Abstract

This paper provides a long overdue update on the global non-revenue water (NRW) estimates, initially published in a World Bank publication more than 10 years ago. The authors use a similar approach based on latest data to estimate the volume of water lost by water utilities around the world. The conclusion of this study is that the 2006 estimates were extremely conservative and that realistic NRW estimates are significantly higher. The global volume of NRW has been estimated to be 346 million cubic metres per day or 126 billion cubic metres per year. Conservatively valued at only USD 0.31 per cubic metre, the cost/value of water lost amounts to USD 39 billion per year. Not only is this an enormous financial concern, but elevated NRW also detracts from water utilities, in a time of increasing scarcity and climate change, from reaching their goals of full service coverage, at a reliable level of service at an affordable price.

INTRODUCTION

In 2006, Liemberger (together with Kingdom and Marin) calculated the volume of water losses around the word and the approximate cost of these water losses. Since then, the numbers have been quoted in presentations and articles around the world. In 2010, Liemberger (together with Frauendorfer) published updated numbers for Asia, which clearly indicated that the assumptions for the 2006 estimates were far too conservative. The purpose of this paper is to calculate new estimates based on the latest available data and provide a new, reliable source of information.

Figure 1

Regional NRW levels.

Figure 1

Regional NRW levels.

Wyatt worked extensively on the analysis of the non-revenue water (NRW) situation in Latin America and the Caribbean (LAC), which will be published in a forthcoming Technical Guide on NRW by the Inter-American Development Bank (Wyatt 2017). His extensive data set has been used to check and calibrate the global NRW model.

With this study, the authors would like to highlight the seriousness of the global NRW problem. Chronic water losses have been the hallmark of water utilities in most parts of the world over recent decades. This may not have been a large concern during an era of assumed plenty. But the rapid growth of the world's towns and cities, coupled with the negative impact of climate change, has meant that there is much less water available than in the past. If the world's volume of NRW was reduced by only one-third, the savings would be sufficient to supply 800 million people (assuming a per capita consumption of 150 litres per day). But this is not all; reducing NRW will improve water service reliability, water supply to the urban poor, water quality, decrease energy consumption, and in some cases delay water supply capacity expansions.

The authors are providing all country data and data sources in the Appendix (available with the online version of this paper), so that the assumptions are transparent and might be improved by others who have access to good data for specific countries.

METHODOLOGY

General

WHO/UNICEF JMP (Joint Monitoring Program) is the custodian of global data on drinking water, sanitation and hygiene (WASH). Total country population with access to piped water supply has been extracted from WASH data.

Average per capita consumption has been determined on a country-by-country basis and the total volume of domestic water consumption has been calculated. A provisional volume of 30% had been added for non-domestic water use.

In the following calculation, it was necessary to use the assumed percentage of NRW as it was the only widely published data available, but as that performance indicator has well-documented shortcomings, the results will be presented as NRW in litres per capita per day:

PS Supplied population   
PCC Per capita consumption (l/c/d) 
CD Domestic consumption (m3/d) 
CND Non-domestic consumption (m3/d) 
CT Total consumption (m3/d) 
NRW Non-revenue water (m3/d) 
SIV System input volume (m3/d) 
%NRW Non-revenue water (% of SIV) 
PS Supplied population   
PCC Per capita consumption (l/c/d) 
CD Domestic consumption (m3/d) 
CND Non-domestic consumption (m3/d) 
CT Total consumption (m3/d) 
NRW Non-revenue water (m3/d) 
SIV System input volume (m3/d) 
%NRW Non-revenue water (% of SIV) 
The different volumes are calculated as follows:  
formula
 
formula
 
formula
 
formula
 
formula
which can then be brought into one equation:  
formula

Per capita consumption

The main data sources are:

  • IBNET

  • IWA

  • AWWA

  • EU

Unfortunately for many countries no information, or at least no reliable information, could be found and the authors had to make estimates based on personal experience, other sources of public information, and data from countries with similar conditions. Details can be found in the Appendix (available with the online version of this paper).

Level of NRW

IBNET was the most important source of information for the level of NRW expressed as a percentage of system input volume, but data quality is partly problematic. Some country data are based only on very few, or even only one, water utility and sometimes the latest available data are already more than a few years old, notably in LAC but also in all other regions. But despite all of this, IBNET is the only global source for NRW data. For all the other countries which are not included in the IBNET data set, the authors had to make estimates based on personal experience and data from countries with similar conditions. Details can be found in the Appendix.

Calculating the value/cost of NRW

The cost or value of one cubic metre of NRW for a water utility depends on a number of factors:

  • The ratio between commercial (apparent) and physical (real) losses.

  • A reduction in commercial losses will lead to increased revenues – consequently commercial losses will be valued using the average tariff. (It has become accepted practice in the USA that sewer service charges, if billed as a function of water volume consumed, are also included in the valuation of apparent losses.)

  • If some of the recovered physical losses can be sold to existing or new customers, then that portion of NRW can also be valued using the average tariff.

  • In case there is no unsatisfied demand, any reduction of physical losses will only lead to a reduction in variable water abstraction, treatment and distribution cost – and those will of course vary extremely depending on the water source – gravity supply of spring water on the one hand and desalination and/or pumping to high altitudes on the other hand.

  • If additional water sources will be needed to meet the increasing demand, then the capital and operating cost of such future water sources will have a significant impact on the value of NRW.

Consequently, it is very complex to calculate the average value of NRW on a country level. Specific studies are required – by location. Therefore, an extreme simplification has been used to provide a country average value of NRW which represents a mix of variable production cost and average tariff.

A simple empirical formula has been developed which is based on the assumption that even in the poorest countries the (realistic) value of a m3 of NRW will not be less than USD 0.20 and then it will increase in some relation to the country's per capita Gross Domestic Product (GDP):  
formula

This formula means that NRW in the poorest countries is valued close to USD 0.20 (for example for Burkina Faso (Africa)) while the value would be USD 0.38 for the USA.

More research is needed to improve this method and try to calibrate with good country data.

Methodology used in Wyatt/IDB

Country-level results from Wyatt/IDB were used to cross-check the results of the methodology used in this paper.

The objective of the Wyatt/IDB effort was to assemble a regional database of NRW variables and indicators, with the following considerations:

  • Include all countries in the region, to the extent possible

  • Focus on urban areas

  • Assemble data from a sample of utilities in each country across a range of sizes and climatic conditions

  • Utilize primary or secondary data which are internally consistent (cross-check calculated indicators and look at trends for inconsistencies).

Data were collected for 109 water utilities in 28 countries. Many of the smaller countries have a single national water supplier, which facilitated data collection. Sources included water utility websites and annual reports, country studies and sector assessments, regulatory documents, IDB, CAF, CDB or World Bank loan preparation documents, primary data collection from field visits, a large but somewhat out of date website/database developed in 2008 by the World Bank and IDB, and the following documents which contain multi-country data: Andres & Guash (2009), Andres et al. (2013), Ducci & Garci (2013), ADERASA (2014), Lentini (2015), Janson (2017).

Data were collected on the following parameters:

  • Situational: connections, mains length, median household income, water stress

  • Operational: billed volume, continuity, pressure (some estimated), extent of micro-metering, burst rates, staff, complaints

  • Financial: revenues, operating costs, tariffs, collection efficiency

  • Water balance: approximate split of NRW components (some estimated).

Data were analyzed to determine the following parameters, which correspond well to this broader study:

  • Billed Water Use

    • o Volume: L/capita/day, L/connection/day, Total in 1,000 m3/day

  • Apparent Loss Indicators

    • o Volume: L/conn/day, Total in 1,000 m3/day

    • o Value: USD/conn/day, Total in 1,000 USD/day

  • Real Loss Indicators

    • o Volume: L/conn/day, Total in 1,000 m3/day

    • o Value: USD/conn/day, Total in 1,000 USD/day

  • NRW Indicators

    • o Volume: L/conn/day, Total in 1,000 m3/day

    • o Value: USD/conn/month, Total in 1,000 USD/day

  • Financial Indicators

    • o Unit Water Cost: USD/m3 produced

    • o Estimated Variable Production Cost: USD/m3 produced

    • o Total Water Cost: USD/m3 sold

    • o Effective Average Tariff: USD/m3 sold.

RESULTS AND DISCUSSION

The global volume of NRW has been calculated to be 346 million cubic metres per day or 126 billion cubic metres per year. To put this in some perspective, this annual volume is about 70% of the average flow of the Niger River – the principal river in West Africa, and nearly 50% of the average flow of the Ganges River in India. But more importantly, the aggregate NRW is 30% of water system input volumes across the world.

The newly introduced NRW indicator – litres/capita/day – was computed by country and by region to understand the level of NRW, independent of country or region size. The introduction of this indicator was necessary because the normal operational water loss performance indicators, like NRW per service connection per day or per kilometre pipeline per day cannot be used because data on network length or number of connections are in most cases not available at country level. The results for the different regions can be seen in Table 1. Details for each region on a country-by-country basis can be found in the Appendix (available with the online version of this paper).

Table 1

NRW volume and cost/value per region

Volume of NRW
Average levelCost/value of
Million m3/dayBillion m3/yearof NRW Litres/capita/dayNRW Billion USD/year
Sub-Saharan Africa 14.1 5.2 64 1.4 
Australia and New Zealand 1.0 0.3 36 0.1 
Caucasus and Central Asia 8.0 2.9 152 0.8 
East Asia 53.0 19.3 42 6.2 
Europe 26.8 9.8 50 3.4 
Latin America and Caribbean 69.5 25.4 121 8.0 
Middle East and Northern Africa 41.2 15.0 96 4.8 
Pacific Islands 0.5 0.2 211 0.1 
Russia, Ukraine, Belarus 9.5 3.5 65 1.1 
South Asia 63.4 23.2 93 6.0 
Southeast Asia 18.4 6.7 81 2.0 
USA and Canada 40.7 14.8 119 5.7 
Total 346 126 77 39 
Volume of NRW
Average levelCost/value of
Million m3/dayBillion m3/yearof NRW Litres/capita/dayNRW Billion USD/year
Sub-Saharan Africa 14.1 5.2 64 1.4 
Australia and New Zealand 1.0 0.3 36 0.1 
Caucasus and Central Asia 8.0 2.9 152 0.8 
East Asia 53.0 19.3 42 6.2 
Europe 26.8 9.8 50 3.4 
Latin America and Caribbean 69.5 25.4 121 8.0 
Middle East and Northern Africa 41.2 15.0 96 4.8 
Pacific Islands 0.5 0.2 211 0.1 
Russia, Ukraine, Belarus 9.5 3.5 65 1.1 
South Asia 63.4 23.2 93 6.0 
Southeast Asia 18.4 6.7 81 2.0 
USA and Canada 40.7 14.8 119 5.7 
Total 346 126 77 39 

It does not come as a surprise that the regional differences are significant. The lowest NRW levels (36 l/capita/d) can be found in Australia and New Zealand, which is due to the big water loss reduction efforts that have been made during the last 10 to 15 years in the attempt to better cope with the long droughts in Australia. The Pacific Islands have the highest NRW level (211 l/capita/d) but data quality is limited and therefore this number is questionable. The average level of NRW in LAC is 121 l/capita/d (Figure 1).

A comparison of the model's results to actual country data from the Wyatt/IDB database shows a very good correlation. The weighted average from the new NRW model is 116 l/capita/day compared with 113 l/capita/day estimated by Wyatt/IDB, which is a difference of only 2% (see Table 2).

Table 2

Comparison of the NRW model results with data from selected LAC countries

 Population servedNRW in l/capita/d
NRW in 1,000 m3/day
Percent difference
ModelWyatt/IDBDifferenceModelWyatt/IDBDifference
Argentina 43,019,152 183 189 7,894 8,113 −219 3% 
Bahamas 369,918 98 98 36 36 0% 
Barbados 278,577 286 300 14 80 84 −4 5% 
Belize 297,949 37 37 11 11 0% 
Bolivia 7,915,294 70 63 −7 554 502 52 −9% 
Brazil 199,750,809 94 90 −5 18,872 17,879 993 −5% 
Cayman Islands 51,572 53 52 0% 
Chile 17,847,512 91 84 −6 1,620 1,505 115 −7% 
Colombia 42,381,111 118 107 −11 4,997 4,547 450 −9% 
Costa Rica 4,768,368 206 220 14 982 1,048 −67 7% 
Dominica 54,444 69 97 28 3.8 5.3 −1.5 40% 
Ecuador 13,833,220 173 174 2,388 2,403 −15 1% 
El Salvador 5,367,444 106 104 −2 568 556 12 −2% 
Grenada 96,055 104 91 −13 10.0 8.7 −1.8 −13% 
Guyana 506,840 332 348 17 168 176 −8 5% 
Honduras 7,213,368 111 106 −4 798 768 30 −4% 
Jamaica 2,259,016 289 261 −28 652 590 62 −10% 
Mexico 119,478,951 104 99 −5 12,426 11,817 609 −5% 
Nicaragua 4,234,942 206 208 874 882 −7 1% 
Panama 3,617,503 204 327 123 738 1,183 −445 60% 
Paraguay 5,936,257 206 193 −12 1,220 1,147 74 −6% 
Peru 26,032,349 94 97 2,437 2,528 −92 4% 
Puerto Rico 3,465,481 349 357 1,209 1,238 −29 2% 
St Lucia 176,678 186 205 19 33 36 −3 10% 
St Vincent and the Grenadines 101,763 94 104 10 10 11 −1 11% 
Suriname 364,343 69 142 73 25 52 −26 105% 
Trinidad and Tobago 1,253,202 310 359 49 389 450 −61 16% 
Uruguay 3,384,329 152 147 −5 514 496 18 −3% 
Total Population 514,056,446        
Weighted Average NRW
l/capita/day 
 116 113 − 3     
Total NRW Volume
1,000 m3/day 
    59,510 58,070 1,430 − 2.4% 
 Population servedNRW in l/capita/d
NRW in 1,000 m3/day
Percent difference
ModelWyatt/IDBDifferenceModelWyatt/IDBDifference
Argentina 43,019,152 183 189 7,894 8,113 −219 3% 
Bahamas 369,918 98 98 36 36 0% 
Barbados 278,577 286 300 14 80 84 −4 5% 
Belize 297,949 37 37 11 11 0% 
Bolivia 7,915,294 70 63 −7 554 502 52 −9% 
Brazil 199,750,809 94 90 −5 18,872 17,879 993 −5% 
Cayman Islands 51,572 53 52 0% 
Chile 17,847,512 91 84 −6 1,620 1,505 115 −7% 
Colombia 42,381,111 118 107 −11 4,997 4,547 450 −9% 
Costa Rica 4,768,368 206 220 14 982 1,048 −67 7% 
Dominica 54,444 69 97 28 3.8 5.3 −1.5 40% 
Ecuador 13,833,220 173 174 2,388 2,403 −15 1% 
El Salvador 5,367,444 106 104 −2 568 556 12 −2% 
Grenada 96,055 104 91 −13 10.0 8.7 −1.8 −13% 
Guyana 506,840 332 348 17 168 176 −8 5% 
Honduras 7,213,368 111 106 −4 798 768 30 −4% 
Jamaica 2,259,016 289 261 −28 652 590 62 −10% 
Mexico 119,478,951 104 99 −5 12,426 11,817 609 −5% 
Nicaragua 4,234,942 206 208 874 882 −7 1% 
Panama 3,617,503 204 327 123 738 1,183 −445 60% 
Paraguay 5,936,257 206 193 −12 1,220 1,147 74 −6% 
Peru 26,032,349 94 97 2,437 2,528 −92 4% 
Puerto Rico 3,465,481 349 357 1,209 1,238 −29 2% 
St Lucia 176,678 186 205 19 33 36 −3 10% 
St Vincent and the Grenadines 101,763 94 104 10 10 11 −1 11% 
Suriname 364,343 69 142 73 25 52 −26 105% 
Trinidad and Tobago 1,253,202 310 359 49 389 450 −61 16% 
Uruguay 3,384,329 152 147 −5 514 496 18 −3% 
Total Population 514,056,446        
Weighted Average NRW
l/capita/day 
 116 113 − 3     
Total NRW Volume
1,000 m3/day 
    59,510 58,070 1,430 − 2.4% 

However, it has to be noted that there are some countries with significant differences:

  • Dominica – high consumption from cruise ships

  • Panama – higher consumption (wastage) due to low level of metering, high illegal consumption

  • St Lucia – high consumption from cruise ships

  • Suriname – most likely because Wyatt/IDB only had data for one region of the country, a predominately urban region, which is probably not representative of all piped water systems in the country.

Data from the 10 largest countries, each with a population of more than five million, show a similarly good correlation, with a difference of −3.7% in the total volume of NRW. In none of the countries, does the difference exceed ±10% (Table 3).

Table 3

Comparison of data of the 10 largest countries in the LAC region

Ten largest LAC countriesPopulation servedNRW in l/capita/d
NRW in 1,000 m3/day
Percent difference
ModelWyatt/IDBDifferenceModelWyatt/IDBDifference
Brazil 199,750,809 94 90 18,872 17,879 993 −5.3% 
Mexico 119,478,951 104 99 −5 12,426 11,817 609 −4.9% 
Argentina 43,019,152 183 189 7,894 8,113 −219 2.8% 
Colombia 42,381,111 118 107 −11 4,997 4,547 450 −9.0% 
Peru 26,032,349 94 97 2,437 2,528 −92 3.8% 
Chile 17,847,512 91 84 −6 1,620 1,505 115 −7.1% 
Ecuador 13,833,220 173 174 2,388 2,403 −15 0.6% 
Bolivia 7,915,294 70 63 −7 554 502 52 −9.4% 
Honduras 7,213,368 111 106 −4 798 768 30 −3.7% 
El Salvador 5,367,444 106 104 −2 568 556 12 −2.1% 
Total 482,839,210        
% of Region 94%        
Average NRW   109 105 − 4         
Total NRW         52,550 50,620 1,930 − 3.7% 
% of Region         88% 87%     
Ten largest LAC countriesPopulation servedNRW in l/capita/d
NRW in 1,000 m3/day
Percent difference
ModelWyatt/IDBDifferenceModelWyatt/IDBDifference
Brazil 199,750,809 94 90 18,872 17,879 993 −5.3% 
Mexico 119,478,951 104 99 −5 12,426 11,817 609 −4.9% 
Argentina 43,019,152 183 189 7,894 8,113 −219 2.8% 
Colombia 42,381,111 118 107 −11 4,997 4,547 450 −9.0% 
Peru 26,032,349 94 97 2,437 2,528 −92 3.8% 
Chile 17,847,512 91 84 −6 1,620 1,505 115 −7.1% 
Ecuador 13,833,220 173 174 2,388 2,403 −15 0.6% 
Bolivia 7,915,294 70 63 −7 554 502 52 −9.4% 
Honduras 7,213,368 111 106 −4 798 768 30 −3.7% 
El Salvador 5,367,444 106 104 −2 568 556 12 −2.1% 
Total 482,839,210        
% of Region 94%        
Average NRW   109 105 − 4         
Total NRW         52,550 50,620 1,930 − 3.7% 
% of Region         88% 87%     

The new model has now been used to run a simulation with 2005 data to compare the result with the World Bank publication (Kingdom et al. 2006) which used 2005 data. The result shows a significant difference which can be explained by the conservative assumptions made in 2006 and the fact that the first model used the supplied urban population only, while this model does not differentiate between urban and rural and includes all piped water supply systems. The difference with the model published by the Asian Development Bank (Liemberger & Frauendorfer 2010) has similar reasons (for all comparisons see Table 4).

Table 4

Annual NRW volume – comparison to previous NRW models

Billion m3 per year2005
2009
2016
WB publicationNew modelADB publicationNew modelNew model
World 48.6 97.5   126 
Asia and Pacific Islands   28.7 47.2 64 
Billion m3 per year2005
2009
2016
WB publicationNew modelADB publicationNew modelNew model
World 48.6 97.5   126 
Asia and Pacific Islands   28.7 47.2 64 

Table 1 also provides estimates of the cost/value of the NRW, by region. Estimating the cost/value of global NRW has been as problematic in the previous models as it is in this model. The global cost/value of NRW has been calculated to be USD 39 billion per year but the number can only be understood as a rough estimate. However, it is a clear indication that the high levels of NRW have a massive negative financial impact on the water supply sector.

If global NRW was reduced by one-third (115 million cubic metres per day), the annual financial benefit would be around USD 13 billion. Assuming an average NRW reduction cost of USD 600 per m3/day NRW reduction, the investment required to achieve this reduction would be USD 69 billion – a payback time of just over 5 years. This may sound too optimistic, and based on the authors' experience, payback times between 7 and 10 years are in many cases more realistic, but it is still difficult to understand why water utilities and governments are so reluctant to invest in NRW reduction.

It is also important to note that the 115 million m3/day would serve 800 million people, assuming an average consumption of 150 litres/capita/day. This could address the needs of customers who currently do not have service, expanding current coverage, or could serve the needs of future customers, allowing water source expansions to be postponed or cancelled.

The analysis in Wyatt/IDB indicated that of the 28 countries, 26 could have projects to significantly reduce NRW in the sample utilities, with a payback period of less than 10 years. In 16 of those countries, the payback period would be less than 5 years. The total cost of the projects would be about USD 11 billion. The countries with less favourable payback periods suffer from low tariffs. The resulting water savings could provide water to over 19 million new water connections, which could address current local coverage issues, or cover urban growth for typically 8 to 10 years (assuming a 4% urban growth rate).

CONCLUSIONS

This new analysis of global levels of NRW shows that the current estimated volume is far higher than previously estimated. This is partly due to over-conservative estimates used in previous estimates, as well as growth in population and expansion of water supply systems. The results of this study place the volume of NRW at 126 billion m3/year, which has a financial cost/value of USD 39 billion/year.

The model has been found to provide very similar results to another study of NRW levels in Latin America – based on a sampling approach. Nonetheless more data and more accurate data would help refine the numbers.

Given climate change, expanding populations' full coverage of improved water service is still a large global challenge. However, NRW can provide many benefits – including reduced operating costs increased revenues, better water resource efficiency and expanded water supply at a cost far lower than new water production facilities.

REFERENCES

REFERENCES
Andres
L. A.
&
Guash
J. L.
2009
Understanding Sector Performance: The Case of Utilities in Latin America and the Caribbean
.
Report #53380
,
World Bank
,
Washington, DC, USA
.
Andres
L. A.
,
Schwartz
J.
&
Guash
J. L.
2013
Uncovering the Driver of Utility Performance – Lessons from Latin America and the Caribbean on the Role of the Private Sector, Regulation, and Governance in the Power, Water, and Telecommunication Sectors
.
World Bank
,
Washington, DC, USA
.
doi: 10.1596/978-0-8213-9660-5
.
Asociacio de Entes Reguladores de Agua Potable y Saneamiento de Las America – ADERASA
2014
Informe Annual 2014 Grupo Regional de Trabajo de Benchmarking (GRTB)
.
ADERASA
,
Buenos Aires
,
Argentina
.
Ducci
J.
&
Garci
L. C.
2013
Principlaes indicadores financieros de entidades prestadoras de servicios de agua potable y saneamiento en America Latina y el Caribe
.
Technical Note # IDB-TN-521
,
InterAmerican Development Bank
,
Washington, DC, USA
.
Janson
N.
2017
Current governance and performance of the water supply and sanitation sector in the Caribbean
. In:
IDB/GOJ Conference: Financing Water Utilities of the Caribbean – Towards Efficient and Financially Sustainable Services
,
Kingston, Jamaica
,
2017
.
Kingdom
W.
,
Liemberger
R.
&
Marin
P.
2006
The Challenge of Reducing Non-Revenue Water (NRW) in Developing Countries – How the Private Sector Can Help: A Look at Performance-Based Service Contracting
.
WSS Sector Board Discussion Paper #8
,
World Bank
,
DC, USA
.
Can be downloaded from the World Bank's webpage
.
Lentini
E
, .
2015
El Futuro de Los Servicios de Agua y Saneamiento en America Latina – Desafios de los Operadores de Areas Urbanas de mas de 300,000 habitantes
.
InterAmerican Development Bank, CAF and ADERASA
,
Washington, DC, USA
.
Liemberger
R.
&
Frauendorfer
R.
2010
The Issues and Challenges of Reducing Non-Revenue Water. Asian Development Bank, Mandaluyong, Philippines
.
WHO/UNICEF JMP (Joint Monitoring Program)
.
Wyatt
A. S.
2017
Improved Non-Revenue Water Management – A Holistic Integrated Approach to NRW Programs
(draft).
InterAmerican Development Bank
,
Washington, DC
,
USA, publication forthcoming
.

Supplementary data