and sincere appreciation to my
major advisor, Dr. John Hoehn for all his support, valuable advice, and encouragement.
Also, my deepest appreciation goes to him for providing the data necessary to complete this
dissertation.

I remain indebted to members of my committee, Dr. Sandra Batie, Dr. Carl Eicher,
Dr. James Oehmke, Dr. Thomas Reardon, and Dr. Jeffrey Wooldridge for the continuous
support, guidance, and constructive comments they provided me. Many thanks go to my
friend and colleague, Dr. Douglas Krieger for listening to my ideas and providing me with
numerous comments and information.

I benefitted a lot from the assistance of staff members and faculty in the department
of Agricultural Economics. Completing my educational program would not have been
possible without them. I especially appreciate the help of Dr. Eric Crawford for the
emotional, financial, and academic support he provided me throughout my stay in the
department. I also thank the college of Agriculture and Natural Resources at Michigan State
University for their financial support. Also, my thanks go to Nicole Alderman for her help
in producing the final document of this thesis.

Special thanks are owed to my fiiends and colleagues in the department for the
studying we have done together, the exchange of ideas, and the smooth introduction to

different cultures they engaged me in.

TABLE OF CONTENTS

LIST OF TABLES ..................................................... viii
LIST OF FIGURES ..................................................... ix
INTRODUCTION ....................................................... 1
1. The Challenges Ahead ........................................ 1
2. Policies in the past ........................................... 3
3. The Findings ............................................... 4
4. Implications for Future Water Policy ............................ 6
References ....................................................... 9

ESSAY 1: URBAN WATER CONSERVATION: INCORPORATING HOUSEHOLD

SERVICE RELIABILITY ......................................... 10
l . Introduction ............................... _ ................ 10
2. Background ............................................... 13
3. Theoretical Model ................... - ....................... l7
4. Data ..................................................... 19
5. Empirical Methods .......................................... 22
6. Results ................................................... 23

6.1. Current Water Service ................................. 23
6.2. Testing the Hypothesis: Hypothesis One and Two ........... 26
6.3. Conservation Analysis: Hypothesis Three and Four .......... 27
7. Conclusion ................................................ 33
References ...................................................... 35
Appendix 1. Per Capita Cost Savings as a Form of Household Benefits ...... 38

ESSAY 2: WILLINGNESS TO PAY FOR WATER RELIABILITY IMPROVEMENT:
A CASE STUDY OF DEFENSIVE TECHNOLOGIES IN CAIRO, EGYPT . . 42

1 . Introduction ............................................... 42
2. Background ............................................... 44
3. Theoratical Model .......................................... 46

3.1. Willingness to Pay Calculations ......................... 47
4. Data ..................................................... 49
5. Empirical Methods .......................................... 52
6. Results ................................................... 53

6.1. Testing the Hypothesis: ................................ 54
7. Conclusion ............................................... 57
Appendix 2 ...................................................... 61

ESSAY 3: WILLINGNESS TO PAY FOR WATER CONNECTIONS:

A FEMALE PERSPECTIVE ........................................ 64
1 . Introduction ............................................... 64

vi

2. Theoretical Model .......................................... 66
3. Data ..................................................... 68
4. Empirical Methods .......................................... 72
4.1. The General Empirical Model ........................... 72

4.2. Using the Estimated Model to Address the
Questions of Interest: .................................. 73
5. Results ................................................... 75
5.2. Lower Willingness to Pay ................................ 82
5.3. Female Control Over Budget ............................ 85
5.3. Female Control Over Budget ............................ 86
6. Discussion ................................................ 88
References ...................................................... 89

vii

Table 1.1

Table 1.2
Table 1.3
Table 1.4
Table 1.5
Table 1.6
Table 1.7

Table 1.8
Table 1.9
Table 2.1
Table 2.2

Table A2.1
Table A2.2
Table A2.3

Table 3.1
Table 3.2
Table 3.3
Table 3.5
Table 3.6
Table 3.7
Table 3.8

LIST OF TABLES

Past, Current, and Projected Leakage and Waste Rates

for the East Bank ........................................... 15
Percentage of Households in Each Duration ...................... 24
Frequency of Water Cutoffs ................................... 24
Time When Cutoff or Low Pressure Starts ....................... 25
Timing of Water Cutoffs During the Day ........................ 25
Willingness to Pay to Eliminate Water Cutoffs .................... 26
Estimates of Conserved Water Translated to Number of

Hours of Additional Service for the Years 1990 to 2000 ............ 28
Estimated Household Benefits for Different Water Allocations ....... 31
Per Capita Cost Savings as a F orrn of Benefit ..................... 33
Percentage of Households in Each Duration ...................... 54
Probit Estimates of Willingness to Pay .......................... 55
Selected Family Characteristics of Sample Households ............. 61
Selected Characteristics of Respondents’ Dwelling ................ 62
Some Variables to Approximate Household’s Income .............. 62
Current Water Sources ....................................... 76
Factors Affecting Effort and Time .............................. 77
Family Characteristics ....................................... 8O
Probit Estimates of Willingness to Pay .......................... 83
Mean Willingness to Pay (WTP) ............................... 85
Characteristics of Dwellings .................................. 86
Characteristics of Interviewed Respondents ...................... 87

viii

LIST OF FIGURES

Figure 1. Willingness to Pay Differentials by Gender ...................... 84

ix

INTRODUCTION

1. THE CHALLENGES AHEAD

More than twenty percent of Egypt’s population, currently estimated at around 60
million people, lives in Cairo in densities as high as one hundred thousands people per
square kilometer [United Nations, 1990]. Furthermore, Cairo’s population is estimated to
grow at its present rate of at least three percent per year. The size of municipal water needs
in Cairo is directly related to the population size, since the latter constitutes the main variable
responsible for the increase in municipal water demands [Khouzam, 1995; CHZM Hill
International, 1990]. The fresh water shortages facing Egypt at large impact all sectors,
including the municipal sector of Cairo. The arid climate of Egypt, characterized by
negligible rainfall (5-200 mm/year) and high evaporation rates (1 500-2400 mm/year), means
the river Nile is the main source of fresh water contributing to 98 percent of the country’s
total water supply [Ministry of Public Works and Water Resources, 1996]. Egypt already
needs more than its annual allotted Nile share of 55 billion cubic meters to meet the
competing water uses including those for municipal purposes. The country’s water
requirements for all purposes (including water losses) are estimated at 63 billion cubic
meters per year. Further, it is unlikely that Egypt Nile water budget will increase in the near
future due to natural and political reasons in the Nile basin [ibid.]. For example, reducing
losses in the river course by building the Jonglei Canal in Sudan can result in an extra
availability of 10-15 billion cubic meters per year. This work, however, is halted due to civil
war in the project area. Also, due to expected economic development Ethiopia will be

claiming more Nile water in the future.

Shortages of water quantities are not the only concerns around the Greater Cairo area.
Due to high population densities, the Nile River has been subjected to considerable pollution
stemming from man made conditions. These include effluents from untreated/semi treated
domestic and industrial wastewater as well as raw sewage from the rapidly growing
unconnected areas. They also include fertilizer and pesticide leaching and irrigation return
flows from the agricultural sector as well as river traffic and waste from tourist and other
activities [United Nations, 1990; Ministry of Public Works and Water Resources, 1996].

The lack of fresh water is only one factor determining the amount of treated water
available to consumers. Due to the lack of production as well as storage capacity and the
aging distribution system of Cairo (first built in 1865), treated water fell short of daily
maximum demand in the municipal sector of Cairo. The maximum daily demand for treated
water in some parts of Cairo water supply and distribution system was about 3.09 million
cubic meters for the year 1990. This amount exceeded the available supply (estimated at 2.27
million cubic meters per day for the same year). Further, the quantity demanded was
projected to increase to 5.7 and 7.1 million cubic meters per day for the years 2000 and 2010
respectively due to the ongoing rapid population grth and the change in consumption rates
and development patterns [CH2M Hill International, 1990].

Other challenges facing GOGCWS include a lack of financial resources. GOGCWS
collects revenue through lump sum administrative charges rather than quantity based prices.
In 1994, administrative charges yielded average revenue of approximately $37 per thousand
cubic meters of processed water. The revenues collected by GOGCWS were the lowest

compared to other water networks of the same size and population coverage [Hoehn and

Krieger, 1996]. These low revenues are due to charging low subsidized water fees as well
as problems with enforcing and collecting such fees.
2. POLICIES IN THE PAST

The General Organization for Greater Cairo Water Supply (GOGCWS) is faced with
two challenges in the municipal sector: to close the rapidly increasing gap between the
limited water resources and the increasing demands on the one hand and to achieve such an
objective using its limited financial resources on the other.

To manage its limited water resource, both supply oriented (such as developing new
sources) as well as demand oriented (such as increasing the efficiency of water use) measures
were utilized by GOGCWS [Ministry of Public Works and Water Resources, 1996]. A
twenty years program was launched in the seventies with an estimated cost of 2.9 billion
dollars [United Nations, 1990]. The main objective of this effort was to expand the Cairo’s
water system’s production and storage capacity, extend and upgrade its distribution network,
and bring the Cairo’s water system up to existing engineering standards for urban water
supply [Hoehn and Krieger, 1996]. The latter included efforts to improve the quality of
treated water, increase water pressure through out the service area, and strengthen the
management capacity of the GOGCWS [CH2M International, 1990]. Also, detecting and
enhancing the use of ground water was included as part of the policy [United Nations, 1990].

On the other hand, demand management efforts such as using water more efficiently
and reducing water losses were taken [United Nations, 1990; CH2M Hill International,
1990]. However, policies regarding more efficient use of water resources targeted other
sectors with little attention given to the household sector [Richards, 1995]. Other means to

conserve water used by GOGCWS were increases in the lump sum administrative charges

3

and public information campaigns. In mid-1985, the flat fee per ton of water in Cairo was
raised from 1.2 to 3 piasters'. Despite this increase, the new water fees were way below the
cost of production estimated at 8 piasters per cubic meter. Also, efforts were neither taken
to change the structure of water pricing (exchange subsidized flat fees with prices that reflect
the value of water in the municipal sector) nor metering was implemented to make sure that
customers pay for the water actually consumed. Due to this customers had little or no
incentives to conserve and only half of the average per capita supply actually reached their
households due to leaking and broken pipes and poor household fittings [Rada, 1994; CH2M
Hill International, 1990; Hoehn and Krieger, 1996; United Nations 1990].

Also, GOGCWS still faces the financial challenge of expanding water supplies to
meet the growing needs. This challenge makes it rely heavily on foreign assistance. Efforts
to expand the water system were usually multi-national including the United States,
Germany, Japan, and France [United Nations, 1990; Hoehn and Krieger, 1996].

Despite all its efforts in the past, today twenty five percent of Cairo households,
mainly on the periphery, have no access to piped water. More over the supply fluctuates over
much of the urban area meaning many do not have water service all hours of the day or do
not have full discharge pressure.

3. THE FINDINGS

The three essays presented in this dissertation had the broad objective of

understanding consumers’ preferences (as reflected by willingness to pay) regarding water

service in the municipal sector of Cairo. The first essay reports on research that found that

 

‘ Each Egyptian pound denoted as LE contains 100 piasters. The 1995 exchange rate is $3.3 to 1 LE.

4

households in Cairo are willing to pay to avoid unreliable water service2 with those who are
hurt the most willing to pay the highest amount. Conservation was seen as one way to
increase water service reliability. Hence, the household willingness to pay for improvement
in reliability was used to estimate the benefits from conservation efforts. The paper
documented substantial conservation potential in Cairo through reducing domestic,
institutional, and system wide leakage. The benefits of such efforts were estimated to be as
high as three million dollars per month in terms of improved reliability for the households
in the service area.

The second essay reports on research that found that when water service is unreliable,
Cairo households invest in water-improving technologies (WIT) such as storage tanks and
electrical pressure boosting pumps. The cost of these defensive technologies as well as their
ability to mitigate the risk of unreliable service influence the magnitude of the households’
willingness to pay for programs aiming at improving water service reliability. Households
with both a pump and a tank were willing to pay up to $1.53 while those using only a pump
were willing to pay only $0.96 per week for a program that improves water service to 24
hours a day at full discharge pressure. These amounts are 2.4 percent and 1.5 percent of
average household monthly income. The essay also finds that the current household weekly
cost of operating and maintaining a pump of $0.32 and both a pump and a tank of $0.38
constitute a lower bound on willingness to pay because the defensive technologies are not

completely successfiil in eliminating inadequate water service.

 

2 Characterized by frequent water cutoffs and/or incidents of low water pressure such that the household’s
water needs cannot be met from the tap at home.

The third essay reports on research that found that households were willing to pay for
residential water connections. Female heads of the household were willing to pay less than
male heads for such connections (with willingness to pay estimated at $5.63 and $6.07 per
month respectively). This finding was despite the fact that such connections provided
females with substantial benefits in the form of time and effort savings as well as improved
health status. 91 percent of the females interviewed did not work outside the home for cash.
This finding implied that nonworking females obtained a fraction of the total household
income from other members of the household who earned income. Lower female willingness
to pay could be explained if females utilized such a smaller budget for domestic expenses
including those for water needs instead of the total household income. The data, however,
did not support such an explanation and the hypothesis that the lower female willingness to
pay resulted from controlling a fraction of the total household income was rejected.

4. IMPLICATIONS FOR FUTURE WATER POLICY

The broad understanding of consumers’ preferences for water service improvement
is vital for policy makers in Cairo. Such an understanding can be utilized by GOGCWS to
address its two challenges. Results reported in the first essay can address the water resource
challenge while results reported in the second and third essays can address the financial
resource challenge.

The first essay research provided a benefit side analysis of conservation efforts. This
research can be used in a full benefit cost analysis to justify conservation programs.
Conservation can play a substantial role in reducing the gap between the available supply and
the municipal water needs. Suitable conservation programs that encourage reductions in

domestic, institutional, and system wide leakage need to be identified and their cost

quantified. Having an estimate of benefits in dollar terms available makes advocating
conservation efforts more plausible. Due to the high value households place on reliability,
household participation in conservation efforts can be enhanced if public campaigns
emphasized and explained the relationship between reliability and conservation.

Previous literature emphasized the role of willingness to pay as a tool for sizing and
financially planning water services. Willingness to pay is found to be a good indicator of the
potential of cost recovery of investments aiming at improving water service [McPhail, 1993;
MacRae and Whittington, 1988]. It also facilitates consumers’ financial participation if
needed. Using the research results reported in the second essay, comparing willingness to pay
of those with both water tanks and pumps with that of those with only water pumps provides
an idea of the magnitude of households’ willingness to pay for private storage. This amount
is equivalent to $2.3 per month per household. Engineering documents indicate that it costs
GOGCWS $2.7 per month per household to provide system wide storage. This comparison
indicates some cost savings if GOGCWS allows households in its service area to provide
their own storage facilities. Although the difference in cost per household is not substantial
(only $0.40), aggregating over the households in part of the GOGCWS service area
(estimated at 2.09 million households in the year 2000) provides substantial amounts.
GOGCWS cost savings might be as high as $836 thousands per month. This policy can also
be advocated on the grounds of equity. The appendix of this essay shows that households
investing in water improvement technologies are more affluent on average. Allowing them
to provide their own storage provides GOGCWS with extra resource and alleviates its

financial burden.

In the third essay, the information on willingness to pay can be used to determine the
level of service provided (in terms of the number of connections) as well as design charges
based on the value of willingness to pay. Females are the prime users of improved Water
sources. The data, however, did not support the notion that lower female willingness to pay
resulted from controlling only a fraction of the total household income. Excluding the
budget factor leaves many other reasons to justify the lower female willingness to pay. One
of these is the lack of alternative opportunities and uses of female spare time. For example,
if women attach a lower value to their time and effort because of the lack of available
employment, then providing high levels of water connections based on their merit of saving

women’s time might overestimate the value of such connections.

References

CH2M Hill International. “The Rod El-Farag Distribution System”. Final Report.
General Organization for Greater Cairo Water Supply. 1990.

Hoehn J. and Krieger D. “Cairo Water and Wastewater Economic Benefit
Assessment”. Final Report. Prepared for the United States Agency for International
Development/Cairo. 1996.

Khouzam R. F. “Municipal and Industrial Water Use Projections for 2025". Ministry
of Public Works and Water Resources. Water Research Center, Strategic Studies Unit.
Winrock International/EPAT. Cairo, Egypt. 1995.

MacRae D. and Whittington D. “Assessing Preferences in Cost-Benefit Analysis:
Reflections on Rural Water Supply Evaluation in Haiti”. Journal of Policy Analysis and
Management, Vol. 7, No. 2. p 246-263. 1988.

McPhail, A. “The Five Percent Rule for Improved Water Service: Can Households
Afford More?”. World Development, Vol. 21. No. 6. p. 963-973. 1993.

Ministry of Public Works and Water Resources. “Strategy Analysis in Water Demand
Management: A Contribution to Water Resources Planning in Egypt for Confronting the
Imminent Water Scarcity”. Proposal for Research. Planning Sector. Cairo, Egypt. 1996.

Rada Research and Public Relations Co. “Environmental Economics Program:
Qualitative Research Study”. Final Report. Cairo, Egypt. 1994.

Richards A. “Management Strategies for C0ping with Water Scarcity: An
Exploration of Issues”. Prepared for Strengthening the Planning Sector Project, Ministry of
Public Works and Water Resources. Cairo, Egypt. 1995.

United Nations. “Population Growth and Policies in Mega-Cities, Cairo”. Department
of International Economic and Social Affairs. Population Policy Paper No. 34. 1990.

URBAN WATER CONSERVATFIDI‘IEINICORPORATING HOUSEHOLD
SERVICE RELIABILITY
1. INTRODUCTION

The evaluation of water conservation efforts in the municipal sector in previous
literature emphasized the advantages of such efforts to households resulting from possible
reductions in quantities used or wasted, while system reliability is held constant‘ [Maddaus,
1987; Planning and Management Consultants Ltd., 1993; Seattle Public Utilities, 1998;
Renwick and Archibald, 1998; Male, Noss, et al., 1985; Moyer, 1985]. Accordingly, a
significant proportion of the households’ benefits was attributed to reductions in water fees
resulting from the water utility’s ability to avoid long and short run costs of production due
to conservation2 [Moyer, 1985; American Water Works Association, 1999].

Water conservation may also be valued for the role it plays in mitigating water
service unreliability. In a context of a residential water system, reliability is measured by the
degree to which households receive their water service within acceptable standards such as
adequate quantity, good quality and maximum and minimum pressure3 [Harberg, 1997]. The
lack of reliability implies losses or damages to households the magnitude of which is

determined by the length of duration of water cutoffs and incidents of low pressure“.

 

' Conservation refers to water demand management efforts that aim to reduce usage and losses of water. The
water authority and/or the household can take such efforts. In this essay we address efforts taken by the water
authority to reduce losses from leakage.

2 These costs include those for water acquisition, treatment, power and pumping, transmission and distribution,
and support services. They also include capital and wastewater costs [10].

3 In this essay unreliability only refers to water cutoffs and incidents of low pressure. Quality issues are not
addressed.

‘ Duration of water cutoffs is not the sole determinant of damages or losses associated with unreliability. The
frequency as well as the timing of water cutoffs and incidents of low pressure are other determinants. More
frequent cutoffs as well as those occurring at bad timing (when the household needs water the most) are
associated with more damage.

10

The unreliability losses and damages to households include the households’ efforts
to secure water from alternative sources that are more expensive, less convenient, and of
lower quality and their attempts to adapt their water-related activities to variable service and
complete domestic chores at less favorable times when water is more available. In cases
when water cutoffs and low pressure are severe these losses or damages can be substantial
[Hoehn and Krieger, 1996].

The purpose of this paper is to conduct a household level benefit assessment of
possible reductions in household, institutional, and system-wide leakage and waste in Cairo,
E gyptS. By making an extra quantity of water available for use, conservation efforts mitigate
water cutoffs and reduce incidents of low pressure, and hence reduce their associated
damages or losses to households. The households’ benefits from conservation efforts can be
quantified by determining the degree by which conservation can facilitate full water service
and by measuring the amounts that households are willing to pay to receive such service and
avoid the losses or damages associated with unreliability.

The paper proceeds by testing the following hypothesis:

1 . Households that are hurt the most by water service unreliability (those facing
cutoffs and low-pressure incidents that are of longer duration) are willing to
pay the highest amount to receive full water service and avoid the losses or
damages associated with unreliability.

2. The households’ willingness to pay for full service consists of a risk

premium, fixed across households, paid to eliminate unreliability and an

 

5 Our choice of household assessment stems from the fact that the majority of the water processed in Cairo is
destined for the residential sector [11]. Thus the primary beneficiaries from programs that enhance water
service are households [9].

ll

amount that varies by household paid to reduce unreliability. The latter
increases with the severity of the households’ unreliability problem (as
measured by water cutoffs duration).

Eliminating the unreliability problem in Cairo is infeasible through water
conservation alone. Hence, allocating the conserved water equally among
households suffering from unreliability underestimates the total benefits of
conservation. Households’ benefits from conservation are largest when the
limited conserved water is allocated starting with those having the least
unreliability problem (shortest duration of water cutoffs). Using the
conserved water in this manner allows some households to eliminate
unreliability. Eliminating water service interruptions gains an extra premium
in terms of willingness to pay.

Conservation analysis suggested in previous literature underestimates the
total benefits of conservation. It limits households’ benefits to fee reductions
resulting from the water utility’s cost savings and ignores the benefits

resulting from mitigating or eliminating unreliability.

This paper is divided into six sections. The background section explains the

household problem under excess water demand and describes the conservation potential in

Cairo. The theory section explains how to derive a measure equivalent to the maximum

amount a household would pay to avoid unreliability. The data section discusses data

collection procedures. The empirical methods section describes the probit model used to

reach empirical results. The results section describes the current water service received by

12

the household and tests the above four hypotheses. The conclusion section provides a
summary of the final findings.
2. BACKGROUND

The maximum daily demand for treated water in the East Bank6 water supply and
distribution system was about 3.09 million cubic meters for the year 1990. This amount
exceeded the available supply (estimated at 2.27 million cubic meters per day for the same
year). Further, the quantity demanded was projected to increase to 5.7 and 7 .1 million cubic
meters per day for the years 2000 and 2010 respectively due to the rapid population growth
and the change in consumption rates and development patterns [CH2M Hill International,
1990]

The increase in available supplies was not as rapid [ibid.]. Hence, a situation resulted
in which many locations of the East Bank’s service area were not supplied with water at
adequate pressure nor in adequate quantity. Water was available only some part of the day
and typically at inconvenient hours [Hoehn and Krieger, 1996; Rada Research and Public
Relations Co., 1994]. Unannounced water cutoffs and incidents of low pressure impose a risk
on the households residing in the area (estimated at 2.09 million households) since these
households neither know when and how often would a cutoff or an incident of low pressure
occur nor for how long would they last [Rada Research and Public Relations Co., 1994].
Cutoffs and incidents of low pressure that are of longer duration are associated with more

losses or damages since adapting to them requires substantial money, time, and effort

 

6 The Nile divides the water supply system of Cairo into the East Bank and the West Bank water supply and
distribution networks. These two networks are operated independently from each other’s with interconnecting
service provided only on emergency basis. This study is concerned only with conservation efforts in the East
Bank.

13

[Nadim, et al., 1980; Hoehn and Krieger, 1996]. These losses or damages include, but not
limited to, buying water from alternative sources, collecting water from sources outside the
home, storing water, and adapting household activities to less favorable times when water
is more available (for example finishing the laundry very late at night).

A number of studies were conducted to project future water demand for Cairo in
general and the East Bank in particular and find the necessary means of meeting it [ES-
Parsons, 1979; American British Consultants, 1981; Sanyu Consultants Inc., 1986]. An
underlying theme of most studies, however, is their advocacy for establishing new supplies
and expanding the Cairo water system. For example, in 1990 it was suggested that for the
year 2000 maximum daily treated water demand to be met, a total of 3 .5 million cubic meters
were to be obtained from expansion and rehabilitation of treatment plants existing in 1990
and from the construction of new treatment facilities. It was also suggested that for the year
2010 maximum daily treated water demand to be met, the East Bank sources of supply will
have to be nearly tripled by the year 2010 [CH2M Hill International, 1990]. Expanding a
water system as large as Cairo’s is not limited to securing additional sources of fresh water
and enlarging treatment plants, a difficult task in itself, but extends further to updating
storage capacities, transmission lines, and distribution and pumping facilities. These
expansions are expensive to implement, as they involve substantial capital investments, and
require time to complete [CH2M Hill International, 1990].

Supply augmentation, however, is not the only available water management practice.
Conservation (or demand management) often constitutes a cheaper and quicker alternative
[Planning and Management Consultants Ltd., 1993]. Substantial conservation potential

exists in the East Bank. Although this water distribution system does not suffer from

14

excessive leakage at the transmission line level, it records high leakage rates at the service
line level due to poor plumbing [Pitometer Associates, 1988; CH2M Hill International,
1990]. Previous studies provided different estimates of leakage and waste rates in the East
Bank area depending on the estimation technique utilized in the analysis. ES-Parsons Master
Plan estimates that domestic plumbing leakage is approximately 50% of total residential
demand [ES—Parsons, 1979]. Another study concluded that domestic (residential) plumbing
leakage is approximately 35% of total residential demand [Sanyu Consultants Inc.,1986]
while a third study provided a range of 38%-47% as an average for the entire East Bank
service area [Montgomery, 1986]. Table 1.1 provides estimates of past, current, and
projected leakage and waste rates in the East Bank water distribution system as calculated
by previous studies [CH2M Hill International, 1990].

Table 1.1 Past, Current, and Projected Leakage and Waste Rates for the East

 

 

Bank
Year Domestic' Domestic Domestic Institutional” System""
(High)” (Medium) (Low)
1987 38 47 47 50 17.43
1990 38 47 47 45 16
2000 34 3 6 34 30 l 3
201 0 3 O 25 20 l 5 1 0

 

'Measured as a percent of total residential demand.

” High, medium and low refer to the water consumption category is associated with specific
education level and a housing standard of the area in which they are located [CH2 Hill International]
”'Measured as a percent of total institutional demand. Institutional demand refers to water used by
occupants of governmental buildings, schools, hospital, mosques, and small commercial
establishments. It also includes irrigation of public green spaces.

"”Measured as a percent of combined residential and industrial demand.

Table l .1 shows that previous studies identify three areas as targets for conservation.
The first area is household plumbing leakage and waste, defined as “domestic” in the table

and refers to the amount of water delivered to residential buildings served by the network,

15

but not beneficially used by their occupants. The second area is institutional plumbing
leakage and waste. This is defined as the amount of water delivered to governmental
buildings, schools, mosques, and small commercial establishments, but not beneficially used
by their occupants. The third area is system wide leakage. This is defined as the amount of
water which is lost from the network prior to reaching the end user and is more related to the
condition of transmission and distribution lines rather than poor plumbing.

Available literature emphasizes that these levels for each category are reasonable
targets [CH2M Hill International, 1990]. The levels targeted for domestic leakage and waste
are found to reflect the relative financial ability of the consumption category to improve
residential building plumbing and reduce leakage. The targets for institutional leakage and
waste are thought of being reasonable as future building plumbing is repaired and upgraded.
This is particularly true for new buildings that are projected to develop. The targets for
system wide leakage are assumed to be reasonable due to suggested government efforts to
replace corroded pipes and repair and replace the leaky ones [Pitometer Associates, 1988].
Although the specific methods by which the General Organization of Greater Cairo Water
Supply can achieve these reductions are not identified in this study, a number of measures
are available to ensure household and government entities’ participation in reducing leakage
[Maddaus, 1987; Planning and Management Consultants Ltd., 1993; Seattle Public Utilities,
1998; American Water Works Association, 1985]. For system wide leakage, leak detection
programs, pipe replacement, rehabilitation, and repair are highly relevant for the East Bank

[Pitometer Associates, 1988].

16

3. THEORETICAL MODEL

The purpose of this section is to provide a characterization of the household
preferences in a situation involving a risk of water cutoffs or low pressure incidents. Assume
that the household derives utility from consuming a numeraire good X and that there is an
adverse event (a water cutoff or an incident of low pressure) over which the household has
no control. Let C = C*(D) represents the occurrence of a water cutoff that of duration D and
C = 0 represents the absence of a water cutoff”. Duration (D) measures the severity of the
cutoff with longer water cutoffs being more severe.

C takes the value C*(D) with some probability (1t) and the value of zero with a

probability (1- 1:). The household utility conditional on C is:

u = U(X,C)
“(X.0) > "(X,C'(D))
Ux > 0 (1)

Ucw) < 0

The household receives an income (M) with certainty out of which it pays a fixed fee
(F) for waters. Suppose that the household makes a choice about whether or not to pay for
a program to receive full water service before the program is actually implemented (i.e. an
ex-anti choice). Also, assume that the household makes its choice so as to maximize its

expected utility given its budget constraint:

 

7 In this section, the word water cutoff is used interchangeably to refer to both situations when the household
does not get water at all or get water with a level of pressure that is too low to fulfil the household’s water
needs.

8 Water is not currently priced in Cairo, the household pays a fixed fee that does not change with the quantity
it consumes [Hoehn and Krieger].

l7

Maximize: E(U) = nou(X,C’(D)) +(l—1t)ou(X,0)
subject to: M = F + PxX (2)

Since X is a numeraire good, its price is suppressed in the following analysis. Given
utility maximization, there is an indirect utility function V= v(M—F, C) which measures the
maximum attainable utility given M“ = M-F, and C. This indirect utility function has the

properties:

> <
vM. 0,vM.M. 0

vc < O,vc > 0 (3)

Define an option price (OP) to be the maximum payment the household would make
to change from the status quo risk to a situation in which C = C' (D) would not occur. This
option price is a form of compensating surplus where the reference point is the household

expected utility [F reeman, 1993]. Option price is the solution to the following equation:

v(M’-0P,0) = trov(M’,C'(D)) + (l —1t)ov(M‘,0) (4)

For equation (4) to hold, the household will be willing to pay a higher option price

to avoid a longer cutoff.

18

4. DATA9

Contingent valuation was used to estimate option prices for improved water service.
The contingent valuation method uses survey questions to directly elicit people‘s preferences
for non-marketed goods by finding out what they would be willing to pay (in money terms)
for specified improvements in them. Respondents are presented with a hypothetical setting
in which they have the opportunity to trade money payments for such improvement [Mitchell
and Carson, 1989]. In this study water reliability is the non-market good and the specific
improvement valued is an increase from current reliability levels to daily 24 hours water
service at full discharge pressure.

In comparison with other survey methods, the contingent valuation is particularly
demanding in terms of questionnaire development and survey design and implementation
[Arrow, Solow, et al., 1993]. Issues of questionnaire development were particularly
important in Cairo due to language and cultural differences and lack of previous research
[Hoehn and Krieger, 1996]. Hence, a multi-stage process of questionnaire design and
reliability testing was incorporated. A total of nine focus groups were conducted with 69
heads of Cairo households (to establish background information necessary to develop a draft
questionnaire), the draft questionnaire was pre-tested in two phases and 57one-on—one
interviews were completed in the offices of a research company in Cairo”. A field pretest
tested the semi-final questionnaire under conditions similar to those proposed for the final

survey. Fifty in-home interviews were conducted to finalize the development of the

 

9 Hoehn and Krieger collect the data used in this study in 1995. For a detailed account of survey design and
data collection refer to the original work by the two authors [Hoehn and Krieger].

‘° Rada Research and Public Relations in Cairo organized and conducted focus groups and facilitated data
collection.

19

questionnaire. The purpose of this multistage process was to construct a plausible,
intelligible, and meaningful questionnaire to the respondent, reduce miscommunication
between the researcher and the respondent, and train interviewers.

GIS was used to establish a database, update Cairo maps, and draw random primary
sampling units (PSU) by utilizing landscape maps from the year 1968 and combining them
with information on administrative units and urban boundaries, population and population
growth rates, and water service conditions. A total of 25 households were interviewed in
each PSU. Eight hundreds and ninety three heads of households completed the final
questionnaire to value 24 hours of reliable water service, 50% of them were females. The
response rate for this questionnaire was 84%.

The final questionnaire had a total of 75 questions that spread into 13 sub-sections.
It contained four separate sections each collecting different type of information. The
screening section determined the eligibility of the household to complete the questionnaire.
The questionnaire was to be completed by households that are connected to the water system
but had unreliable water service (defined as having water cutoffs and/or a level of water
pressure that is too low to obtain water for meeting household needs for all or part of the day,
once a week, several times a week, and everyday). The second section collected information
on the level of current water service, duration, timing and frequency of water cutoffs. The
third section collected data on health status, housing, and socio-economic characteristics of
the household.

The fourth section, the valuation scenario, allowed households to elicit their choice
for reliability improvement (in the form of receiving daily 24 hours of service at full

discharge pressure) and obtained the maximum amount of money they would be willing to

20

give up to receive it. The valuation scenario expressed the possibility of improved service,
described the hypothetical program that would lead to such improvement, reminded the
respondent with baseline conditions (current water pressure and payment) and the post
program conditions (including the elimination of water cutoffs and low water pressure and
the change in payments to the water authority). The decision setting included reminders of
some of the reasons other respondents accepted or rejected the program, of budget
constraints, and of program costs.

The total sample of respondents was subdivided into 9 sub-samples each randomly
receiving the program at a different initial cost. Respondents in each sub-sample were asked
to accept or reject the program at the proposed program cost. Each sub-sample was proposed
the program at only one of the following nine costs: 20, 40, 60, 80, 100, 175, 250, 400 or 600
piastersll paid weekly for five years. These costs were based on a distribution of value
responses obtained during the pre-testing stage [Hoehn and Krieger, 1996]. The lowest cost
threshold was selected so that approximately 90% of the respondents would accept the
project if they were offered it at that cost. The highest cost threshold was selected so that
approximately 90% would reject the project if they were offered it at that cost. A respondent
rejecting the program at the initial cost was offered it again at a lower cost and that accepting
the program at the initial cost was offered it again at a higher cost. A debriefing section
followed to allow the respondents the chance to discuss the reasons for their choice decision.

This open-ended section encouraged respondents to reflect upon the scenario, slowed the

 

" One Egyptian pound (denoted as LE) contains 100 piasters. The 1995 exchange rate was approximately 33
LE to 1 US dollar.

21

pace of the interview, and provided later information to judge if respondents gave careful
consideration to the economic tradeoffs and consequences of their decisions.
5. EMPIRICAL METHODS

The data collected provided a set of 893 accept/rej ect responses based on 9 different
cost categories. The probit model is used to analyze the data, calculate the mean willingness
to pay for obtaining 24 hours of water service, and test if it varies by duration of cutoffs and
low pressure incidents. In this section we assume that the unobserved willingness to pay (w)
can be explained using the variable D representing duration, according to the following

equation:

wt : W(Di’€i) (5)

Where 1' goes from 1 to N, and e, is an error term. In this section we argue that
household 1' accepts (votes yes) for the reliability improvement program if its willingness to
pay is greater than the cost (1’) at which the program was offered to it. The (t) goes from 1
to 9 in our analysis depending on the sub-sample interviewed. The household will reject
(votes no) otherwise. Hence, the following must be true for a household that accepts the

improvement program:

was.) ”2 (6)

If we choose a linear function to represent the relationship between willingness to pay
and duration, the probability of the respondent accepting the program at a given program cost

could be presented as follows:

22

Probi(Yes/Yr) = Probflio + BiDi + 6:) > Yt] (7)

We assume that s, is identically and independently distributed as normal with a mean
of zero and a standard deviation of 6, move Y,to the other side of the inequality, and use a

standardized normal distribution to obtain the following equation:

Probl(Yes/Yr) = Prob[([30 + 11,1),- - Y‘.)/O] > go) (8)

The probit regression is used to estimate the values of the coefficients in equation (8)

and the standard deviation (0) which maximize the log likelihood function:

LogL = g {d,.log[1 - <I>((Y,.-D,.B)/0)l + (1 — dglogiewz-Dryon} (9)

Where (n) is the total sample size (equals to 893), d, is a dummy variable (equals one
if the respondent accepts the program and zero otherwise), D, is a vector representing
duration, Y, is the cost threshold, 8 is the standard deviation of the estimated distribution, and
<D(.) is the cumulative density function of the normal distribution [Cameron and James,
1987] i
6. RESULTS
6.1. Current Water Service12

Seventy three percent of the households reported witnessing water cutoffs or low
pressure incidents within the last four weeks prior to the interview. The average duration of

water cutoffs or low pressure incidents is 11.4 hours per day, with a minimum of one hour

 

'2 In this section water cutoffs and incidents of low pressure are used interchangeably.

23

and a maximum of 23.9 hours. Table 1.2 provides the percentage of households at each
cutoff or low pressure incidents’ duration. The table shows lower concentration of
households at extreme water cutoffs or low pressure incidents’ duration (too short or too
long)

Table 1.2 Percentage of Households in Each Duration

 

 

Duration of Cutoffs Percentage of Households
(Hours per Day) (%)
1 To 3 Hours 6.98

3.5 To 6 Hours 14.82

6.5 To 9 Hours 18.62

9.5 To 12 Hours 17.73
12.5 To 15 Hours 15.99
15.5 To 18 Hours ’ 15.26
18.5 To 21 Hours 7.12
21.5 To 24 Hours 3.48

 

Table 1.3 provides information about the frequency of water cutoffs for the
households interviewed. The table shows that more than 70% of the cutoffs and low pressure
incidents occurred every day and only around 3% occurred less than once a week.

Table 1.3 Frequency of Water Cutoffs

 

 

Frequency of Cutoff And Low Pressure Percentage of Cutoffs
(%)
Less Than Once Each Week 2.60
About Once Each Week 5.40
Several Times Each Week 19.60
Every Day 72.18
Other 0.22

 

Table 1.4 provides information about the timing of the water cutoffs and low
pressure. The table shows that 96% of the cutoffs and low pressure occurred during the day,
and only 4% occurred when household members are asleep (between 9 pm. to 5 am). Focus

groups’ participants provided an example of how this affected the household’s activities.

24

Women in focus groups reported that they were often forced to do the laundry late at night,
when water was available. This meant that some of them stayed awake all night. This was
reported to cause headache and fatigue and affect women’s productivity the next morning
[Rada Research and Public Relations Co., 1994].

Table 1.4 Time When Cutoff or Low Pressure Starts

 

 

Time Cutoff Starts Percentage of Cutoffs
(%)
Day (5 am. to 9 pm.) 96.37
Night (9:05 pm. to 4:55 am.) 3.63

 

Within the day, water cutoffs do not carry the same weight. Focus group participants
provided some examples of inconvenient times in which water cutoffs occurred when the
household needed water the most. These included times around meals, Muslim prayers", and
when household members are preparing to go to school or work. Table 1.5 provides
information about the timing of water cutoffs and low pressure incidents during the day. The
table indicates that the majority of the cutoffs (more than 78%) start in the morning between
five and eleven o’clock and 92% of the cutoffs start prior to two in the afternoon.

Table 1.5 Timing of Water Cutoffs During the Day

 

 

Time Cutoff or Low Pressure Starts Percent of Cutoffs

(%)

0:00 am. to 5:00 3.45

5:05 am. to 8:00 am. 36.90

8:05 am. to 11:00 am. 41.52

11:05 am. to 2:00 pm. 13.66

2:05 pm. to 5:00 pm. 3.32

5:05 pm. to 9:00 pm. 1.15

 

 

‘3 Muslim prayers require washing one’s self five times a day. The Friday noon prayer (equivalent to Sunday
church in Christian religion) has special importance and most Muslims shower or wash themselves extensively
prior to practicing it. Focus group participants specifically distinguished the time around the Friday noon
prayer to be a hard time for obtaining water [13].

25

6.2. Testing the Hypothesis: Hypothesis One and Two
One: Households that are hurt the most by water service unreliability
(those facing cutoffs and low-pressure incidents that are of longer duration)
are willing to pay the highest amount to receive full water service and avoid
the losses or damages associated with unreliability.
Two: The households’ willingness to pay for full service consists of
a risk premium, fixed across households, paid to eliminate unreliability and
an amount that varies by household paid to reduce unreliability. The latter
increases with the severity of the households’ unreliability problem (as
measured by water cutoffs duration).
To test these two hypothesis a relationship between willingness to pay to receive full
service (daily 24 hours water service at full discharge pressure) and duration of water cutoffs
is estimated". Table 1.6 shows the probit estimates of this relationship.

Table 1.6 Willingness to Pay to Eliminate Water Cutoffs

 

 

Variable Estimates

(Standard Errors)
Intercept 137.28 (11.23)*
Duration of Cutoff (hours) 4.18 (2.13)*
Log-likelihood Function Value -3 72.30
McFadd 0.22
Incremental Chi-Square (df) 208.7(2)
Correct Predictions (%) 74

 

-N=688. (*) Indicates significantly different from zero at less than 5% level.

 

'4 In addition to duration the variables timing and frequency of water cutoffs determine the magnitude of the
households’ losses and damages. The latter two variables are dropped assuming that reducing the duration of
water cutoffs to zero (by providing daily 24 hours of water service) automatically eliminates the risks
associated with frequency and timing. Also, this step is taken to simplify the following conservation analysis
of the different allocation mechanisms.

26

Table 1.6 shows that households that are hurt the most, those with longer water
cutoffs are willing to pay the highest amount to receive full water service and avoid
unreliability losses since the coefficient on “duration of cutoffs” is positive and significantly
different from zero at less than 5% level. The intercept can be interpreted as measuring what
the household is willing to pay to obtain daily 24 hours of water service regardless of its
cutoff duration. It is a measure of the risk premium to completely avoid cutoffs and low
pressure. This amount is fixed across households including those with the minimum duration
of cutoffs (one hour). The intercept is positive and significantly different from zero at all
levels.

Table 1.6 also shows that the willingness to pay equation contains an amount that
varies by household. It is represented by the coefficient on “duration of cutoffs” multiplied
by the number of hours during which the household currently does not obtain water. Since
the coefficient on “duration of cutoff” is positive, this amount increases with the length of
duration.

6.3. Conservation Analysis: Hypothesis Three and Four
Three: Eliminating the unreliability problem in Cairo is infeasible
through water conservation alone. Hence, allocating the conserved water
equally among households suffering from unreliability underestimates the

total benefits of conservation. Households’ benefits from conservation are

largest when the limited conserved water is allocated starting with those

having the least unreliability problem (shortest duration of water cutoffs).

Using the conserved water in this manner allows some households to

27

eliminate unreliability. Eliminating water service interruptions gains an extra

premium in terms of willingness to pay.

To support the first part of the above hypothesis, the amount of water saved through
conservation is calculated, reduced to quantities received at the household tap, and
transferred to additional hours of service. Table 1.7 shows the feasible suggested reductions15
in leakage and waste for the East Bank water system for the years 1990-2000. Also, a
combined program is evaluated assuming that all the leak reduction efforts are taken
simultaneously. All estimates are per day.

Table 1.7 Estimates of Conserved Water Translated to Number of Hours of
Additional Service for the Years 1990 to 2000

 

 

Leakage Source Percent Cubic Meters at the Additional Hours of
Reduction Tap Service

Domestic 9.33 316,962.20 3 .63

System Wide 3 78,478.93 0.90

Institutional 15 91,883.89 1.05

Combined - 576,384.90 6.60

 

The column “cubic meters at the tap” records the estimates of cubic meters per day
that are made available to the household (at the tap within the household dwelling) as a result
of the conservation efforts. These are not equal to the saved cubic meters from conservation.
This is because the leakage and waste savings occur at different stages of the water network
prior to the household dwelling. Hence, it was necessary to calculate savings in cubic meters
after accounting for possibility of leakage and waste at each stage prior to the household

dwelling.

 

'5 These reductions were suggested in engineering documents that were conducted between the years 1979-
1990 and were seen as feasible given the technical conditions of the East Bank area [15, l7, 19, 12].

28

To translate the conserved water to number of hours of additional service (or
inversely to cutoff duration), an assumption was needed as to which amount of water
constitutes 24 hours of water service. The General Organization for Greater Cairo Water
Supply estimated the required average daily demand for the East Bank area. This estimate
specified the projected amount of water required to meet the full average day domestic
demand (after accounting for household leakage) to be 2,096,582 cubic meters per day for
the year 2000. This amount is assumed to be what is needed (at the household tap) to
achieve daily 24 hours of service [12]. In this study we assume that if the water authority is
able to deliver this amount to the household at the tap, then this amount can be used as a base
to calculate the additional number of hours that could be made available through
conservation. This, however, does not say anything about the logistics of making this amount
available at the household’s tap i.e. no assumptions are made as to how the water authority
would go about producing, storing, and transporting the water needed to achieve this goal '6.
Table 1.7 shows that conservation provides a limited amount of water such that eliminating
the unreliability problem entirely is not feasible. Although, each household can obtain some
extra hours of service and reduce their duration of water cutoffs as a result of the
conservation efforts, not every household is able to enjoy full service. Only households with
water cutoff duration less than or equal to the additional hours of service provided through
conservation will be able to enjoy 24 hours of service and eliminate their unreliability losses

or damages entirely.

 

‘6 Knowing the logistics of providing water at the household tap is necessary if future research is concerned
with the cost side of this analysis.

29

To test the second part of hypothesis three above the paper uses Table l .6 to calculate
willingness to pay at each cutoff category and obtains benefit estimates associated with three
mechanisms, the first two of which allow for equal allocation of the conserved water among
the households suffering from unreliability. In the first mechanism, conservation benefits the
household by providing every household with an equal number of extra daily hours of
service. This valuation does not account for the households that are able to leap the daily 24
hours service threshold as a result of obtaining extra hours of service. Hence, the marginal
value of extra hours (as described by the coefficient of the “duration of cutoffs” variable in
Table 1.6) is used to calculate the household benefits. A marginal value of 4.181 is
multiplied by the third column of Table 1.7 for each household in the service area.

In the second allocation mechanism, every household obtains the same number of
daily extra hours of service. This brings some households, those with current daily hours of
cutoffs equivalent to or less than the extra hours provided, to 24 hours of water service. The
water provided to this percentage of households will be evaluated at the marginal value of
hours plus the risk premium of completely eliminating water cutoffs (as measured by the
intercept in Table 1.6). The rest of the households, with cutoffs that of longer duration than
the extra hours of service provided through conservation, drive a marginal value similar to
that obtained in the first mechanism.

In the third mechanism, water saved from conservation is allocated starting with
those with the shortest cutoffs. The percentage of households in each cutoff category obtains
water that brings it up to 24 hours of service until all the water is utilized. Since the
conserved water is not enough to provide all the households with daily 24 hours of service,

some households are left at their initial reliability level. The benefits in this case are

30

calculated as the risk premium plus the marginal value for all the households achieving daily
24 hours of water service. Table 1.8 provides benefits estimates of each of the leakage
reduction efforts and for each water allocation mechanism. All figures are in US dollars per
month for the specific percentage of households accommodated under each allocation

mechanism.

Table 1.8 Estimated Household Benefits for Different Water Allocations (Per

 

 

Month in US. Dollars)
Leakage Equal Allocation Equal Allocation Allocation Minimum
Source (100% at Marginal (Some % at Marginal Cutoff (Some % at 24
Value) Value) Hours Service)

Domestic 403,234.82 654,444.80 2,228,129.51
System Wide 100,021.56 116,126.10 882,006.06
Institutional 116,691.82 132,721.96 967,769.29
Combined 733,268.60 1,471,978.00 3,425,765.82

 

Under the first mechanism, 100% of the households are included in the analysis with
each getting the marginal value from reducing their current cuto ff duration by the extra hours
of service provided through conservation. Under the second mechanism, the percentage of
households taken over the 24 hours of service threshold depends on the number of extra
hours made available through conservation. This is 7.14 % for domestic, 0.44% for both the
institutional and system wide leaks, and 21.82% for the combined effort. The remaining
households obtain a marginal value similar to that in the first mechanism. Under the third
mechanism, 21.43% (for system wide), 23.38% (for the institutional), and 73.87% (for the
combined program) of the households are brought over the daily 24 hours water service
threshold with no one else getting additional hours as a result of reducing domestic leakage

and waste.

31

In conclusion, Table 1.8 shows that equal allocations of the conserved water among
households suffering from unreliability underestimate the total benefits of conservation while
the allocation that utilized the conserved water starting with households with the least
unreliability problem (shortest cutoff duration) maximizes such benefits. This is because the
latter mechanism allows the maximum number of households to enjoy daily 24 hours of
service, completely eliminates their unreliability problem, and captures their risk premium.

Four: Conservation analysis suggested in previous literature
underestimates the total benefits of conservation. It limits households’
benefits to fee reductions resulting from the water utility’s cost savings and

ignores the benefits resulting from mitigating or eliminating unreliability.

Previous literature emphasized the household benefits from conservation resulting
from possible reductions in quantities used or wasted while system reliability is held
constant. Under such an analysis, conservation benefits the household through a reduction
in water fees resulting from the cost savings from avoided production. The water authority
does not re-distribute the water saved through conservation. Water production is reduced by
the amount of water conserved and so is the water authority’s cost. It is assumed that the
water authority distributes the cost savings equally among the households served”. Table 1 .9
provides a comparison between the water allocation mechanism that starts with households
having the least cutoffs and the per capita cost savings suggested in previous literature. All
figures are in US dollars per month for the percentage of households accommodated under

each analysis.

 

‘7 For a detailed analysis of the calculations involved in this hypothesis refer to Appendix A.

32

Table 1.9 Per Capita Cost Savings as a Form of Benefit

 

 

Leakage Source Per Capita Cost Savings Allocation Minimum Cutoff
(Some % at 24 Hours Service)

Domestic 929,072.50 2,228,129.51

Institutional 471,289.10 967,769.29

System Wide 291,647.60 882,006.06

Combined 1,526,176.67 3,425,765.82

 

Table 1.9 shows that the per capita cost savings produce a lower estimate of the
household benefits. This is despite the fact that 100% of the households are provided with
per capita cost savings but only 7 3.87% to 21.43% are provided with daily 24 hours of water
service under the second allocation mechanism.

7. CONCLUSION

This study estimated the benefits of reducing domestic, institutional, and system wide
leakage and waste in mitigating the household’s water service unreliability in the residential
sector of Cairo. It shows that households are willing to pay for daily 24 hours of water
service and eliminate the risks of unannounced water cutoffs and low pressure incidents.
Households are willing to pay more to avoid longer cutoffs and low pressure incidents.

Due to water quantity limitations, the mechanism taken to allocate the conserved
water among the households with unreliable service affects the value of the estimated total
households ’ benefits. Allocating the limited conserved water equally among households with
unreliable water service does not bring maximum benefits. The maximum benefit is obtained
when the households are ranked and the conserved water is allocated starting with those with
the least unreliability problem. Such allocation allows more households to obtain daily 24
hours of water service and captures the risk premium they are willing to pay to entirely

eliminate unreliability.

33

Limiting conservation benefits to supply cost savings, suggested in previous
literature, is found to underestimate the value of conservation programs to the household.
Such an analysis does not account for those households that are willing to pay to completely

eliminate the risks associated with unreliability.

34

References

American British Consultants (AMBRIC). 1981 . Rehabilitation and Expansion of the Cairo
Wastewater System. Final Report. Cairo, Egypt.

American Water Works Association (AWWA). 1996. “Water: \Statsl 996: The Water Utility
Database.” Denver, CO.

American Water Works Association (AWWA). 1999.Water Audits and Leak Detection:
Manual of Water Supply Practices. Denver, Co.

Arrow, K., R. Solow, et al. 1993. Report of NOAA: Panel on Contingent Valuation. Final
Report, pp. 4602-4614.

Cameron, T. and M. James. 1987. “Efficient Estimation Methods for Closed-Ended
Contingent Valuation Surveys,” Review of Economics and Statistics 69: 269-76.

Center for Development Information and Evaluation. “Capital Projects: Economic and
Financial Analysis of Nine Capital Projects in Egypt”. A.I.D. Report #19. Cairo,
United States Agency for International Development.

CH2M Hill International. 1990. Rod El-Farag Distribution System. Final Report. Cairo,
General Organization for Greater Cairo Water Supply.

Clark, R. 1976. “Water Supply Economics,” Journal of the Urban Planning and
Development Division 102: 213-224.

Clark, R. and J. Gillean. 1977. “The Cost of Water Supply and Water Utility Management.”
Environmental Protection Agency. USA.

Clark, R. and R. Stevie. 1981. “A Water Supply Cost Model Incorporating Spatial
Variables,” Land Economics 57(1).

Dajani, J. and R. Gemmell. 1973. “Economic Guidelines for Public Utilities Planning,”
Journal of the Urban Planning and Development Division 99: 171-182.

ES—Parsons. 1979. Waterworks Master Plan. Final Report: a Joint venture in association
with ECG—Cairo, Cairo.

Ford, L. and L. Warford. 1969. “Cost Functions for the Water Industry,” Journal of
Industrial Economics 18: 53-63.

35

Freeman, M. 1993. “The Measurement of Environmental and Resource Values: Theory and
Methods.” Washington DC, Resources for the Future.

General Organization for Greater Cairo Water Supply. 1994. Presentation to the United
States Agency for International Development. Final Report. Cairo.

Harberg, R. 1 997. Planning and Managing Reliable Urban Water Systems. American Water
Works Association. Denver, CO.

Hoehn, J. and D. Krieger. 1996. Cairo Water and Wastewater Economic Benefit Assessment.
Final Report. Prepared for United States Agency for International Development,
Cairo.

Koeni g, L. 1966. “The Cost of Water Treatment by Coagulation, Sedimentation, and Rapid
Sand Filtration.” US Public Health Service, Division of Water Supply and Pollution
Control, Technical Services Branch. United States.

Linaweaver, P. and C. Clark. 1964. “Costs of Water Transmission,” Journal of the
American Water Works Association 56: 1549-60.

Maddaus, WC. 1987. Water Conservation. The American Water Works Association.
Denver, CO.

Male, 1., R. Noss, et al. 1985. Identifidng and Reducing Losses in Water Distribution
Systems. Noyes Publications. New York.

Marks, D., C. Revelle, et al. 1970. “Mathematical Models of Location: A Review,” Journal
of the Urban Planning and Development Division 96: 81-93.

Mitchell, C. and R. Carson. 1989. “Using Surveys to Value Public Goods: The Contingent
Valuation Method.” Resources for the Future. Washington DC.

Montgomery J. 1986. Water Waste and Metering Report. Vol. VII. Final Report. Consulting
Engineers, Cairo.

Moyer, E. 1985. The Economics of Leak Detection: A Case Study Approach. American
Water Works Association. Denver, Co.

Nadim, et. a1. 1980. “Living Without Water”. Cairo Papers in Social Science. The American
University of Cairo, Cairo.

Pitometer Associates, Engineers. 1988. Investigations for Leakage and Flow Measurements.
Final Report. Montclair, New Jersey. USA.

36

Planning and Management Consultants Ltd. 1993. Evaluating Urban Water Conservation
Programs: A Procedures Manual. American Water Works Association. Denver, Co.

Rada Research and Public Relations Co. 1994. Environmental Economics Program:
Qualitative Research Study. Rada Research and Public Relations Company. Cairo.

Renwick, M. and S. Archibald. 1998. “Demand Side Management Policies for Residential
Water Use: Who Bears the Conservation Burden?” Land Economics 74(3): 343-359.

Sanyu Consultants Inc. 1986.Review of Master Plan in Greater Cairo Water Supply
Improvement Project Phase 111. Final Report. Tokyo, Japan.

Seattle Public Utilities. 1998. “Water Conservation Potential Assessment.” Seattle. USA.

37

Appendix 1. Per Capita Cost Savings as a Form of Household Benefits

Three steps are taken to estimate per capita cost savings as a form of household
benefits. In the first step, a water authority’s cost function is estimated using regression
analysis. This cost function relates total maintenance and operation costs to the quantity of
water produced. In the second step, the estimated cost function is used to calculate cost
savings due to conservation. In the third step, a rule is developed to allocate the cost savings
among the households in the service area.

Theoretical Background

Physically, it is possible to separate the water supply system into two components:
the treatment plant and the delivery (transmission and distribution) system [Clark, 1976;
Clark and Stevie, 1981; Koenig, 1966]. Economics of scale (at the treatment plant level)
may be offset by diseconomies of dispersion and spatial arrangement at the distribution level
[Dajani and Gemmell, 1973; Marks, Revelle, et al., 1970; Linaweaver and Clark, 1964; Ford
and Warford, 1969].

Despite that, empirical studies suggest that the relationship between costs (TC) and
the amount of water produced (Q), given input prices, can be reduced to an exponential

fimction of the following form [Clark and Gillean, 1977]:

TRC = TC(Y) = aQ", and
aTC/6Q>0 (10)
or > 0 and l 2 [i 2 0

Estimates of the cost function’s parameters (or and B) are obtained by using a

logarithmic functional form in which the log of the dependent variable (TC) is regressed on

38

the log of the independent variable (Q). After estimating the parameters, it is a matter of
substituting data on Q (the quantity of water conserved) to obtain the cost savings required.

Second, a rule is established to distribute the estimated cost savings among the
households in the service area. The rule applied in this analysis assumes that the water
authority divides the cost savings of not producing the conserved water equally among the

households served:

(11)

Where N is the total number of households receiving water in the service area.
Data Requirement

With the exception of one year, detailed cost data was not available fi'om the General
Organization of Greater Cairo Water Supply [Center for Development Information and
Evaluation, ; CHZM Hill International, 1990]. Available cost data was not adequate to
estimate both the parameters (or and B) of the cost fimction specified in equation (1). The
following two steps were taken instead: First, the parameter [5 is estimated using detailed US
data. The B parameter reflects the returns to scale to the technology employed by the water
authority. In Cairo, such a technology is mostly supplied through U.S. investments in capital
and personnel training [CH2M Hill International, 1990]. Hence, US. data provides a close
substitute for Cairo data. The American Water Works Association collected detailed data
on cost and quantities of water produced from 898 member water utilities. Utilities used in

this analysis were closer to that of Cairo’s in their sources of supply (the use of ground and

39

surface water) and their size ranging to those serving 16 million people [American Water
Works Association, 1996].

Second, after the parameter B is estimated, data obtained from the Central
Organization of Greater Cairo Water Supply is used to calculate the parameter or in equation
(1). This parameter reflects the level of cost rather than the type of technology and how it is
utilized in the production process. Hence, the use of Cairo data was required to obtain a
reliable estimate.

Results
The following cost function was estimated using a combination of US and Cairo

data'3:

TC = 801.1 Q"-83
N = 579 (12)
R2 = 0.73

Where TC represents operating and maintenance costs (including routine and
preventive maintenance, salaries, wages, and benefits, chemicals, power, and other process
inputs, but do not include funds to replace capital investments)”. Q is the quantity of water
produced in million gallons per day and N is the number of water utilities used in the
estimation. The cost function in equation (12) is used to obtain per capita cost savings based

on the projected population of the East Bank for the year 2000 (estimated at 12,531,032

 

‘3 Results are also checked against the cost of water per cubic meter in Cairo as reported in USAID documents
[31]. Results obtained from the cost estimates in this study were within $0.0026 of those reported in USAID
documents.

'9 The assumption of using operation costs rather than total costs stems from the fact that capital costs do not
change by much with the change in output for water utilities in the short run [Clark and Gillean].

40

people) and using an average household size of 6 individuals (average family size in the

sample).

41

WILLINGNESS TO PAY FOR vsisTlERIIELIABILITY IMPROVEMENT:
A CASE STUDY OF DEFENSIVE TECHNOLOGIES IN CAIRO, EGYPT
1. INTRODUCTION

In the presence of adverse environmental effects, households affected often take
voluntary defensive measures, including the adoption of technologies, that can help reduce
the risk of such adverse effects. In previous literature the cost of technology and its ability
to mitigate the risk faced by those affected were emphasized as major determinants of
adopting defensive technology [Courant and Porter, 1981; Gerking and Stanley, 1986;
Harford, 1984; Bartik, 1988; Murdoch and Thayer, 1990]. Adopting defensive technologies
affects the value of the household benefits (as measured by willingness to pay) derived from
programs aiming at reducing the adverse environmental effects [Abdalla, 1990; Lee and
Moffitt, 1993; Dardis, 1980; Watson and Jaksch, 1982; Jakus, 1994]. A review of the
previous literature identifies two issues of interest regarding the relationship between
willingness to pay, cost, and risk mitigation.

First, a tradeoff exists between defensive technologies’ cost and their ability to
mitigate risk and the value of willingness to pay. On the one hand the defensive expenditure
literature argues that the higher the ongoing costs of defensive technologies, the higher are
the benefits from programs aiming at reducing the adverse environmental effects, and the
higher is the willingness to pay for them. Hence, higher ongoing costs have an upward effect
on willingness to pay [Courant and Porter, 1981; Gerking and Stanley, 1986]. On the other
hand, the technology adoption literature argues that the higher the ability of defensive

technologies to mitigate the risk associated with the adverse environmental effects, the

42

higher will be the demand for such technologies. In that case, willingness to pay for
programs aiming at reducing the adverse environmental effect is reduced since households
already obtain the program’s benefits from investing in defensive technologies. Hence, the
higher ability of defensive technologies to mitigate risk have a downward effect on
willingness to pay.

Second, the costs of defensive technologies can be used to obtain some information
on the magnitude of willingness to pay for programs aiming at reducing adverse
environmental effects'. The defensive expenditure literature argues that the cost of defensive
technologies may constitute an upper or lower bound on willingness to pay depending on
their success in mitigating the risks of the adverse environmental effect compared to
proposed programs aiming at achieving the same purpose. If defensive technologies are not
successful in reducing the adverse environmental effect, then their costs constitute a lower
bound on willingness to pay. Willingness to pay for programs aiming at reducing the
adverse environmental effects exceeds the defensive technologies’ on going costs to reflect
the willingness of households to pay a premium to enjoy more protection against risk
[Harford,l984].

The purpose of this paper is to address how households’ adoption of defensive
technologies, such as water storage tanks and electrical water pressure boosting pumps,

affects their willingness to pay for a program that aims at reducing water service

 

‘ One of the major benefits of defensive expenditure models is that they allow willingness to pay be estimated
using only information on protective opportunities. This benefit is particularly useful when prior information
on protective opportunities is more reliable than information on willingness to pay or if it is costly to collect
willingness to pay data directly [Bresnahan and Dickie].

43

unreliability. It collects data from heads of households in Cairo, Egypt to test the following
two hypothesis:
1. Willingness to pay for better water service is lower for households investing
in two defensive technologies if compared with willingness to pay of
households investing in only one technology. This indicates that the upward
effect of higher costs (associated with a double technology) on willingness
to pay is stronger than the downward effect of more riSk mitigation using two
technologies.
2. Operation and maintenance costs of defensive technologies constitute a lower
bound on willingness to pay for more reliable water service. Willingness to
pay will be higher than such costs implying that defensive technologies are
not entirely successful in eliminating the risk of inadequate water service.
This paper is divided into six sections. The first section provides a background of
the problem. The second section develops a theoretical model. The third section discusses
the data. The fourth section describes the empirical methods. The fifth section presents the
results. The sixth section provides some concluding remarks.
2. BACKGROUND

The one certainty in Cairo today is its critical residential water shortage. The demand
for residential water exceeds the system’s capacity. The available water is barely enough to
meet the growing uses [Hoehn and Krieger, 1996]. Those who are connected to the
municipal service often receive water only part of the day and typically at inconvenient
hours. Inadequate water service is characterized by routine reductions in water pressure or

complete cutoffoffs. Low water pressure and complete cutoff off in service are frequent

44

enough that households need to supplement from other sources outside the home [Tecke,
Oldham and Shorter, 1994]. Sources outside the home are often of lesser quality and of
higher cost. In an effort to protect themselves against inadequate water service, Cairo
households often engage in voluntary defensive actions to mitigate or totally avert the
consequences of poor water service [Hoehn and Krieger, 1996; Rada Research and Public
Relations Co., 1994].

Households have a set of defensive actions available to them to reduce the effects of
inadequate service. Households adapt their activities by performing water intensive chores
at less convenient times when sufficient water is available, often late at night [Hoehn and
Krieger, 1996]. In addition, households buy bottled water and store water for essential uses
such as drinking or cooking [Tecke, Oldham and Shorter, 1994]. They also store larger
amounts of water for other uses. Storing water is messy, takes up space, inconvenient, and
time consuming [Hoehn and Krieger, 1996]. Storing can also reduce the quality of water as
well as involve the cost of purchasing storage containers [Rada Research and Public
Relations Co., 1994]. Households also buy larger amounts of water and obtain free water
from areas with more adequate supply. The effort, time, and money to haul and purchase
water are significant costs.

Households with unreliable service often invest in water improving technologies
(WIT). The main WIT used in Cairo are rooftop tanks and electrical pumps. Rooftop tanks
fill when water pressure is high and provide water when system pressure fell [Hoehn and
Krieger, 1996]. Electrical pumps boost pressure and are connected directly in the household
dwelling [Tecke, Oldman and Shorter, 1994]. Tanks are expensive to install, difficult to

maintain, and leak substantial amounts of water [CH2M Hill International, 1990]. They are

45

also associated with lower water quality. Electrical pumps are expensive to install, maintain,
and Operate. They also create noise, increase electricity bills [Hoehn and Krieger, 1996].
Installing, maintaining, and operating both an electrical pump and a storage tank (a double
investment) is more costly than a single investment [Rada Research and Public Relations
Co., 1994]. Also, both the water tanks and electrical pumps involve substantial fixed costs
at the installation phase.
3. THEORATICAL MODEL

The model discussed in this section is designed to obtain a relationship between the
cost of defensive technologies, their ability to mitigate unreliability, and the household
benefits (as described by its willingness to pay) from obtaining full water service. A
proposed water reliability improvement program is assumed to change the water service level
from a lower level of W" to a full level of Wm. W0 is the level of water service initially
delivered by the water authority. W24 is the water level that is proposed with the reliability
improvement program and is equivalent to water service 24 hours a day at full pressure. The
current level of water service received by a household is denoted by W and can be any
where between W0 and W24 depending on whether the household invested in a WIT and on
whether such a WIT is successful in providing more reliable service. It can be described by

the following equation:

W = pW" + (1 -B)W24 (1)

 

2 Water service can be viewed as the number of hours water is available at good discharge pressure.

46

Where 0 <B<l (i.e. 0 can fall any where between 0 and 1) . The level of 0 determines
the success of the WIT in mitigating the household risk of unreliability. If [3:1, then
household’s WIT is totally failing in mitigating unreliability, and the household receives a
level of water service equals to W. If [3:0, then the household’s WIT it totally successful
and the household is able to reduce all its risk to zero and receives W 2" i.e. the household is
able to replicate the level of water service available alter the improvement program.

Assume that a household that uses a WIT derives its utility from consuming a current

water service level W), and a good (X) according to the utility function:

U = U(W,X) (2)

The household decision to use the WIT depends on whether such a use improves its
well being. The household optimizes its position by maximizing utility subject to a budget

constraint:

max U = U(W,X) subject to M = PxX + qW + K (3)

Where M is total income, q is ongoing operating and maintenance costs of providing
additional water using a WIT, and K is the fixed installation cost of the WIT. The Optimal

level of X consumed will then be defined by:

X = x(P,,M-q -K.W) (4)

3.1. Willingness to Pay Calculations
The value of the water service reliability improvement program, as reflected by the

household’s willingness to pay, stems from the households’ ability to trade some of the

47

desirable goods and services it currently consumes to enjoy the new level of water service
[Hoehn and Krieger, 1996]. If the household is willing to pay for the reliability improvement
program, then a comparison of its current level of indirect utility with its post improvement
program indirect utility will be a good estimate of its Hicksian3 compensating measure of
benefits [Freeman, III, 1993]. Substituting back the level of optimal X from equation (4) into
the utility function provides the indirect utility function which measures the maximum level
of well-being the household achieves given M, P,, q, K, and W. This indirect utility function

is defined by:

V : V(Px,M-q—K,W) (5)

The household’s willingness to pay (wtp) solves the following equation:

V = v°(Px,M-K-q,W) = v‘(P,,M-K-wtp.W”) (6)

Note that households that invested in a WIT have already suffered a sunk cost of (K)
whether or not they get improved water service from the water agency. In the above
equation (wtp) for firll water service will include both the value of eliminating the ongoing
cost of operating the defensive technology (q) as well as the value of obtaining an additional
level of water service (reliability gains) measured by the difference between W and W“ .
From the above equation, the higher the ongoing operating costs of the defensive technology

(q) the higher is willingness to pay. On the other hand, the smaller the reliability gains from

 

3 Hicksian measures, as compared to Marshalian measures, assume that the individual’s level of utility is kept
unchanged. For more details refer to references [Freeman, III; Nicholson].

48

the water improvement program (i.e. the higher the defensive technology’s ability in
mitigating unreliable service) the lower is willingness to pay.
4. DATA

Contingent valuation was used to estimate wtp required to solve equation (6). The
contingent valuation method uses survey questions to directly elicit people's preferences for
non-market goods by finding out what they would be willing to pay (in money terms) for
specified improvements in them. Respondents are presented with a hypothetical setting in
which they have the opportunity to trade money payments for such improvements [Mitchell
and Carson, 1989]. In this study, water reliability is the non-market good and the specific
improvement valued is an increase from current reliability levels to daily 24 hours water
service at full discharge pressure.

In comparison with other survey methods, the contingent valuation is particularly
demanding in terms of questionnaire development and survey design and implementation.
Issues of questionnaire development were particularly important in Cairo due to language
and cultural differences and lack of previous valuation research [Hoehn and Krieger, 1996].
Hence, a multi-stage process of questionnaire design and reliability testing was incorporated.
A total of nine focus groups were conducted with 69 heads of Cairo households (to establish
background information necessary to develop a draft questionnaire), the draft questionnaire
was pre-tested in two phases and 57 one-on-one interviews were completed in the offices of

a research company in Cairo“. A field pretest tested the semi-final questionnaire under

 

‘ Rada Research and Public Relations in Cairo organized and conducted focus groups and facilitated data
collection.

49

conditions similar to those proposed for the final survey. Fifty in-home interviews were
conducted to finalize the development of the questionnaire.

The purpose of this multistage process was to construct a plausible, intelligible, and
meaningful questionnaire to the respondent, reduce miscommunication between the
researcher and the respondent, and train interviewers. GIS was used to establish a data base,
update Greater Cairo maps, and draw random primary sampling units (PSU) by utilizing
landscape maps from the yearl 968 and combining them with information on administrative
units and urban boundaries, population and population grth rates, and water service
conditions. A total of 25 households were interviewed in each PSU. Eight hundred and
ninety three heads of households completed the final questionnaire to value 24 hours of
reliable water service, fifty percent of which were females.

The final questionnaire had a total of 75 questions that spread into 13 sub-sections.
It contained four separate sections each collecting different type of information. The
screening section determined the eligibility of the household to complete the questionnaire.
The questionnaire was to be completed by households that are connected to the water system
but had unreliable water service (defined as having water cutoffs and/or a level of water
pressure that is too low to obtain water for meeting household needs for all or part of the day,
once a week, several times a week, and everyday). The second section collected information
about the level, cost, and quality of the water service currently received by the household,
whether the household invested privately in an electrical pump to boost water pressure, a
tank to store water or both, and the expenses incurred to improve current water service

reliability (including the cost of operating and maintaining water-improving technologies).

50

The third section collected data on health status, housing, and socio-economic characteristics
of the household.

The fourth section, the valuation scenario, had the central objective of allowing
households to elicit their choice for reliability improvement and obtaining the maximum
amount of money they would be willing to give up to receive it. The valuation scenario
expressed the possibility of improved water service, described baseline conditions (current
water pressure and payment), and post program conditions (including the elimination of
water pumps or storage tanks usage, and the change in payments to the water authority). The
decision setting included reminders of some of the reasons other respondents accepted or
rejected the program, of budget constraints, and of the cost of the program.

The total sample of respondents was subdivided into 9 sub-samples each randomly
receiving the proposed maintenance program at a different initial cost. Respondents in each
sub-sample were asked to accept or reject the program at the proposed program cost. Each
sub-sample received only one of the following nine costs: 20, 40, 60, 80, 100, 175, 250, 400
or 600 piastersS paid weekly for five years. These costs were based on a distribution of value
responses obtained during the pre-testing stage [Hoehn and Krieger, 1996]. The lowest cost
threshold was selected so that approximately 90% of the respondents would accept the
project if they were offered it at that cost. The highest cost threshold was selected so that
approximately 90% would reject the project if they were offered it at that cost. A debriefing
section followed to allow the respondents the chance to discuss the reasons for their choice

decision. This open-ended section encouraged respondents to reflect upon the scenario,

 

5 One Egyptian pound (denoted as LE) contains 100 piasters. The 1995 exchange rate was approximately 3.3
LE to 1 US dollar.

51

slowed the pace of the interview, and provided later information to judge if the respondents

gave careful consideration to the economic tradeoffs and the consequences of their decisions.

5. EMPIRICAL METHODS

The data collected provided a set of 893 accept/rej ect responses based on 9 different
cost categories. The probit model is used to analyze the data, calculate the mean willingness
to pay for the water reliability improvement program, and test if it varies by the usage of
WIT. Assume that the unobserved willingness to pay (w) can be explained using a set of
indicator variables S (whether the household uses a WIT, and which one they use if any) and

Y (demographic and socio-economic characteristics):

wt : W(St’ Yi’ er) (7)

Where i goes from 1 to N, and e, is an error term. In this section we argue that
respondent i accepts (votes yes) for the improvement program if her willingness to pay is
greater than the cost (C,) at which the program was proposed to her. The (t) goes from 1 to
9 in our analysis depending on the sub-sample interviewed. She will reject (votes no)
otherwise. Hence, the following must be true for a respondent who accepts the water

improvement program:

w(Si,Y 6'.) > C, (8)

52

If we choose a linear function to represent the relationship between willingness to pay
and its explanatory variables, the probability of the respondent accepting the program at a

given program cost could be presented as follows:

Probi(Yes/C‘) = Prob[([30 + BIS, + 1121’, t E) t (3,] (9)

We assume that e, is identically and independently distributed as normal with a mean
of zero and a standard deviation of 0, move C, to the other side of the inequality, and use a

standardized normal distribution to obtain the following equation:

Probi(Yes/Ct) = Prob{[([l0 + BIS, + BZY‘. — C)/0] > ‘E/O} (10)

The probit regression is used to estimate the values of the coefficients in equation

(10) and the standard deviation (0) which maximize the log likelihood function:

LogL = 2: id, log[1-<I>((K,B-C,)/o)] + (1—d,)log[<1>((K,ts—C,.>/o>1} (11)

Where (n) is the total sample size (equals to 893), d,- is a dummy variable (equals one
if the respondent accepts the program and zero otherwise), K, is a vector of explanatory
variables (S,, Y,), o is the standard deviation of the estimated distribution, and <I>(.) is the

cumulative density function of the normal distribution [Cameron and James, 1987].

53

6. RESULTS6

In our sample 19.6% reported using some form of a WIT to improve their water
service (a pump, a tank, or both). The capital cost of installing a water pump or a storage tank
ranged from 1400 LE ($424.24) to 500 LE ($151.52) depending on capacity (equivalent to
160%-57% of average monthly income)7. Operating costs are usually higher for a double
investment (both a pump and a tank). It costs an average of 4.35 LE ($1.32) per month to
maintain and operate water pumps and 5.35 LE ($1.62) to operate both water storage tanks
and pumps. These costs are for achieving the current level of water service (less than 24
hours).

Table 2.1 provides the percentage of households at each cutoff duration. The table
shows lower concentration of households at extreme water cutoffs duration (too short or too
long). The table also indicates that using defensive technologies does not necessarily
eliminate the unreliability problem since not even a single household in the sample reported
a zero cutoff duration.

Table 2.1 Percentage of Households in Each Duration

 

 

Duration of Cutoffoffs Percentage of Households
(Hours per Day) (%)
l to 3 hours 6.98
3.5 to 6 hours 14.82
6.5 to 9 Hours 18.62
9.5 to 12 Hours 17.73
12.5 to 15 Hours 5.99
5.5 to 18 Hours 15.26
8.5 to 21 Hours 7.12
1.5 to 24 Hours 3.48

 

 

° For descriptive statistics see appendix A.
77 Based on an annual income of 10447LE ($3165.76) in 1995 prices.

54

6.1. Testing the Hypothesis:

One: Willingness to pay for better water service is lower for
households investing in two defensive technologies if compared with
willingness to pay of households investing in only one technology. This
indicates that the upward effect of higher costs (associated with a double
technology) on willingness to pay is stronger than the downward effect of
more risk mitigation using two technologies.

Table 2.2 provides the probit estimates of willingness to pay as a function of selected
independent variables.

Table 2.2 Probit Estimates of Willingness to Pay

 

 

Variable Definition Estimated Coefficients and Standard Errors
Probit on Accept/Reject Significant at a= 0.00, 005*
Base Group: Used both a Pump and a Tank Significant at a= 0.10“
N=l 75
Constant Term 5.83 (1.5)*
Used only a Pump -1.87 (1.1)"
University Degree 1.75(0.80)*
Large Apartment 1.78(0.81)*
Frequent Water Bills -4.3 1(2.1)*
Standard Deviation 3.39(O.66)*
Log Likelihood Value -120.95
McFadden’s R Square 0.20
% Correct Predictions (Total) 71%
% Correct Predictions (Accept) 77%
% Correct Predictions (Reject) 65%
Chi-square (df) 47.73 (5)

 

The coefficient on “used only a pump” variable was negative and significantly
different from zero at less than the 10% level indicating that respondents who used only a
pump are willing to pay less for the water reliability improvement program compared to

those who used both a pump and a tank. Since it costs less to operate only a pump than both

55

a pump and a tank and since it is expected that a double technology is more successful in
mitigating unreliable service when compared to a single investment, the results indicate that
the upward effect of higher costs on willingness to pay is stronger than the downward effect
of more risk mitigation using two technologies.

Those with a university degree were willing to pay more for the reliability
improvement program. Having higher education might imply that the respondent earns more
income. Also, it might imply that the value of the respondent’s time is higher. Since
adapting to water shortages requires a lot of time and effort, the respondent with a university
degree might value water reliability more. Also, a more educated respondent might have
more concerns about water quality. In the case of unreliable water service, respondents
might need to store water and rely on other sources that are of lesser quality. Respondents
with larger apartments were willing to pay more for the reliability improvement program.
This might be due to the fact that respondents with larger apartments needed more water to
clean, or due to the fact that having a larger apartment indicates earning more income.

Finally, if the respondent paid for piped water more frequently (once or twice a
month) she will be willing to pay less for the reliability improvement program. Water is not
metered in Cairo and respondents have to pay a fixed water fee regardless of the amount of
water they actually consume. If the fixed fee is paid more often the amount of money
available for other goods, including the reliability program, is reduced. If the respondent had
to pay for water according to the amount she consumes, then unreliable water service implies
lesser quantities consumed, and frequent water bills need not necessarily mean higher water
cost. However, when water is not metered and fixed fees are paid, the higher frequency of

water payments implies a higher water cost. Note that the cost at which the reliability

56

program was proposed does not include paying for the water consumed and hence the
household needed to set aside some money for paying for water before eliciting its
willingness to pay.

Two: Operation and maintenance costs of defensive technologies
constitute a lower bound on willingness to pay for more reliable water
service. Willingness to pay will be higher than such costs implying that the
defensive technologies are not entirely successful in eliminating the risk of
inadequate water service.

Using the estimated coefficients, a mean willingness to pay value was calculated for
those who used only a pump and those who used both a pump and a tank. The former group
is willing to pay LE 3.18 ($0.96) per week for the water reliability improvement program
compared to LB 5.05 ($1.53) for those with both a pump and a tank. The average yearly
income for the household is reported to be LE 10,447 ($3165.76). Using this figure, the
willingness to pay amounts are 2.4% of monthly income for those who used both a pump and
a tank, and 1.5% for those who used only a pump. These willingness to pay figures are
higher than the ongoing costs of operating and maintaining defensive technologies reported
above. These results imply that the defensive technologies are not entirely successful in
eliminating the risk of inadequate service. This is also supported in table 1 using sample data.
7. CONCLUSION

In the case of unreliable water service, Cairo households invest in water improving
technologies (WIT) to defend themselves against the adverse effects of not having enough
water. The defensive expenditure literature suggests that the higher the costs currently

incurred by the household when using defensive technologies, the higher would be the

57

household’s benefit from a program that improves reliability. On the other hand, the
technology adoption literature suggests that risk reduction is a major determinant for the
demand for technology. If the technology is successful in mitigating the risk (in our case
inadequate water service), the household’s willingness to pay is less likely to be high for a
program that provides more reliable water service. The household already obtains the
program’s benefit through their investment.

This paper finds that willingness to pay for full water service is lower for those
households investing in two WIT as compared to willingness to pay of those investing in
single WIT. Households with both a water storage tank and an electrical pressure boosting
pump are willing to pay LE 5.05 ($1.53) while those using only a pump are willing to pay
LE 3.18 ($0.96) per week. These estimates are found to be 2.4% and 1.5% using average
household income in Cairo. A double technology (both a pump and a tank) costs more to
operate and maintain but is more likely to be successful in mitigating inadequate water
service. The findings, however, indicate that the upward effect of higher costs of a double
investment on willingness to pay is stronger than the downward effect of more risk
mitigation using two technologies.

The defensive expenditure literature also suggests that the cost of defensive measures
constitutes a lower bound on willingness to pay to avoid adverse environmental effects if the
defensive measure is not successful in totally eliminating the adverse environmental effects.

The paper finds that operation and maintenance costs of defensive technologies
constitute a lower bound on willingness to pay for full water service. Willingness to pay for
both groups (those investing in a single and a double technology) was higher than the costs

incurred by the household to obtain the current level of water service (less than 24 hours).

58

It costs an average of LE 1.04 ($ 0.32) per week to maintain and operate water pumps and
LE 1.27 ($0.38) per week to operate both water storage tanks and pumps. Such finding
imply that the defensive technologies used by households are not completely successful in

eliminating the risk of inadequate water service.

59

References

Abdalla, C. 1990. Measuring Economic Losses from Ground Water Contamination: An

Investigation of Household Avoidance Cost.” Water Resources Bulletin Vol. 26, no.
3.

Altaf, M.S., D. Whittington, H. Jamal, V.K. Smith. 1993. Rethinking Rural Water Supply
Policy in the Punjab, Pakistan. Water Resources Research, Vol. 29, no. 7, 1943 -
1954.

Bartik, T. 1988. “Evaluating the Benefits of Non-Marginal Reductions in Pollution Using

Information on Defensive Expenditure,” Journal of Environmental Economics and
Management, No. 15, 1 11-127.

Breshnahan, B. and M. Dickie. 1995. “Averting Behavior and Policy Evaluation”. Journal
of Environmental Economics and Management, Vol.29, 378-392.

Cameron, T. and M. James. 1987 “Efficient Estimation Methods for Closed-Ended
Contingent Valuation Surveys,” Review of Economics and Statistics, 69. p. 269-276.

CH2M Hill International. 1990. Rod El-Farag Distribution System. Final Report. General
Organization For Greater Cairo Water Supply. Vol. II.

Courant, RN. and RC. Porter. 1981. “Averting Expenditures and the Cost of Pollution,”
Journal of Environmental Economic and Management. No. 8, 321-329.

Dardis, R. 1980. “The Value of a Life: New Evidence from the Market Place,” The American
Economic Review, vol. 70, no. 5.

Freeman, III. 1993. The Measurement of Environmental and Resource Values: Theory and
Methods. Resources for the Future, Washington DC.

Gerking S. and L. Stanley. 1986. “An Economic Analysis of Air Pollution and Health: The
Case of St. Louis,” The Review of Economics and Statistics, Vol. 68, Issue 1.
February.

Harford, J. 1984. “Averting Behavior and the Benefits of Reduced Soiling,” Journal of
Environmental Economics and Management. Vol. 8, 296-302.

Hoehn, J. and D. Krieger. 1996. Cairo Water and Wastewater Economic Benefit Assessment.
Report. Prepared for the United States Agency for International Development/Cairo.

J akus, P. 1994. “Averting Behavior in the Presence of Public Spillovers: Household Control
of Nuisance Pests,” Land Economics, vol. 70, no. 3.

60

Khouszan, RF. 1995. Municipal And Industrial Water Use: Projections for 2025. Report for
the Ministry of Public Works and Water Resources. Cairo, Egypt.

Lee, L. and L. J. Moffitt. 1993.“Defensive Technology and Welfare Analysis of
Environmental Quality Change with Uncertain Consumer Health Impacts,” American
Journal of Agricultural Economics, No. 75.

Mitchell, R. and R. Carson. 1989. Using Surveys to Value Public Goods: The Contingent
Valuation Method. Resources for the Future. Washington, DC.

Murdoch, J. and M. Thayer. 1990. “The Benefits of Reducing the Incidence of
Nonmelanoma Skin Cancers: A Defensive Expenditures Approach”. Journal of
Environmental Economics and Management. No. 18, 107-119.

Nicholson, W. 1989. Microeconomic Theory: Basic Principles and Extensions. Fourth
Edition. The Dryden Press. Chicago.

Rada Research and Public Relations Co. 1994. Environmental Economics Program:
Qualitative Research Study. Final Report. Cairo, Egypt.

Tecke, B. L. Oldham, and F. Shorter. 1994. A Place to Live: Families and Child Health in
a Cairo Neighborhood, The American University in Cairo Press.

Watson, W.D. and J.A. Jaksch. 1982. “Air Pollution: Household Soiling and Consumer
Welfare Losses,” Journal of Environmental Economics and Management, Vol. 9,
248-262.

Whittington, D., A. Okore, and A. McPhail. 1990. Strateg» for Cost Recovery in the Rural

Water Sector: A Case Study of Nsukka District, Anambra State, Nigeria. Water
Resources Research, Vol.26, No. 9, pp. 1899-1913.

61

Appendix 2

Table A2. 1 provides some family characteristics of interviewed households. The table
shows that those who used a double investment were on average more educated. The table
also shows that those with a double investment had smaller families with a lower number of

children under the age of 5.

Table A2.] Selected Family Characteristics of Sample Households

 

 

Characteristic Whole Using a Using a Pump
Sample Pump and Tank
Senior Female:
Age (Years) 40 39 40
Completed at Least Primary Education (%) 60 72 84
Obtained at Least a University Degree (%) 20 27 50
Senior Male:
Age (Years) 44 44 49
Completed at Least Primary Education (%) 69 79 87.5
Obtained at Least 3 University Degree (%) 35 47 62.5
Number of Household Members:
All Ages (Number) 4.7 4.7 4.6
Females 14 Years of Age or Older (Number) 1.7 1.6 1.9
Males 14 Years of Age or Older (Number) 1.6 1.6 1.7
Less Than 5 Years Old (Number) 0.4 0.5 0.3

 

Table A2.2 shows some selected characteristics of the households’ dwellings. On
average, a higher percentage of those with a double investment lived in apartments as
opposed to rural houses. They also had a higher percentage of owners as opposed to renters.
Their dwellings were larger on average (more number of rooms) and more likely to have
special use rooms such as dining rooms. The table also shows that this group had a higher
percentage of dwellings with finished floors (either wood, stone, tiles, carpet, or linoleum)

and with finished walls (either plaster, paint, wallpaper, or wood). The table indicates that

62

those with a double investment had more access to public infrastructures and services such
as paved streets, public transportation, and sewer connections.

Table A2.2 Selected Characteristics of Respondents’ Dwelling

 

 

Dwelling Characteristics Whole Using a Pump Using a Pump
Sample and a Tank

Apartments (%) 89 94 97
Rural Houses (%) 10 6 3
Owner Occupied (%) 47 45 59
Renter Occupied (%) 53 55 41
Number of Rooms 5.8 6.1 6.3
Separate Dining Room (%) 22 34 41
Finished Floors (%) 96 100 100
Finished Walls (%) 92.4 92.7 100
Building Facing Paved Street (%) 42 48 81
In-home Sewer Connection (%) 81 84 100
Average Walk Time from Nearest

Public Transportation (min) 11.1 10.3 7.2

 

Table A2.3 provides information on some of the variables that might approximate the
household’s income level.

Table A2.3 Some Variables to Approximate Household’s Income

 

 

Variable Whole Using a Using a
Sample Pump Pump and
Tank
Head Female Employed as a Professional (%) 4 6 9
Head Male Employed as a Professional (%) 17 24 31
Number of Eamers in the Household 1.2 1.3 1.2
Total Annual Expenditure (000LE) 6.9 7.2 7.5
Value of Occupied Dwelling (000LE) 44.3 47.8 76.7

 

Table A2.3 shows that on average households that used both a water pump and a tank
did not have more members working outside the home for cash. However, this result does

not mean they might have a lower household income. This conclusion is because the heads

63

of households in this group had higher quality jobs. A higher percentage of heads in this
group are employed as professionals (for example doctors, engineers, and accountants) . On
average, those who employed a double investment had more valuable dwellings (in terms of
expected selling price). Total annual expenditure for those with both a pump and a tank was
higher than the rest of the groups. Hence, on average it can be argued that households with

a double investment were more affluent that other groups.

64

ESSAY 3:
WILLINGNESS TO PAY FOR WATER CONNECTIONS:
A FEMALE PERSPECTIVE
1. INTRODUCTION

In most cultures, fetching water and adapting to inadequate water service is a job for
women (Whittington et. al, 1990; Narrowe, 1989; Nadim et. al, 1980). When their
households are not connected to the municipal water system, women are the prime suppliers
of water for domestic use (IWTC, 1990). Hence, the provision of improved water sources,
such as residential water connections, often have complex direct and indirect effects on many
facets of women’s lives, their households, and the community’s economy (Swarna and
Whittington, 1994; Bolt, 1994).

The direct benefits of a residential water connections include the reduced time and
effort spent by women hauling water from remote sources to the home and the improved
health and socio-economic status. A chain of indirect benefits can also be initiated. For
example, if women are no longer spending time fetching and hauling water from traditional
sources, they may reallocate the saved time to different activities such as food preparation,
agricultural work, childcare, and leisure. The time saving reallocated to childcare, food
preparation, and agricultural work may reduce child mortality and morbidity, and increase
the household’s income (MacRae and Whittington, 1988; Blumberg, 1989).

Despite that, direct and indirect benefits do not always translate into women’s higher
willingness to pay for improved water sources. A mixed evidence is found in the literature.
Some studies found that females are willing to pay more for improved water sources

compared to males (Whittington et al 1990; Ministry of Energy and Water Resources,

65

Zimbabwe ,1985), while others docrunented the opposite results (Hoehn and Krieger 1996;
Krieger and Hoehn, forthcoming).

In some communities males dominate market-oriented production and achieve
effective ownership of and control over the household resources (Femandez-Kelly, 1981;
Nash, 1977; Caulfield, 1981; Afonja, 1981; Linares, 1984; Bossen, 1984). This is especially
true when females do not contribute to any of the household earnings. When that happens,
women might receive a fraction of the household income which they spend on purchasing
items for domestic consumption including water needs (Schuster, 1982; Browner 1986;
Nash, 1977). If willingness to pay estimates are conditional on the income distribution as
suggested in previous literature (Swarna and Whittington, 1994), one reason for lower female
willingness to pay for improved water sources might be the females’ lack of control over
household resources.

This paper has the broad objective of adding to the limited literature by helping
explore some of the determinants of female willingness to pay for improved water sources.
The program under study provides residential water connections. The paper specifically
answers the following two questions:

1. Do the high female benefits from obtaining residential water connections in

Cairo translate into higher willingness to pay for such connections?

2. If not, does the female lower willingness to pay for residential water

connections results from controlling a fiaction of the total household monthly

income‘ .

 

' 1n Egyptian culture it is not unusual for non-working women (housewives) to receive a certain fraction of the
total household income. This smaller budget is used to cover domestic needs such as paying for water, buying
food items, and other household needs.

66

This paper is divided into five sections. The first section outlines the theoretical
model. The second section describes data collection procedures. The third section provides
the empirical methods used to reach final results and the fourth section lays out the results.
The fifth section discusses the results and provides concluding remarks.

2. THEORETICAL MODEL

The absence of a residential water connection imposes substantial costs on
individuals. These costs include the money, time, and effort required to obtain water from
other sources. Furthermore, the lack of a residential water connection forces individuals to
consume less water than they would do if they had a water connection at home. These
factors affect the individual’s utility. We can define the individual’s utility level prior to

obtaining a residential water connection (Uo) using the following equationzz

U0 = u°(Y-c,E-e,W°) (1)

Where, 1’ is total income, E is the effort endowment, (c) is the money cost, and (e)
is the amount of effort required to obtain water from current sources. W0 is the amount of
water the individual consumes at the present and prior to obtaining a residential water
connection.

If the individual obtains a residential water connection, the money cost and effort
required to obtain water from available sources will reduce to zero. The individual will pay

a fixed fee (F) for consuming the new level of water’. Also, the individual will be able to

 

2 For similar formulation see Krieger and Hoehn, forthcoming.
3 Water is not currently priced in Cairo. Consumers are charged a fixed fee that does not vary with the amount
of water they consume.

67

consume a higher level of water service W. The individual’s utility after obtaining such a

connection can thus be described by the following equation“:

U' = u‘(Y-D,E,W‘) (2)

The economic value of a residential water connection is measured by the maximum
amount of income the individual is willing to sacrifice to obtain such a connection and still
maintains her pre connection level of utility. Such an economic value will be equivalent to

the willingness to pay (wtp) that satisfies the following equation:

u°(Y—c,E—e,W°) = ul(Y-F-wtp,E,Wl) (3)

In this formulation, willingness to pay for a residential water connection includes the
value of money cost and effort minimized (c, e), plus the value of the higher level of water
the individual is able to consume as measured by the difference between W1 and W0.
Willingness to pay will be higher the higher is the amount of water consumed with the
residential water connection (W‘) and the larger the money cost (c) and effort (e) minimized.

The general equation provided in (3) can be specified for male and female heads of
households as follows:

u°(Ym-cm,E—em,W°) = ul(Ym—F-wtpm,E,Wl)

° _ _ _ l __ _ 1 (4)
u (Y! cfE efWo) _ u (Yf F wtpr.W)

 

“ Water quality is kept constant in this analysis. We assume that the water the individual consumes now has
the same quality as the water she will get with the residential water connection. This might not be true in the
real analyzed situation.

68

Where Y", and Yfare income for males and females respectively, (cm) and (C) are the
male and female money cost and (em) and (ef) are the male and female amount of effort
required to obtain water from current sources. Willingness to pay by males and females are
denoted by wtp", and wtpf respectively.

If females and males are endowed with the same effort E, and if W0 (the amount of
water used before residential water connections), W'(the amount of water used after
residential connections), and the amount of income (I) do not differ by gender (i.e. if Y,,, =
Y]: Y) , then we expect wtp", to be less than wtpf. This is because the money cost and effort
minimized due to residential connections are higher for females (i.e. cf>, ef> em). However,
if data reveals that wtp", is more than wtp, , then equation (4) indicates that Yf< Ym. In this
paper we hypothesize that female respondents control only a fraction of total household
income (such that Yf = orY) where 0<or <1, while male respondents control the total
household income (such that Y,,, = 1’). Hence, Y", >Yf.

3. DATA5

Three aspects are of interest in this section: questionnaire design and development,
sample selection, and survey administration and implementation. Contingent valuation was
used to estimate (wtp) required to solve equation (4) above. The contingent valuation method
uses survey questions to directly elicit people's preferences for non-market goods by finding
out what they would be willing to pay (in money terms) for specified improvements in them.

Respondents are presented with a hypothetical setting in which they have the opportunity to

 

5 The data used in this paper was collected by Hoehn and Krieger. Summary provided in this section is obtained
directly from the original work of the two authors. For more details refer to Hoehn and Krieger, 1996.

69

trade money payments for such improvements (Mitchell and Carson, 1989). In this study a
residential private water connection is the non-market good.

In comparison with other survey methods, the contingent valuation is particularly
demanding in terms of questionnaire design and development, sample selection, and survey
administration and implementation. Issues of questionnaire design and development were
particularly important in Cairo due to language and cultural differences and lack of previous
valuation research (Hoehn and Krieger, 1996). Hence, a multi-stage process of questionnaire
design and reliability testing was incorporated. A total of six focus groups were conducted
with 46 heads of Cairo households who are not connected to the municipal water system (to
establish background information necessary to develop a draft questionnaire), the draft
questionnaire was pre-tested in two phases and 50 one-on-one interviews (37 in the first
phase and 13 in the second) were completed in the offices of a research company in Cairo“.

The purpose of these pre-tests was to improve the questionnaire, especially the
valuation scenario and enhance the interviewers’ skills and information delivery. A field
pretest tested the semi-final questionnaire under conditions similar to those proposed for the
final survey. Approximately 50 residential interviews were conducted to finalize the
development of the questionnaire. The purpose of this multistage process was to construct
a plausible, intelligible, and meaningful questionnaire to the respondent, reduce
miscommunication between the researcher and the respondent, and train interviewers.

The sample selection stage was also challenging due to the lack of updated Cairo

maps. GIS was used to establish updated Greater Cairo maps and draw random primary

 

‘5 Rada Research and Public Relations, Heliopolis, Cairo, Egypt organized and conducted focus groups, and
facilitated data collection.

70

sampling units (PSU). Landscape maps constructed in 1968 (the most recent at the time of
the survey) were combined with information on administrative units and urban boundaries,
population and population grth rates, water network connection rates to establish the GIS
database. A total of 25 households were chosen in each PSU. Candidate PSU had a water
connection rate of less than 40%. One thousand and two respondents completed the final
questionnaire to value the provision of water network services to households previously not
connected to the water system. The response rate was 73%.

The final questionnaire had a total of 74 questions that spread into 12 sub-sections.
It contained four separate sections each collecting a different type of information. The
screening section determined the eligibility of the household to complete the questionnaire.
The questionnaire was to target households that are not connected to the government piped
water system. The second section collected information about the sources of both free and
purchased water, the money, time and effort cost of acquiring water from current sources,
and the perceptions of the respondents regarding piped water and how it compares to water
obtained fi'om current sources. The third section collected data on the characteristics of the
respondent’s housing and property value, family composition and health status, and
characteristics of family heads. The questionnaire also provided the interviewers the chance
to collect visual observations about the respondent’s building and neighborhood.

The fourth section, the valuation scenario, had the central objective of allowing
respondents to elicit their choice for having a residential water connection and obtaining the
maximum amount of money they would be willing to sacrifice to receive it. The valuation
scenario expressed the possibility of residential connections, described a hypothetical

program that would install a water network in the respondent’s neighborhood. It further

71

described baseline conditions (current water sources with no payments to the water authority)
and the post program conditions (including renewed payments to the water authority for the
coming five years).

The decision setting included reminders of some of the reasons other respondents
accepted or rejected the program, of budget constraints, and of the cost of the program. The
total sample of respondents was subdivided into 9 sub-samples each receiving the program
at a different initial cost. Respondents in each sub-sample were asked to accept or reject the
program at the proposed program cost. Each sub-sample received only one of the following
nine costs: 100, 150, 200, 250, 300, 450, 600, 750, or 900 piasters7 paid weekly for five
years. These costs were based on a distribution of value responses obtained during the pre-
testing stage (Hoehn and Krieger, 1996). The lowest cost threshold was selected so that
approximately 90% of the respondents would accept the project if they were offered it at that
cost. The highest cost threshold was selected so that approximately 90% would reject the
project if they were offered it at that cost. In addition to covering the cost of network
installation, these payments covered the cost of the water consumed.

The provision of water service was conditional upon the continuity of payments.
Also, to reduce free riding, respondents were assured that other households in the
neighborhood will pay their share of the program cost. A debriefing section followed the
valuation process to allow the respondents the chance to discuss the reasons for their choice
decision. This open-ended section encouraged respondents to reflect upon the scenario,

slowed the pace of the interview, and provided later information to judge if the respondent

 

7 One Egyptian pound (denoted as LE) contains 100 piasters. The 1995 exchange rate was approximately 3.3
LE to 1 US dollar.

72

gave careful consideration to the economic tradeoffs involved and the consequences of their

decisions.
4. EMPIRICAL METHODS
4.1. The General Empirical Model

The data collected provided a set of 1002 accept/rej ect responses based on 9 different
cost categories. The probit model is used to estimate the mean willingness to pay for the
water connection and test if it varies by gender. We assume that the unobserved willingness
to pay (w) can be explained using a set of indicator variables S (gender) and Y (other

characteristics):

wt : W(St’ Yi’ei) (5)

Where i goes from 1 to N and e,- is an error term. In this section we argue that
respondent i accepts (votes yes) for the improvement program if her willingness to pay is
greater than the cost (C,) at which the program was offered to her. (t) goes from 1 to 9
depending on the sub-sample interviewed. She will reject (votes no) otherwise. Hence, the

following must be true for a respondent who accepts the improvement program:

w (81" Y1 6:) > C1 (6)

If we choose a linear function to represent the relationship between willingness to pay
and its explanatory variables, the probability of the respondent accepting the program at a

given program cost could be presented as follows:

Probi(Yes/Cr) = Pr0b[(Bo+BlS,+DzYi+€,) t (3,] (7)

73

We assume that e,- is identically and independently distributed as normal with a mean
of zero and a standard deviation of a, move C, to the other side of the inequality, and use a

standardized normal distribution to obtain the following equation:

Probi(Yes/C() = Prob {[(130 +0181 +02 Y! - C‘)/O] > - 61/0} (8)

The probit regression is used to estimate the values of the coefficients in equation (8)

and the standard deviation (as) which maximize the log likelihood function:

LogL = g {d,log11 -<I>((K,15—C,)/oi+(1——d,)log1<I>((K,B—C,)/o)1} (9)

Where (n) is the total sample size (equals to 1002), d, is a dummy variable (equals
one if the respondent accepts the program and zero otherwise), K is a vector of explanatory
variables (5,, Y,), o is the standard deviation of the estimated distribution, and <I>(.) is the
cumulative density function of the normal distribution.

4.2. Using the Estimated Model to Address the Questions of Interest:

Cameron and James, 1987 argued that the above empirical formulation produces a
willingness to pay equation that can be interpreted analogously to the results obtained from
ordinary least squares OLS regression. This willingness to pay equation takes the following

general form:

W.- = 130 + 13K + 6.- (10)

where K,- is a vector of explanatory variables (S,, Y,).

74

Substituting for some of the variables in the vector (K), two of the resulting equations

are of interest:

wtp = [30 + [31 household cash incomei + [33 other factors + et (11)

where “household cash income” is a continuous variable that measures the monthly

household cash earnings.

If female =1, then the above equation reduces to:

wtpi = [[30 + [32] + B: household cash incomei + [33 other factors + e‘, (12)

The intercept is now [[30 + 02] instead of [30. The intercept is smaller than that in
equation (1 1) if females are willing to pay less for residential connections (i.e. if Bz< 0). The
marginal effect of the variable “household cash income” is the same for male and female
respondents and is measured by the coefficient ([3,).

The assumption that females lower willingness to pay results from controlling a

fraction (or) of “household cash income”, can be tested using the following equation:

wtp, = [30 + [31 household cash incomei + [32 femalei +
[31(1-a) femaleiir household cash incomei + [33 other factors + 61- (13)

or alternatively,

wtp, = [30 + [5, household cash incomei + [32 femalei +
6 femalei * household cash incomei + 03 other factors + (at.

(14)

when female =1, then equation (13) reduces to:

75

wtp‘, = [[50 + [32] + ([31 * or) household cash incomei
+ [33 other factors + ex, (15)

In this formulation, we assume that the intercept [[30 + [32], and the marginal effect of
the variable “household cash income” as measured by ([3,) are the same for male and female
respondents. Equation (15) estimates (wtp) if female respondents control a fraction (or) of the
household cash income. Hence, to answer the second question of whether lower female
willingness to pay results from controlling a fraction of total household income, the
coefficient on the interaction term “female*household cash income” as measured by (6*) in
equation (14) needs to be negative and statistically significant.

5. RESULTS
5.1. Numerous Benefits8

The benefits from residential water connections to women in Cairo are mainly the
elimination of the time and effort required to secure water from sources outside the home and
the potential reduction in adverse health impacts. Table 3.1 shows the current sources of

water used by the households in the sample.

 

8 Results reported here are from sample data, focus group and qualitative results, and previous research.

76

 

Table 3.1 Current Water Sources

 

 

Current Water Sources Responses
Water At Cost
All Sources (%) 18
Filled Containers at a Private Tap (%) 2
From A Vendor (%) 66
Paid Someone to Fill Containers (%) 31
Paid Someone to Fill Water Tanks (%) 1
Total for Water At Cost (%) 100
Water At No Cost
All Sources (%) 82
From Hand Pumps (%)' 60
From Local Property Owners" (%) 25
From Public Taps (%) 8
From A Mosque or School (%) 7
Total for Water At No Cost (%) 100
Total for All Current Water Sources (%) 100

 

' These are usually installed inside the household premises.
” From a neighbor, shot, café, or garage.

The table indicates that only 18% of the households in the sample obtained their
water at a monetary cost. This cost is either to cover the purchase of water, the labor to
transport it, or both. In the above table, only the three categories: paying someone to fill
containers, paying someone to fill water tanks, and obtaining water from hand pumps require
minimum or no household labor. Hence, the current water sources utilized by the sample
place a heavy demand on the household’s labor as they depend on household members
transporting water from the source to the home. In Cairo, 90% of this labor is provided by
females with more than 80% of the household water supplied by adults and 10% by female
children (Nadim eta], 1980; Rada, 1994).

The time and effort spent collecting water are determined, among other things, by the

number of trips completed to the water source, the number of people available to help bring

77

water to the home, the distance to the water source, the terrain to be traversed, and the
waiting time at the water source. Table 3.2 provides information on some of these factors.

Table 3.2 Factors Affecting Effort and Time

 

 

Factor Water Water
At No Cost At Cost
Trips Per Week 10 5
Help in Last Trip (Persons) 1.5 1
Time Last Trip (Min.) 56 48
Hours Spent Per Week (HH) 8.4 4
Hours Spent Per Week (fem) 7.56 3.6

 

Those who obtained water at no monetary cost completed more trips to water sources
outside the home but at the same time they had a larger number of household members to
help transport the water they needed. They also spent more time per trip on average. This
might be due to the fact that they used public taps for 8% of their water. In addition to the
substantial waiting time, it might take more than 15 minutes or a kilometer to walk to the
nearest public tap (ibid).

Obtaining water at no monetary cost does not mean that it is cost-less. Other costs,
although difficult to quantify in money terms, can be substantial. In addition to the time and
effort, the embarrassment and humiliation of asking for free water and the inadequate water
quality of some of the free sources are a major concern. Women in focus groups repeatedly
complained about being embarrassed and humiliated when they asked for free water. They
also complaint about the quality of pump water and described it as being “chemically harder”
than network water. Pump water tasted bad, lessened appetite when used for cooking, and

changed color when boiled and used for tea or drinking. Its chemical properties caused hair

78'

loss and made it coarse and dry, and wasted soap if used for washing clothes as it did not
generate any lather (Rada, 1994).

The walk to and from the water source is often over rough terrain and in temperatures
as high as 110° Fahrenheit in the summer. 99% of the households in the sample lived in
buildings that faced unpaved streets. In addition to increasing the effort of walking, dirt
streets can be slippery (especially close to the water source) increasing the risk of falling
while carrying water. This creates a health hazard for the female and her children when they
accompany her to the water source.

The slope of the terrain over which water is transported substantially increases the
amount of effort required by women to haul water. Previous research considered the
maximum possible distances to walk to water sources carrying an average weight in terms
of calories of food energy consumed per day. It was found that the distance has to be
shortened by almost 70% for an increase in slope of only 12 degrees (White et al., 1972). In
some Cairo neighborhoods, women have to walk up the hill carrying their water containers
(N adim et. al, 1980; Rada, 1994) as well as carrying water to upper floors using the stairs.
The average height of the buildings in the sample was 2.97 levels and only 2% of the sample
lived in buildings with elevators. In addition to such effort, the waiting time at the water
source might be substantial, especially at public taps. In some Cairo neighborhoods it
sometimes takes a woman an hour to gain access to a tap (ibid).

Adverse health impacts associated with the lack of a residential water connection can
result from injury due to carrying water loads, injury due to other accidents (at the water
source or during transportation), intake of contaminated water, and possible reductions in

water use for personal hygiene. The last two factors affect women directly and indirectly

79

through their effect on other members of the household (especially children). In Cairo 98%
of the women transport water using a cylindrical container with two handles at the top
usually made of zinc or tin known as the “bastilla”. These containers usually hold 20 to 25
liters and when filled with water weigh more than 20 kilograms (Hoehn and Krieger, 1996;
Rada, 1994; Nadim et.al, 1980). Head loading is the most popular method of carrying
among women in Cairo. Medical studies found that this method is one of the least efficient
in terms of Oxygen consumption per minute (Curtis, 1986). The heavy jars that women
balance on their heads may eventually cause pelvic disorders, complications at childbirth,
headaches, and hair loss and bald spots (Rada, 1994). Also, load carrying might cause knee
problems and feet ache. Women in focus groups repeatedly complained about fatigue and
loss of energy due to hauling water (ibid).

Injuries due to other accidents during transportation include slipped discs, paralysis,
injury to carried children, and broken backs. These are particularly probable when women
walk along slippery terrain. Injuries due to other accidents at the water source can occur due
to violence or effort needed to secure one’s turn in line and successfully bring water to the
home. Violence and fights at the water source are frequent in Cairo due to congestion.
Women in focus groups reported that in the process of getting water, they are exposed to
pushing, stepping on each other’s feet, spilling water on one another, and getting hit by water
containers. To avoid congestion times, women sometimes adapt by obtaining water very late
at night (N adim et al., 1980; Rada, 1994). In Cairo, several women reported that they had
suffered miscarriages (due to being exposed to violence at the tap and being pushed by other
women), bruises, cuts, and wounds (ibid). This violence, although not quantified in money

terms, is a major psychological burden on women.

80

Table 3.3 provides some family characteristics from the sample and compares them
to characteristics of other households in Greater Cairo that are connected to the water
system9.

Table 3.3 Family Characteristics

 

 

Characteristic Connected Sample
Average Family Size (#) 4.7 5.5
Household Water Consumption (Lit/day) 263 115.5
Per-capita Consumption 56' 21"
Females Older than 14 years (#) 1.6 1.6
Children less than 5 years (#) 0.4 0.7
Children less than 5 With Diarrhea in Last 24 13 23

Source (%)

' Lowest estimate for Liters per day for connected households in Cairo (CH2M Hill International,
1990).

” Lowest estimate for Lters per day for unconnected neighborhoods in Cairo (N adim et al., 1980).

The average household size of the sample is higher than that of connected
households. Despite the fact that the sample has a larger family size, total household water
consumption is lower. Per-capita water consumption is 50% lower than the connected
households. Lower water consumption rates can be justified on two grounds: it is not only
that obtaining water is difficult but also disposing of it after use. Only 12% of the sample
reported being connected to the sewer system (as compared to 88% in connected
neighborhoods). Amounts used for personal hygiene might be affected if the household is
not able to secure enough water for daily usage and this affects the health of its members
(IWTC, 1990).

The table also shows that households in the sample reported a lower female to child

ratio of 0.7 than those connected to the water system of 1.6. This might indicate that lesser

 

9 Information on connected households is obtained from Hoehn and Krieger, 1996.

81

 

care is provided to children (if females are the prime providers of childcare). The lack of
enough females to care for children is especially intensified if the limited number of
available females is heavily engaged (in terms of time and effort) with water collection. The
lower ratio might also indicate that children are more often taken to the water source,
increasing their risks of injuries and attracting diseases.

The table also shows that the sample reported 10% higher incidents of diarrhea
disease among children below the age of five. The lack of clean and adequate water is one
of the major causes of incidents of diarrhea and death among children (Oduor, 1992; IWTC,
1990). This affects women indirectly since they are assumed as having the prime role in
health care of children in Egypt (El-Katsha and Watts, 1993). The sample reported an
average travel time to medical facilities to treat diarrhea of 44 minutes and an average
waiting time of 102 minutes to receive such treatment. This time cost, to women and their
households, does not include the money required to receive the treatment. Also, women
might get emotionally distressed when their children get sick or die due to illness.

In the sample respondents had unambiguous positive expectations of the water they
would received after installing a residential water connection. This was especially true for
women. Table 3.4 shows how respondents perceived piped water in comparison to their
current sources.

Table 3.4 Piped Water Compared to Current Sources

 

 

Perception Female Male
. (% yes) (% yes)
Piped Water Makes It Easier to Clean Clothes 95 91
Piped Water is Less Likely to Make Someone Sick 93 91
Piped Water Saves a Lot of Time 94 90
Piped Water is Less Likely to be Dirty 92 84

 

82

An average female respondents had higher expectations of the quality and
convenience of piped water. Compared to male respondents, a higher percentage of females
believed that piped water cleans clothes easier, reduces the risk of getting sick, saves a lot
of time and effort, and contains lesser dirt. This table also suggests the possibility that a
female might be willing to pay more (a premium) for piped water based on her perception
of its quality and convenience.

5.2. Lower Willingness to Pay

The objective of this section is to answer the question “do the high female benefits,
as described above, translate into higher willingness to pay for residential water connections
compared to male respondents?”. To test if willingness to pay differs by gender a probit
model is used to estimate willingness to pay for residential water connections using the
following independent variables: 1) Household Cash Income: is monthly household cash
income (in Egyptian pounds) as measured by the pooled cash earnings of all household
members who worked outside the home for cash. The average number of those who worked
outside the home for cash was two persons per household. This variable does not include
household assets, remittances received from abroad, and other forms of household income.
The average monthly household income for the sample is $213.25. 2) Female: is an
indicator variable that equals one if the respondent is a female. Since the interview was
conducted with heads of households, female refers to the female head of the household. 3)
Rough Terrain: is an indicator variable that describes the topography in the street where the
respondent lived. This variable was based on interviewers’ observations. Interviewers were
asked to note whether the street facing the respondent’s residence is rough or smooth.

Hence, the variable rough terrain equals one if the interviewer reported a rough street. This

83

was seen as an important factor to include based on previous literature (Curtis, 1986; IWTC,
1990). 4) Elevator: is an indicator variable that equals one if the respondent lived in a
building with an elevator and is used to capture some of the workload required to haul water
to upper floors. It is assumed that if those hauling water used the stairs, instead of an
elevator, this increases their workload substantially. 5) Interaction Term: This variable is
generated by multiplying the variables “female” and “household cash income”. Table 3.5
provides the results of using these variables to estimate willingness to pay for residential
water connections.

Table 3.5 Probit Estimates of Willingness to Pay

 

 

Variable Name Estimates (standard errors)
Constant 504.63 (37.14)*
Household Cash Income 0.14 (0.07)*
Female -84.82 (38.03)*
Rough Terrain 144.83 (78.40)‘
Elevator -264.6 (136.39)*
Interaction Term 0.05 (0.10)
Log-likelihood Function Value -407.23
McFadden R2 0.22
Incremental Chi-Square (df) 227.10 (6)
Correct Predictions (%) 75

 

~N=792. (*) indicates significantly different from zero at 5% level.

Table 3.5 shows that all the variables used (other than the interaction term) were
statistically significant at the 5% level and had the expected signs as explained below. Those
who lived in buildings with elevators were willing to pay less for residential connections
than those without ones. The existence of an elevator reduces the effort required to haul
water to upper floors and hence reduces willingness to pay for residential connections. Those
with higher pooled household cash income and lived in building facing rough streets were
willing to pay more for an in-home water connection. The higher the household income the

higher is the willingness to pay. This finding is supported by previous literature (Swarna and

84

Whittington, 1994). Also, rough terrain increase the effort of hauling water and hence their
existence increases willingness to pay for residential connections.

The table also indicates a difference in willingness to pay between male and female
respondents keeping the level of income and other variables constant i.e. it measures the
ceteris paribus effect of changing the gender of the respondent. The coefficient on “female”
is negative. The hypothesis that the difference in willingness to pay due to gender is not
significantly different from zero was tested using a student-t test and was rejected at the 5%
significance level. The above estimated equation assumes a constant willingness to pay
differential across different levels of income and hence represents a difference in intercepts
between male and female respondents. The intercept for male respondents is 504.62 while

that for female respondents is 419.80. This can be seen graphically in Figure 1.

 

Willingness to Pay
WT P,,,
WTPf
419.8
Income

 

Figure 1. Willingness to Pay Differentials by Gender

Using the estimated equation in Table 3.5, willingness to pay for residential water
connections for male respondents and female respondents are calculated. Table 3.6 provides
the estimated results. The table shows that the estimated female willingness to pay for

residential connections is lower than that of males.

85

Table 3.6 Mean Willingness to Pay (WTP)

 

 

(Per month)
Category LE S
Male WTP 20.28 6.07
Female WTP 16.72 5.63

 

5.3. Female Control Over Budget

Tables 3.7 and Table 3.5 shows that all the variables used (other than the interaction
term) were statistically significant at the 5% level and had the expected signs as explained
below. Those who lived in buildings with elevators were willing to pay less for residential
connections than those without ones. The existence of an elevator reduces the effort required
to haul water to upper floors and hence reduces willingness to pay for residential
connections. Those with higher pooled household cash income and lived in building facing
rough streets were willing to pay more for an in-home water connection. The higher the
household income the higher is the willingness to pay. This finding is supported by previous
literature (Swarna and Whittington, 1994). Also, rough terrain increase the effort of hauling
water and hence their existence increases willingness to pay for residential connections.

The table also indicates a difference in willingness to pay between male and female
respondents keeping the level of income and other variables constant i.e. it measures the
ceteris paribus effect of changing the gender of the respondent. The coefficient on “female”
is negative. The hypothesis that the difference in willingness to pay due to gender is not
significantly different from zero was tested using a student-t test and was rejected at the 5%
significance level. The above estimated equation assumes a constant willingness to pay

differential across different levels of income and hence represents a difference in intercepts

86

 

between male and female respondents. The intercept for male respondents is 504.62 while
that for female respondents is 419.80. This can be seen graphically in Figure 1.

Using the estimated equation in Table 3.5, willingness to pay for residential water
connections for male respondents and female respondents are calculated. Table 3.6 provides
the estimated results. The table shows that the estimated female willingness to pay for
residential connections is lower than that of males.

5.3. Female Control Over Budget

Tables 5 .7 and 5.8 indicate that those with no water connections lived in below
average conditions compared to the typical resident of Greater Cairo and that within this
sample females had lesser favorable conditions. Table 5.7 compares the sample to typical
housing conditions in Greater Cairo").

Table 3.7 Characteristics of Dwellings

 

 

Characteristic Typical Sample
Living in Rural Houses(%) 8 2
Living in Apartments(%) 88 64
Number of Rooms 5.6 5.5
Finished Floors (%) 93 79
Finished Walls (%) 93 76
Building Faces Paved Street(%) 38 1

In Home Connection to Sewer(%) 88 12

 

Sample respondents are more likely to live in rural houses in the less densely settled
agricultural-urban fringe of Cairo. Although the sample has below average dwellings (in
terms of structure and access to infrastructures) compared to typical conditions, individuals

in the sample pay higher rent for those dwellings. The average rent paid is LE 40 per month

 

'° Information on typical conditions were obtained from Hoehn and Krieger, 1996.

87

compared to LE 22 for a typical rented dwelling”. The expected selling price of owner
occupied dwellings, however, gave a better indication of dwellings’ values and reflected
local economic conditions. The mean value of owner occupied dwellings is LE 25.4
thousands for the sample compared to LE 37.3 thousands for the typical case. Table 8
provides a comparison between male and female respondents.

Table 3.8 Characteristics of Interviewed Respondents

 

 

Characteristic Female Male
Number Interviewed 501 501
Age (years) 37 41
University Degree (%) 5 10
Primary Education'(%) 33 48
Professional Employment (%) l 3
Housewives 91% -

 

The sample is 50% divided between male and female respondents. These respondents
are heads of households in the sense of making the financial decisions for the family. On
average male respondents were older and had more education. The level of education is also
reflected in employment opportunities. The majority (91%) of the female sample are
housewives who do not work for cash outside the home. Only 1% of females worked as
professionals.

The information presented in the above table indicated that females might be facing
a tighter budget. The fact that 91 % of the females interviewed reported they were housewives
and did not work outside the home for cash suggested that they may be bidding from a
fraction of the total household budget that is allocated for them from earners in the household

for the purpose of domestic spending including water needs.

 

” Rent control laws are less in effect in the urban fringe of Cairo. Hence, landlords are able to charge higher
rents therein (Hoehn and Krieger, 1996).

88

 

To test for whether lower willingness to pay resulted fiom only controlling a fraction
of the household total budget as measured by the variable “household cash income”, the
interaction term defined in the above section is used. From equation (14), the coefficient on
the interaction term (5) needed to be negative and statistically significant to support the
tested hypothesis. Using the results reported in table (5), the coefficient on the interaction
term had the wrong sign and was not significantly different from zero at all levels. Hence,
the data failed to support the hypothesis that females lower willingness to pay for residential
connections results from controlling a fi'action of the total household income.

6. DISCUSSION

Female respondents are willing to pay less than male respondents for residential
water connections. This is despite the fact that such a program provides female respondents
with substantial benefits in the form of time and effort savings as well as improved health
status. The data did not support the hypothesis that the lower female willingness to pay

results from controlling only a fraction of the total household income.

89

 

References

The Economic Benefits of Potable Water Supply Projects to Households in Developing
Countries. Swama V, Whittington D (1994). Asian Development Bank. Economic
Staff Paper Number 53.

“The Sexual Division of Labor, Development, and Women’s Status”. Femandez-Kelly,
Maria (1981). Current Anthropology 22 (4): 414-419.

“Women and Development: Dependency and Exploitation”. Nash, June. 1977. Development
and Change 8: 161-182.

“Equality, Sex, and the Mode of Production”, in G. Berremen ed. Social Inequality:
Comparative and Development Approaches. New York: Academic Press. p. 201-219.

“Changing Modes of Production and Sexual Division of Labor Among the Yoruba”. Afonj a,
Simi. 1981 . Signs 7(2): 299-313.

“Women in Modernizing Societies”. Bossen, Laurel. 1975. American Ethnologist 2(4): 5 87-
601.

“Gender Roles and Social Change: A Mexican Case Study”. Browner, Carol. 1986.
Ethnology 25(2): 89-106.

“Households Among the Diola of Senegal: Should Norms Enter the Front or the Back
Door?”. Linares, O. 1984. In R. Netting, R. Wilk, and E. Arnold (eds), Households:

Comparative and Historical Studies of the Domestic Group. Berkeley: University of
California Press. p. 407-445.

“Recent Research on Women in Development”. Schuster, Ilsa. 1982. Journal of
Development Studies 18 (4): 51 1-535.

Together for Water and Sanitation: Tools to Apply a Gender Approach. Bolt, E. (1994). The
Hague, The Netherlands, IRC International Water and Sanitation Center.

“Estimating Willingness to Pay for Water Services in Developing Countries: A Case Study
of the Use of Contingent Valuation Surveys in Southern Haiti”. Whittington, D., J.
Brisco, et a1. (1990). Economic Development and Cultural Change 38(2).

All You Have to Do is Teach Me. Narrowe, J. (1989). Stockholm, Department of Social
Anthropology.

“Living Without Water”. Nadim, A., Nawal El Messiri, Sohair Mehanna, John Nixon.
(1980). The Cairo Papers in Social Science 3(3).

90

“Assessing Preferences in Cost Benefit Analysis: Reflections on Rural Water Supply
Evaluation in Haiti”. McRae, D. and D. Whittington (1988). Journal of Policy
Analysis and Management 7(2).

Making the Case for the Gender Variable: Women and the Wealth and Well-being of

Nations. Blumberg, R. L. (1989). United States Agency for International
Development.

“The Empowerment of Women: Water and Sanitation Initiatives in Rural Egypt”. El-Katsha,
S. and Susan Watts (1993). Cairo Papers in Social Science 16(2).

Women and Water. International Women's Tribune Center IWTC (1990). New York, United
Nations.

Water T artfi' Study. Ministry of Energy and Water Resources Development “(1985).
Zimbabwe.

“Cairo Water and Wastewater Economic Benefit Assessment”. Hoehn, J. and Douglas
Krieger (1996). Cairo, United State Agency for International Development.

“The Economic Value of Reliable Urban Water Services: A Case Study of Cairo, Egypt”.
Krieger, D. and J. Hoehn (Forthcoming).

“Using Surveys to Value Public Goods: The Contingent Valuation Method”. Mitchell, C.
and R. Carson (1989). Resources for the Future. Washington DC.

“Efficient Estimation Methods for Closed-Ended Contingent Valuation Surveys”. Cameron,
T. and M. James (1987). Review of Economics and Statistics 69: 269-76.

“Environmental Economic Program: Qualitative Research Study”. Rada Research and
Public Relations (1994). Focus Groups Results. Cairo, Rada Research and Public
Relations Company.

Drawers of Water. White, G. F., D. J. Bradley, et a1. (1972). Chicago Press.

Women and the Transport of Water. Curtis, V. (1986). London, Intermediate Technology
Publications.

“Rod El-Farag Distribution System”. CH2M Hill International. (1990). General
Organization For Greater Cairo Water Supply. Vol. II.

“Pockets of Poverty: Linking Water, Health, and Gender-Based Responsibilities in South

Kamwango”. Oduor-Noah Elizabeth, Thomas-Slayter Barbara. (1992). ECONGEN
Case Study Series. Ecology, Community, Organization, and Gender project. Clark
University.

91

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

11[I[iljjjijjjjjjjjjll