NOTE NUMBER 242 P U B L I C P O L I C Y F O R T H E

privatesector
APRIL 2002

A Water Scorecard
Nicola Tynan and Bill Setting Performance Targets for Water Utilities
Kingdom
Using data from 246 water utilities in 51 developed and developing
Nicola Tynan is an
countries, this Note highlights the wide variation in performance on key
assistant professor of
economics at Dickinson indicators: unaccounted-for water, labor costs, the working ratio, service
T H E W O R L D B A N K G R O U P PRIVATE SECTOR AND INFRASTRUCTURE NETWORK

College in Pennsylvania.
coverage, water prices and connection costs, and continuity of service. On
Bill Kingdom is a senior
water and sanitation the basis of the performance of the top 25 percent of developing country
specialist at the World utilities, the Note proposes “best practice” targets for developing countries.
Bank. He leads the
Bank’s efforts in water
The performance of water utilities can be assessed account for 70 percent of asset value but are
and sanitation utility
using four broad measures: efficiency of invest- often neglected in favor of more visible assets.
benchmarking.
ment, efficiency of operations and maintenance, Direct measures of asset maintenance are
Preparation of this Note
financial sustainability, and responsiveness to cus-
was funded by the Energy Figure Unaccounted-for water
tomers. This Note uses data from the World
and Water Department of
the World Bank.
Bank’s Benchmarking Water and Sanitation
Utilities database and from the Asian Develop-
ment Bank to review the performance of a sam-
1 Developed country utilities
Unaccounted-for water
as a share of water
supplied (percent) Utilities
Developing country utilities

ple of utilities. The sample includes utilities in all 0 5 10 15 20 25 30 35

regions and in countries at all income levels; half 0
(123) are in 44 developing countries. The utilities 5
10
range from small (serving a population of less 15
than 125,000) to medium (125,000–500,000) and 20
25
large (more than 500,000), with at least 30 utili- 30
ties from developing countries and 30 from devel- 35
40
oped countries in each category. 45
50
Efficiency of investment 55
60
Investment in new assets should occur only when 65
absolutely necessary—and to ensure efficient, 70
75
long-run operation of existing assets, daily main- 80
tenance is critical. Maintenance is particularly Source: Authors’ calculations, based on data from World Bank (2001) and McIntosh and
important for pipe networks, which typically Yñiguez (1997).
A W A T E R S C O R E C A R D SETTING PERFORMANCE TARGETS FOR WATER UTILITIES

ing country utilities report more than 20 staff

Figure Working ratio
per 1,000 connections. The mean value for

2 0.3
0
Developed country utilities

10 20
Utilities
Developing country utilities

30 40 50
developed country utilities is 2.1 staff.
These high ratios in developing countries
may mean that single water connections are serv-
0.4 ing multiple households. Or they may reflect
0.5 loose employment practices, often a result of
0.6 political interference in the water company’s
0.7
0.8 operation. (Recent water sector reforms, such as
2
0.9 those in Buenos Aires and Manila, show that util-
1.0
1.1
ities can sustain services with significantly fewer
1.2 employees.) The staff ratios achieved by the top
1.3 25 percent of developing country utilities in the
1.4
1.5 sample suggest that a target of 5 or fewer staff per
1.6 1,000 connections is achievable.
1.7 A second indicator of operational efficiency
1.8
1.9 is staff per 1,000 people served, which elimi-
2.0 nates the distortion caused when single water
More
connections serve multiple households. This
Note: See the text for a definition of working ratio.
Source: Authors’ calculations, based on data from World Bank (2001) and McIntosh and
measure too suggests excess labor in developing
Yñiguez (1997). country utilities. So does a third measure com-
normally unavailable, so surrogates are needed. bining wages and staffing to give personnel costs
A crude measure of asset maintenance is as a share of total operating costs: 29 percent in
unaccounted-for water—the difference between developed countries and 39 percent in devel-
water supplied and water sold as a percentage of oping countries.
water supplied (figure 1). This measure captures
not only physical losses but also commercial Financial sustainability
losses, due to inefficient billing or illegal con- Failure to cover costs leads to underinvestment
nections. Thus high levels of unaccounted-for in assets, weakened operations, and declining
water indicate poor system management and
Figure Collection period
poor commercial practices as well as inadequate
pipeline maintenance.
Based on the performance of the top 25 per-
cent of developing country utilities, a target for
3 Developed country utilities
Accounts receivable in
months of sales Utilities
0 5 10 15 20 25 30 35 40 45 50
Developing country utilities

unaccounted-for water of less than 23 percent is 0

recommended. (The mean for developed coun- 1
2
tries is 16 percent.) Achieving this target requires
3
reducing both physical and commercial losses. 4
5
Efficiency of operations and maintenance 6
7
Operational efficiency is defined as the lowest- 8
cost use of inputs—labor, energy, water, and 9
materials—in the daily operation of a utility. 10
11
The most efficient combination of inputs 12
depends in part on local input prices and past 13
capital investment decisions. 14
15
To measure operational efficiency, analysts
16
use ratios of inputs to outputs. One such ratio is 17
staff per 1,000 connections. A high ratio may Source: Authors’ calculations, based on data from World Bank (2001) and McIntosh and
Yñiguez (1997).
indicate inefficient use of staff. Many develop-
 Responsiveness to customers
Connection fees for developing country
Figure utilities Customer focus is assessed on the basis of cover-

4
age, affordability, and quality of service.
Connection fee as
a share of per capita
GDP (percent) Utilities Coverage
0 5 10 15 20 25 In most urban settings a pipe network is the
5
10
cheapest and most effective means of supplying
15 water—whether through individual house con-
20 nections or shared yard connections. The share
25 3
30 of households covered by pipe networks differs
35 significantly between developed and developing
40
country utilities. In developed countries cover-
45
50 age rates exceed 99 percent, and all but two util-
55 ities have 100 percent coverage. In developing
60
More countries coverage rates range from 100 percent
to a low of 18 percent. For sewerage, the worst
Source: Authors’ calculations, based on data from World Bank (2001) and McIntosh and
Yñiguez (1997). coverage rates are even lower. The proposed tar-
get is 100 percent coverage with appropriate levels
service quality. Definitions of cost recovery vary, of service for each household. The top quartile of
with much debate on issues relating to capital developing country utilities in the sample have
asset values and rates of return on investment. achieved this target.
A simple measure of cost recovery is the
working ratio—the ratio of total annual opera- Affordability
tional expenses, excluding depreciation and Two indicators of access to new connections are
debt service, to total annual pretax collections important: the cost of a new connection and the
from billing and subsidies. A working ratio of waiting time for a connection. No data are avail-
more than 1 means that a utility fails to recover able on waiting times, but figures for connec-
even its operating costs from annual revenue, tion fees, roughly normalized against annual
while a ratio of less than 1 means that it covers per capita GDP, show that for some of the
all operating costs plus some or all of its capital
costs. The proposed target working ratio for Figure Affordability of water
developing country utilities is 0.68, the per-
formance achieved by the top quartile of utili-
ties in developing countries as well as the mean
for those in developed countries (figure 2).
5 Developed country utilities
Annual cost of 20 liters
a day as a share of per
capita GDP (percent)
Developing country utilities

Utilities
0 10 20 30 40 50 60
Financial sustainability also requires timely 0.00
collection of payments. A common measure of 0.02
efficiency in this area is the collection period— 0.04
0.06
accounts receivable as a share of annual rev- 0.08
enues, expressed in months of sales (figure 3). 0.10
The recommended collection period is 3 0.12
months or less. At first glance that target looks 0.14
0.16
ambitious: the top quartile of developing coun- 0.18
try utilities in the sample achieves a performance 0.20
of 9.7 months. But closer inspection of the 0.22
0.24
results for the top quartile reveals a clear division 0.26
into clusters, with one group achieving a collec- 0.29
tion period of 4 months or less, and another a 0.30
or more
collection period of more than 8 months. The Source: Authors’ calculations, based on data from World Bank (2001) and McIntosh and
developed country average is 1.8 months. Yñiguez (1997).
A W A T E R S C O R E C A R D SETTING PERFORMANCE TARGETS FOR WATER UTILITIES

show the burden on consumers in developing

Figure Continuity of service
countries and underline the need to cut costs.

6 0
Developed country utilities
Hours of service
a day
30 60
Utilities
Developing country utilities

90 120 150
Quality
The quality of service has several dimensions—
viewpoint
2 water availability, water quality, water pressure,
4 and customer relations. But the only one for
6
8 which the sample provides sufficient data is
is an open forum to
10 water availability, as captured by the continuity
12 encourage dissemination of
of service (hours of service a day). Based on the
14 public policy innovations for
16 performance of the top quartile of developing private sector–led and
18 country utilities, the recommended target is 24
20 market-based solutions for
22 hours a day (figure 6). development. The views
24 published are those of the
Source: Authors’ calculations, based on data from World Bank (2001) and McIntosh and Conclusion authors and should not be
Yñiguez (1997).
The indicators and proposed targets capture a attributed to the World
broad range of performance measures for utili- Bank or any other affiliated
developing country utilities the fees are clearly ties. They are not comprehensive. More work is organizations. Nor do any of
unaffordable (figure 4). In some cases they needed to provide a more complete assessment the conclusions represent
exceed 60 percent of per capita GDP. of utility performance, expanding the measures official policy of the World
The connection fee is often what prevents to governance and accountability, to capital effi- Bank or of its Executive
people from obtaining piped water supplies— ciency, and to better measures of responsiveness Directors or the countries
once connected, consumers can usually pay to the needs of the poor. Still, the key point is that they represent.

their usage fees. The data in figure 4 suggest a the target indicators covered here are being
rule of thumb for connection fees: most utilities achieved by 25 percent of developing countries To order additional copies

in developing countries should charge connec- in the data set. These targets can be achieved by contact Suzanne Smith,

tion fees equivalent to no more than 20 percent the rest by strengthening the focus on customers, managing editor,
Room I9-017,
of per capita GDP. improving governance, providing incentives for
The World Bank,
Water prices depend in part on local condi- utility managers to lift performance—something
1818 H Street, NW,
tions. Because family income, household con- they often lack—and finally, if necessary, raising
Washington, DC 20433.
sumption, and consumers per connection vary tariffs.
considerably among countries, most summary
Telephone:
tariff measures (such as average cost per cubic
001 202 458 7281
meter) are more useful for comparing utilities
Fax:
within countries than for comparing them References
001 202 522 3181
across countries. A rough measure for compar- McIntosh, Arthur C., and Cesar E. Yñiguez, eds. 1997.
Email:
ison across countries, however, is the afford- Second Water Utilities Data Book. Manila: Asian Development
[email protected]
ability of the minimum water requirement set by Bank.
the World Health Organization—the annual World Bank. 2001. Benchmarking Water and Sanitation
cost of 20 liters a day as a share of per capita Utilities database. [http://www.worldbank.org/html/fpd/ Printed on recycled paper
GDP. water/topics/bench_network.html].
Calculations based on the average tariff for
each utility in the sample show that in devel- Nicola Tynan, Economics Department, Dickinson College
oped countries customers of the highest quar- ([email protected]), and Bill Kingdom, Water and
tile of utilities (those charging the highest Sanitation Unit, World Bank ([email protected]).
prices) pay the equivalent of 0.036–0.120 per-
cent of per capita GDP for 20 liters of water a
day (figure 5). In developing countries cus-
tomers of the highest quartile pay more than 0.2
percent of annual per capita GDP. These results
This Note is available online:
www.worldbank.org/html/fpd/notes/