performance improvement planning designing an effective ......1 tata steel (formerly the tata iron...

Reducing nonrevenue water and maintaining it at target levels is important for any service providerlooking to improve quality of service, financial soundness, and creditworthiness. In the longer term,the effectiveness of nonrevenue water programs is a function of managerial efficiency andinstitutional accountability.

Performance Improvement Planning

Designing an Effective LeakageReduction and Management Program

April 2008

Field Note

The Water and Sanitation Programis an international partnership forimproving water and sanitation sectorpolicies, practices, and capacities toserve poor people

More than 40 percent of waterproduced in many Indian cities does notearn any revenue, be it water lost beforereaching the consumer or high volumesof water not being billed for or both,therefore contributing to poor costrecovery and hence poor service qualityand coverage. By no means could thisestimate be considered accurate, giventhe absence of any meters—bulk orcustomer meters—in most cities.Ineffective metering is not the only causeof nonrevenue water; old and corrodedpipes and fittings, poor billing, freeflowing services to many, and a lack ofwillingness to improve upon billing andcollection practices also contribute to it.Many providers also lack the ability tounderstand the baseline situation ofNRW—a critical first step if one is tomove towards effective NRW reduction.

High levels of nonrevenue water arealso a result of poor institutionalarrangements under which serviceproviders currently operate. Theincentives to be more accountable, tomanage and perform better, andto have a revenue strategy thatencourages a commercial orientation toservices currently remain very poor inthe country. The enabling financialincentives for encouraging serviceimprovements are also lacking, sinceproviders count on virtuallyunconditional financial support fromgovernment or from subsidies withinmunicipal accounts.

High levels of nonrevenue water result from huge volumes of water lost throughleaks or water not invoiced to customers or both, and this seriously affectsfinancial viability of water providers.

Executive SummaryOne of the principal concerns today for the deteriorating qualityof water supply and sanitation services in India is the highlevels of nonrevenue water (NRW)—the difference between theamount of water put into the distribution system and the amountof water billed to consumers. High levels of NRW result fromhuge volumes of water lost through leaks or water not invoicedto customers or both, and this seriously affects financialviability of water providers through lost revenues, increasedoperational costs and, eventually, increased capital costs.

Reducing nonrevenue water levels does not necessarilycompromise on a service provider’s ability to subsidize servicesfor the poorer sections of the population; rather, it allows forimproved transparency and accountability, and for ensuringbetter targeting of subsidies such that they are more equitableand actually reach poor people.

Reducing NRW and maintaining it at target levels is hence animportant task for any service provider looking to improvequality of service, financial soundness, and creditworthiness.The approach to reducing NRW includes a set of activitiesaimed at the optimization of water supply through improvedoperations, maintenance, and sound management practices ofdistribution networks as well as governance and managementreforms through specific operational and commercial strategies.However, if NRW programs are to remain sustainable in thelong run, they must target institutional and organizationalreform while encouraging efficiencies through technicaland managerial improvements and enhancing humanresource capacity.

This field note captures the core principles for the effectiveimplementation of nonrevenue water programs through realworld examples of service providers implementing suchprograms in India and Vietnam. The case studies demonstratethat successful leakage reduction and its effective andsustainable implementation has not been a result of erratic andirregular technical exercises to reduce physical losses.

On the contrary, these nonrevenue water programs have beensuccessful because of the continuous effort to control physicaland commercial losses while also bringing in managerialefficiency and institutional accountability.

2

Context

Performance ImprovementPlanning for UpgradingUrban Water Services

Performance Improvement Planning:Designing an Effective LeakageReduction and Management Program

Improvingand SustainingNRW LevelsReducing nonrevenue water, and henceimproving cost recovery, is an importanttask for any service provider since it isconsidered a crucial step towardsimproving its financial soundness andcreditworthiness. In spite of its potentialbenefits, NRW reduction is not easy toimplement, especially in an intermittentservice environment. Reducing NRW isnot technically difficult; it is, however,challenging in a governance sense: forinstance, illegal connections thatcontribute to NRW can only beeliminated when utilities and serviceproviders have autonomy anddiscipline, and when they are madeaccountable for keeping illegalconnections under control. Theapproach to reducing NRW levelsincludes a set of technical programsand activities aimed at the optimizationof water supply through improvedoperations and maintenance, soundcommercial practices and networkrefurbishment, rehabilitation and

management established in themanagerial, organizational, andinstitutional environment.

The field note uses case studies todemonstrate that successful leakagereduction is not a result of sporadic andirregular technical exercises to reducephysical losses. Such programs shouldrather be a continuous effort and must

control commercial losses since thelatter helps improve the revenue streamof the water company almostimmediately. These programs will,however, need to be implemented inthe broader context of tackling themanagerial, organizational, andinstitutional environment so that theyremain sustainable and effective in thelong run.

3

Box 1: The Criticality of Nonrevenue Water Management

The basic aim of the performance improvement series is to help water utilities and service providers understand andadopt mechanisms that promote cost recovery and sustainable revenue strategies, as well as help achieve financiallyviable and sustainable improved services. The objective is to be able to focus not only on specific performanceimprovement areas by advancing technical, commercial, and operational efficiency—such as leak reduction, billing andcollection, customer service, and tariff setting, among others—but also ensure that such improvements remainsustainable and viable in the long term through arrangements such as performance agreements, monitoring,and evaluation.

This issue, No. 3 of the series, discusses the importance of the reduction and management of nonrevenue water (NRW)as a critical step towards improving the financial soundness of any water utility. Its objective is to promote, among Indianwater providers, a good understanding of the causes of NRW and ways to design and manage a system to control it. Theseries also showcases some real world examples of NRW management in India and East Asia.

Box 2: Nonrevenue Water

Nonrevenue water is the differencebetween system input volumes andbilled authorized consumption. Itconsists of:

• Unbilled authorized consumption(including water for fire fighting orfree water distributed at standpipesor provided to religious institutions).

• Apparent losses (includingunauthorized consumption andmetering inaccuracies).

• Real losses (including leakagesfrom transmission or distributionmains, leakages and overflow fromutility storage and balance tanks,and leakages in reticulationsystems to the point of metering).

Understandingthe Magnitudeof the ProblemPart of the problem in understandingthe magnitude of NRW lies in the lackof a meaningful standard approach toestimate and report NRW levels. Verylittle data is available because waterutilities lack the systematic reportingand monitoring systems, such as watertable calculations, for assessing waterlosses and NRW performance. A waterbalance calculates NRW into itsrespective components and helpsidentify where the main problem areaslie. A water audit helps review the watersystem and assesses performance byquantifying total water losses andleakages in the network as thedifference between the water input todistribution and the sum of the variouscomponents of water consumed. Acomplete system appraisal helps arriveat a water balance calculation, throughan estimation of real losses, apparentlosses, and unbilled authorizedconsumption (Table 1).

As subsequent case studies indicate,water audits and system appraisalswere undertaken so that providerscould understand and analyze themagnitude of the problem, identify theareas that required specific attentionand undertake technical andcommercial measures to target andkeep NRW under control. A water auditin Jamshedpur indicated huge waterlosses in one of the main trunk lines ofthe network which, when repaired,immediately brought down the levels ofNRW. Similarly, in Bangalore, a wateraudit helped determine that maximumleakages were occurring at the point ofservice connections, which were thenaccordingly controlled.

Box 3: Synopsis of Three Case Studies

Bangalore, capital of the Indian state of Karnataka, has a population of about5.3 million (2007). The city’s water needs are met by the Bangalore WaterSupply and Sewerage Board, a publicly owned water utility. It isresponsible for providing water supply, sewerage system, and sewagedisposal as well as for preparing and implementing plans and water supplyschemes in its service area.

The Board was losing over 40 percent of its water until it undertook somerobust strategies for reducing and controlling water leakages that included apilot nonrevenue water project. These have helped the Board in increasingservice coverage and quality as well as improving the commercial viability ofwater services.

Jamshedpur is an industrial city in the East Singhbhum district of Jharkhandin India. It has a total population of about 0.7 million people and is home tothe first private iron and steel company of the country. The JamshedpurUtilities and Services Company Limited is a wholly owned subsidiary ofprivate company Tata Steel that was created on August 25, 2003, out of TataSteel’s Town Division, to improve the quality of civic services in Jamshedpur.1

The service provider in Jamshedpur provides comprehensive municipalservices including water and wastewater management to the township. It hasrecently undertaken various performance improvement measures including aproactive bulk metering practice and a pilot District Metered Area forcontrolling water losses in its distribution system. As a result of theseinitiatives the utility has been able to serve new neighborhoods with watersupply through saved water from controlling nonrevenue water.

The Haiphong Water Supply Company is a public utility that provideswater supply services to the urban population of Haiphong city—the thirdlargest city in Vietnam—and has a population of about 1.7 million. It is ownedby Haiphong Peoples’ Committee and is regulated by the Department ofTransport and Public Work Services. The city was receiving water supply of325 lpcd (or liters per capita per day) but, given very high levels ofnonrevenue water, the city’s demands were not being met.

The company piloted its ‘Phuong model’ in 1993 and simultaneously instilledinstitutional and management changes (new utility management team,considerable changes in the organizational structure and operating rules andmanagement style, including administrative decentralization), all of whichhelped achieve reduced nonrevenue water (from 70 percent in 1993 to32 percent in 2002), increased water revenues of 5.5 times during1993–2001 and improved quality and standards of service.

1 Tata Steel (formerly the Tata Iron and Steel Company Limited, TISCO) is Asia’s first and India’s largest integrated private sector steelcompany that was established in 1907 in the city of Sakchi, which was originally the name of the town that came to be morefamously known as Jamshedpur.

A water balance calculates nonrevenue water into its respectivecomponents and helps identify where the main problemareas lie.

4



System input volume Authorized Billed metered(measured by input consumption consumptionmeters, after checkingfor their accuracy)

Billed meteredconsumption

Billed unmeteredconsumption

Unbilled meteredconsumption

Unbilled unmeteredconsumption

Unauthorized consumption

Metering inaccuracies

Leakages on transmissionand/or distribution mains

Leakage overflows at utilitystorage tank

Leakage on serviceconnections up to point ofconsumer metering

Revenuewater

Nonrevenuewater

Table 1: Water Balance Calculation

5

Source: IWA, 2000.

Unbilled authorizedconsumption

Apparent lossesWater losses

Real losses

Componentsof NRWAddressing this question requires thatutility managers understand the leveland sources of NRW through asystematic review of the network, howit is being operated, and what thecommercial practices of the serviceprovider are (Box 4 and Table 2).

• A review of the network conditionand management helps determinephysical losses2 that result fromfaulty, corroded, and old watertransmission and distribution mains,overflowing service reservoirs,and utility storage tanks. Theyare a result of poor operationsor maintenance, lack ofactive leakage control, anddeteriorating or poor quality ofunderground assets.

• A review of commercial practicesdetermines the nonphysical orapparent losses. This involves,among others, an analysis ofcompleteness and accuracy ofcustomer database, billingprocedures and frequency, tariffstructure and subsidies, meteringpolicy, handling of illegalconnections, and response tocustomer queries.

• Unbilled authorized consumptioncomprises water for operationalpurposes, firefighting, and forservices to special consumergroups as may be authorized in

Box 4: Understanding the Components of Nonrevenue Water

Several levels of analysis would need to be undertaken for a nonrevenuewater review:

• Discussions with or among senior staff to review current managementpractices, financial and political constraints and influences, attitudetowards leak detection and reduction, customer databaserecordkeeping, and so on.

• Discussions with or among operational staff on how the systems areoperated, including how water leakages are detected and managed.

• Review of technical and commercial data to validate the issuesdiscussed with management and staff and, combined with the resultsof the Water Audit, to get a broad picture of the situation.

• Gather more data to complement the existing ones and implementa systematic collection of key data, if that is not a currentpractice already.

2 Reviewing network conditions and management helpsunderstand the levels of physical losses. This is done througha review of the network age, materials used, hours of supply,service pressure, valves operation, frequency of breakages,particular local characteristics influencing water loss (corrosivesoil, and so on), leakage repair, attitude and level oftechnology available for monitoring and detecting leakages,staff capabilities, skills and awareness levels, among others.

Utility managers will need to understand the levels and sources ofnonrevenue water through a systematic review of the network and how it isbeing operated.

6



Nonphysical(‘management’ or‘apparent’) losses

Physical (real)losses

Source::::: Adapted from Leakage Management and Control: A Best Practice Training Manual. WHO, 2001, and IWA’s best practice standard water balance.

specific tariff regimes. It alsoincludes the difference betweenactual and estimated consumptionwhere there is a volumetric tariffbut no functional meters or waterconsumed by legal customerswhere billing is based not inactual or estimated terms butaccording to certain physicalhousehold characteristics.

7

Table 2: Components and Causes of Physical and Nonphysical Water Losses

Losses Components Causes

Unbilled authorized Unbilled metered consumption Water delivered to special customers that are not billedconsumption although they are metered

Water used for operational purposes (flushing anddisinfection, if metered)

Unbilled unmetered consumption Water delivered to special customers that are neitherbilled nor metered

Difference between actual and estimated consumption

Water used for operational purposes (flushing anddisinfection, if not metered)

Water used for firefighting

Unauthorized consumption Illegal connections where there is no access

Illegal connections to properties that have legalconnections

Illegal connections of vendors selling water

Metering inaccuracies Under-registration of customer meters

Poor quality, inaccurate meters

Inadequate meter maintenance or replacement policy

Stopped meters

Data handling errors

Leakages on transmission and/or Burst pipes (sudden rupture of pipe)

Leaking joints and fittings

Leakage overflows at utility storage tank Seepage from old masonry or concrete walls

Float-valves not working

Leakage on service connections up Burst pipes (sudden rupture of pipe section or joint)

Leaking joints and fittingsto point of consumer metering

distribution mains

AddressingEach of the NRWComponentsOnce Theyare IdentifiedGiven that NRW is the differencebetween water supplied to the systemand water billed to the customers,the value of these two variables iskey for determining NRW levels.These could be estimated by installingmeters at the system entry pointand at each and every consumptionpoint as well.

Macro-meters can be installed atthe outlet of the water treatment plantand storage or distribution tanks and,normally, are also installed in everytakeoff in the main conveying system.3

Customer meters estimateconsumption at the consumer end.These are sometimes not installeddue to technical, economic, orpolitical reasons.4

Once input and consumption figuresare known or estimated (taking intoaccount its consequences), NRWcan be calculated and the searchfor the specific solutions canthen begin.5

Box 5: Technical Solutions for ControllingNonrevenue Water (Bangalore)

A study undertaken by Larsen & Toubro-Thames Water consortium on theuseful life of the Bangalore Water Supply and Sewerage Board’s pipesrevealed that the pipes were in good condition and would remain so for 30more years. The consortium, however, recommended in-situ replacementsand rehabilitation, slip-lining of pipeline to improve the flow and reducecorrosion in existing pipes which, if implemented, would increase pipe lifesignificantly two or three times more than their estimated remaining useful life.These recommendations were adopted for pipes with diameter up to3 inches that were heavily corroded and encrusted. About 150 km of pipelinehas been rehabilitated, with 700–800 km still pending.

The Bangalore Water Board also initiated a study in 2000 that indicated faultyand inaccurate meters and leakages in 30- to 40-year-old galvanized ironpipes feeding house connections as major causes for high nonrevenue water.Accordingly, the utility undertook replacement of all service pipes connectedto corroded galvanized iron pipes with medium density polyethylene pipes.The Board also made it mandatory for house service connections in future touse medium density polyethylene pipes only. The costs of rehabilitationworked to US$506 per house connection and were to be paid by consumersthrough 10 monthly installments of US$5 each. The Board has replaced over150 km of 100 mm galvanized iron pipes and 50,000 house serviceconnection pipework with average length of between 3 and 5 meters. Whilethe per meter cost of replacing the galvanized iron pipes worked to US$26,the per meter cost of replacement for the service pipes cost US$13.75. Thishas been undertaken by Larsen & Toubro and Shaw Technical Consultants,costing about US$3.75 million in all.

3 If not installed, the input should be estimated using othermethodologies like considering hours of pumping combinedwith pump curves, weir constants, Parshall channels, hydraulicmodels, among others. If the input to the system is estimated,the possibilities are that either it is overestimated orunderestimated.4 If customer meters are not installed, estimates are used.These can be derived using different methodologies: by meansof the unit consumption per person per day, more complexcalculations considering all the appliances existing in thehouseholds, conducting sample metering in pilot areas andextrapolating the results, and so on. Again, there are twopossibilities. If customers’ consumption is overestimated theNRW calculated will be lower than the actual and vice versa.The difference between actual and estimated consumption, ifpositive, contributes to the unbilled authorized consumption.5 Similar principles are used in the District Metering Areaapproach of leakage management, as discussed later.6 US$1 = INR 40 (as of October 2007).

Leakage management for tackling technical losses can be through: passivecontrol; regular leakage surveys; and leakage monitoring in zones through theDistrict Metered Area approach.

8



TechnicalSolutions forControlling NRWPhysical losses: Any NRW programmust address technical solutions fortackling physical losses. Such solutionswill involve network rehabilitation,refurbishment, and improvedmanagement; they also depend onhow such leakages are currentlydetected and managed,their frequency and location, theirassessment and management, howlong it takes to achieve and maintaintarget leakage levels,7 how expensivesuch an exercise is (capital andoperating costs), and whether it issustainable in the long term. The easewith which such solutions are foundand implemented will depend onwhether network drawings are in place,how old the network is, and on theexistence of continuous water supply inthe target area for facilitating detectionand repair of leaks.

Leakage management for tacklingtechnical losses is categorized intothree groups: passive control; regularleakage surveys; and leakagemonitoring in zones through the DistrictMetered Area approach.

Passive control is undertaken inresponse to visible leakages in thesystem that arise due to bursts or

drops in pressure, typically reported bycustomers or by linesmen. These aresometimes part of the day-to-daycontrols that are undertaken by thestaff of water utilities in response toregistered complaints.

Regular surveys are a method ofinspection undertaken for a distributionsystem where leakages are looked forwith listening devices8 on pipelines andfittings, or by reading and analyzingmetered inflows to temporarily isolatedareas as leakages may be indicated byhigh night flow rates. These areundertaken on a more regular basisand form part of most leakagemanagement strategies undertaken bywater utilities.

Leakage monitoring in zones is amore advanced, focused, andcomprehensive strategy that monitorsleakages by creating a hydraulic supplysystem divided into operational zonesdefined on the basis of servicereservoirs, pumping stations, pressurezones or other operationalconsiderations. These operationalzones are further divided into DistrictMetered Areas within which leakagelevels are constantly watchedand monitored.

The District Metered Area approach isa way to study an intricate problemaffecting a large (unmanageable) areaby dividing it into smaller (moremanageable) ones. In a sense, it means

breaking down a big network into morenumerous but smaller areas, the DistrictMetered Areas. Then, for each area thetotal net inflows and the volume ofwater billed are measured and theNRW is calculated as the differencebetween the two figures. Once this isdone, areas where NRW levels arehigher than desired are analyzed todetermine the appropriate solutionsand implement them according to aprioritized program.

District Metered Areas are discretehydraulic areas created within thesupply zones or service stations of thecity or town with a defined andpermanent boundary that typicallycover about 500 to 3,500 connections.The establishment of District MeteredAreas includes preliminary design,detailed site survey and data collectionand an understanding of the feedermains, networks valves and meters inthe area, and the supply pattern.Depending on its supply scheme, eachDistrict Metered Area is installed withone or more district meters, whichshould be read and recorded regularly(ideally it would either have a datalogger or it would be linked to a centralcontrol station for continuous datarecording). For each area, the serviceprovider can:

• Produce a water balance;

• Monitor night flow-rates;

• Perform ‘Step Tests’;

• Monitor pressure levels; and

• Verify consumption anomalies.

These actions allow the serviceprovider to define the relevance of eachNRW component (physical andapparent losses and unbilledauthorized consumption), locate leaks,

7 Target leakage levels are often based on the Bursts andBackground Estimates, which were developed from a reviewof international literature and consideration of the manyparameters that influence losses. The original Bursts andBackground Estimates leakage management concepts andsoftware were developed in 1992–94, during the UnitedKingdom’s National Leakage Initiative. Such techniques aredeveloped on four principal leakage management issues,namely logging and analysis of Minimum Night Flows, theeconomics of leakage and leakage control, pressuremanagement and benchmarking of leakage, and auditing ofnonrevenue water. More details on this concept are found inLeakage Management: Introduction to WRC Tools to ManageNonrevenue Water, by R. S. McKenzie and J. N. Bhagwan.

8 Some of them are (a) sounding sticks that are simple acousticor electronically amplified instruments used for sounding on allfittings, valves, and hydrants and confirming the position of aleak; (b) ground microphones that involves placing themicrophone on the ground at intervals along the line of thepipe and noting changes in sound amplification as themicrophone nears the leak position; (c) leak noise correlatoris the most sophisticated of the acoustic leak locationinstruments. It relies on the velocity of sound made by theleak as it travels along the pipe wall towards each of twomicrophones placed on conveniently spaced fittings. Thelatest versions of correlators have the capability for frequencyselection and filtering and some have a noise detection headthat is inserted in the pipe for increased accuracy.

9

illegal connections and changes incustomer categories and detect otherproblems like pressure increasesbeyond desirable levels, among others.It is assumed that, as part of the regularoperations of the service provider,customer meters are read on a monthlybasis for billing purposes (its accuracychecked according to a metering policy)and the data also used for the waterbalance for each District Metered Area.

When invisible leaks are suspected,monitoring night flow-rates allowsestimating leakage volume fromdistribution mains, service connectionsand fittings, after subtracting legitimatecustomer night use. Best practiceanalysis of District Metered Area flowsrequires the estimation of leakage whenthe flow into the District Metered Areaand customer demand is at itsminimum, occurring between12 midnight and 4 am. Analysis ofnight flow-rates determines whetherconsumption in any District MeteredArea is much higher than expected or ifit has increased over time, indicatingbursts or undetected leakage.Suspected areas can be inspectedmore thoroughly using the ‘step testing’procedure in which pressure in thenetwork is monitored while sequentiallyclosing valves until pressure risessignificantly, showing that the leak inthe last line is closed.

Detection of leakage in each DistrictMetered Area is also carried out usingequipment such as basic or electroniclistening sticks, ground microphonesand leak noise correlators. Waterpressure levels within a District MeteredArea are also continuously monitoredthrough pressure loggers deployed atdistrict meter points, elevated zonesand major branching points inthe network.

Box 6: Using District Metered Areas and Bulk Metering toControl for Nonrevenue Water (Jamshedpur)

The drive to reduce water loss levels comes from Jamshedpur Utilities andServices Company Limited’s realization that every m3 of water lost hasresulted in huge revenue losses for the company. Given total potable waterproduction of 180 million liters per day, a percent reduction in NonrevenueWater per day amounts to a saving of 1,800 m3 per day or 54,000 m3 permonth which, cost at US$0.20 per m3, amounts to a monthly saving ofUS$0.01 million. Accordingly the utility created a District Metered Area with24-hour supply in its Circuit House Area. The network spans over 1.83 km2, isover 75 years old and serves 350 households. The District Metered Area isconstantly monitored to see if pressure and leakage levels through 24-hourwater supply are maintained at optimal levels. It took the company about sixmonths to design and implement the District Metered Area.

The utility is also undertaking a systematic bulk metering program of allsystem inlet connections, large consumers, and industrial connections andoutlets. Bulk meters are periodically read and the data downloaded foranalysis and calculation of losses in the distribution system. The bulkmetering program was undertaken in two phases, costing an estimatedUS$1 million, where Phase I was undertaken over a period of 13 months andinvolved the installation of 41 bulk meters at the supply end and up to watertreatment including metering of all inlets and outlets at its pump houses andwater treatment plants as well as metering of branch networks and watertower inlets. Phase II involved installation of 89 meters over eight monthsincluding metering tower outlets and connections to associated companiesand industrial units.

Bulk metering and constant monitoring of its mains has enabled the utility inanalyzing leakages and in surveying pipelines for illegal and unknownconnections. This has helped the utility in identifying two extra lineconnections in the central network which, when metered, brought aboutreduced unaccounted-for water.

These initiatives have brought about impressive results in reduced leakages.While it is difficult to get appropriate data for NRW given the absence ofconsumer metering, unaccounted-for water in the rising mains and theindustrial water network has fallen; unaccounted-for water in the rising mainsfell from 23 percent in 2004 to 10 percent in 2006.

Thus, new neighborhoods have been now served with water that was savedthrough these initiatives. The utility is now working on implementingcustomer meters for all 40,000 household connections in order to reduceleakages in house connections as well. So far the majority of the domesticconsumers were paying a fixed monthly rate of US$3.5. The utility feels thatmetering of all domestic connections will minimize leakage and promoteefficient use of water, while allowing it to divert surplus water to other areas ofthe city.

10

When invisible leaks are suspected, monitoring night flow-rates allows estimatingleakage volume from distribution mains, service connections and fittings, aftersubtracting legitimate customer night use.



Leakages, once identified, must bebrought to controllable ‘target’levels, through a mix of pressuremanagement, and repairs orrehabilitation of distribution mains andhouse connections. These targetleakage levels are formulated onestimated losses from calculatingindividual components of losses inthe distribution systems and on thecustomers’ pipework, either as part ofan annual water balance or from a nightflow-rate analysis. Target leakage levelsvary on a case-by-case basis and aredetermined by the type of leakagecontrol policy, the size and number ofDistrict Metered Areas, the optimalpressure levels to be maintained, andthe staffing policy. Typically, leakagetargets are based on an ‘economic’level defined as a point of minimumtotal costs when costs of both leakingwater and leakage control are takeninto account.

For successful functioning and longterm sustainability of the DistrictMetered Area approach, leakage andNRW levels must be continuouslymonitored and staff awareness,capacity, and motivation kept at thehighest possible levels. All theinformation generated through thisapproach is optimal for feeding ahydraulic model of the network that, inturn, will provide feedback to the NRWmanagement program.

Unbilled authorized consumption isa result of certain legal, institutional,and other social considerations thatare not always sufficiently justifiedor compensated. As indicated inTable 2, there are mainly fourreasons for unbilled authorizedconsumption. They are discussed here,together with some ideas on how totackle them.

Box 7: Using a Pilot District Metered Area toControl Leakages (Bangalore)

The Bangalore Water Supply and Sewerage Board piloted a nonrevenuewater reduction and control program in May 2003. The project, funded bythe Japanese Bank for International Cooperation, spans across 32,000connections and 270 km of pipeline (of diameter 50–600 mm and age five to70 years) in central Bangalore.9 Cost at US$12 million, it was awarded to theLarsen & Toubro-Thames Water consortium. Thames Water providedtechnical inputs while Larsen & Toubro was responsible for ground-levelpipeline rehabilitation work. The project was implemented in two phases, withPhase I focusing on NRW reduction, establishment and maintenance of targetNRW levels in each District Metered Area and staff training, and Phase IIfocusing on maintaining pressure and NRW levels at target levels asdetermined in Phase I.10

Phase I involved the formation of 21 District Metered Areas within five servicestations in the pilot area.11 The District Metered Areas covered about 1,000 to3,500 consumer connections, with the exact number of consumers based onestablished boundaries and a network analysis. Once District Metered Areaswere created, levels of NRW were regularly monitored. Leakage detection wasundertaken through noise logging, leak noise correlators, and groundmicrophones, especially during the night hours. Repair work includingrehabilitation and replacement of pipes using medium density polyethylenepipes was undertaken. The project controlled for apparent losses through arobust check of meters for accuracy in terms of functionality, water tightnessand security and their subsequent replacement.12 Physical losses weremeasured using the net night flow method and were monitored on a monthlybasis. These calculations also took into consideration the impact of pressureas monitored in a few identified locations through data loggers. Ideal targetleakage level rates were also established based on international bestpractice data estimates of water service providers in the United Kingdom inthe 1980s.13

Phase II involved constant monitoring, leak detection surveys, repair of leaksand rehabilitation of pipelines, and retesting of leakage levels in the DistrictMetered Area till water loss levels were reduced and maintained at target levels.The pilot project has improved water service standards and quality. It isproposed that this pilot will be extended to all areas of the city. The JapaneseBank for International Cooperation has extended its loan component toUS$100 million. The project is designed to cover the balance 0.36 millionconsumer connections that were outside the pilot zones.

9 This covered areas of High Grounds, Coals Park, Johnson Market, Ulsoor, and Clive Lines Reservoir.10 Each phase was for 18 months, with a three or four month overrun for complete implementation of Phase I.11 Initially the plan was for 16 District Metered Areas, which was later increased to 21.12 The project tested 5 percent or a minimum of 100 working meters per District Metered Area for accuracy. Faulty consumermeters were identified through customer complaints, billing data and a physical survey of meters in each District Metered Area.Of about 31,808 consumer meters surveyed during the pilot, about 30 percent meters were of a specific make; 69 percent of thesemeters were in good condition and 31 percent of them were either not working or were not readable. All nonworking meters of thatmake were replaced. In July 2004, about 3,124 meters were replaced.13 These estimates were inferred through a graph iteration based on a typical situation of leakage management in theUnited Kingdom, deducing what should be the target leakage level for each District Metered Area given a baseline Net Night Flow.

11

Water delivered to special customersthat are not billed: Certain institutionscould be regarded as deserving specialtreatment and, therefore, it may havebeen decided that these institutionsshould not pay for their waterconsumption. However, without makingany judgment on the adequacy of thiskind of decision, it is desirable that theyare based on a predetermined andsound policy applied objectively. In anycase it is also desirable that the numberof cases should be kept to a minimum.It is advisable that these users aremetered and sent the bill that indicatestheir consumption and gives them anestimate of how many people couldhave been supplied with the water thatthey are using. This could lead them topay attention to any anomaly in theirconsumption that could indicatewastage and/or internal water lossesthat could be fixed. Another optioncould be to send crew to search forlosses inside the premises if the billingsystem detects an abnormality in theconsumption of any of these specialusers. Yet another means could be toprovide coupons to certain consumergroups such as governmentinstitutions, with the coupons allowingfor budgeted consumption.

Certain service providers opt fordelivering a relatively limited volume ofwater to poor people for free or servicethem through public standposts on theunderstanding that they cannot affordany other kind of service. In fact, notonly can poor people afford to pay forwater services but most of them wouldprefer to have a reliable service thatwould allow them to avoid buying waterfrom vendors to cover for shortcomingsin the public supply. As discussed inBox 8, the Bangalore Water Board,over a period of five years,

experimented with providing services topoor people and found that theywillingly paid to connect to a reliableservice. Similarly, other Indian cities ofBaroda, Kerala, and Dehradun as wellas Kathmandu in Nepal indicatedthrough various surveys thatconsumers were willing to pay more forreliable, safe, and adequate watersupply services.14 In many cases,paying for water service allows poorpeople to get other services as theyhave a proof of residency and solvency.For service providers, connectingpoorer sections of the populationimproves both revenue and service

quality, as often the public standposttaps are removed and the water flowsuninterruptedly during supply hours,thus lowering pressure and increasingwastage. Some service providers haveresolved this through innovative meanslike rationalized or subsidized waterconnection fees and subsidizedmonthly rates, simplified proceduresand application rules for a newconnection so that poor people areindeed encouraged to apply for it,providing jobs to community memberson network construction andadministration of applications for newconnections, and so on.

Many times, when the service provideris a public entity, governmentinstitutions are not billed nor do theypay for the water service as this isregarded as money that goes from one

Box 8: Connecting Poor People to the Network(Bangalore, India)

The Bangalore Water Supply and Sewerage Board has undertaken some

commendable work in connecting the city’s poor people to piped water services.

The Board found innovative ways to connect them through subsidized connection

fees, options for group connections, and simplified and easy methods for applying

for a new connection. The program was jointly partnered with AusAID during

2000–02 wherein the project’s Community Development Component examined

and tested options for improved service delivery to urban poor people in three pilot

slums: Cement Huts, Sudhamanagar, and Chandranagar.

A total of 850 connections (individual and shared) were installed during the pilot

phase. After successful implementation of these pilots, the Bangalore Water Board

decided to replicate the results in other slums. The Board’s Social Development

Unit undertook extension of water supply and sewerage services through a

‘Package Program’ in coordination with nongovernmental organizations and

community-based organizations, whereby slum dwellers were allowed to legally

connect to services, provided they paid for the connection. Today there are about

6,000 connections across 46 slums in the city. Households in poor settlements are

willingly paying regular water charges. By connecting all slums to piped water, the

Board is reducing nonrevenue water and hopes to slowly phase out the over

15,000 public taps within city limits.

14 (a) WSP-SA. June 1999. Willing to Pay but Unwilling toCharge—Do Willingness-to-Pay Studies Make a Difference?;(b) Chetan. 1995. Study on Willingness to Pay for Water andSanitation Services—Case Study of Baroda; (c) Whittington,Dale, S.K. Pattanayak et al. 2001. Willingness to Pay forImproved Water Supply in Kathmandu Valley, Nepal,Research Triangle Institute, Final Report.

12

Not only can poor people afford to pay for water services but most of them wouldprefer to have a reliable service so that they can avoid buying water from vendorsto cover for shortcomings in the public supply.



pocket to another. However, eventhough there may be no physicaltransfer of the money, the paymentshould be accounted for to ensure thatthe revenue and the cash are availableto the service provider. Of course,this requires that the accountsof the service provider are clearlyseparated from those of otherpublic entities.

Difference between actual andestimated consumption: There aretwo ways to reduce this element ofnonrevenue water: improving thequality of the estimates (which could bevery difficult), and installing and readingcustomer meters, recording thereadings in a database and calculatingthe difference between readings andactual consumption. This is importantto capture, especially in cases wherebilling is not based on meter-basedcharging practices and where a flat ratehas been imposed. In such cases itwould be good to be able to estimateconsumption and see the differencebetween the two since the impositionof a flat rate also encourages illpractices of no water conservation.This estimate may also be able toindicate that the meters are not inworking condition.

Water used for operational purposes:Unbilled authorized consumption alsoresults from water being used foroperational purposes such as flushingand disinfection functions. In suchcases it would be important to at leastkeep track of the volume of water usedso that it is kept within reasonable limitsand the utility can inculcate somehabits of controlled use. More efficientflushing and disinfection operationscan help reduce this component.Measuring the volume of water used forthese kinds of operations can help if

combined with incentives to the crewsthat can perform these operationssuccessfully with lower volumesof water.

Water used for firefighting: The waysto reduce this component are beyondthe scope of this series. However, theservice provider could help authoritiesfind ways to reduce the number of firesor to extinguish them more rapidlyand effectively.

Apparent losses includeunauthorized consumption (basically,illegal connections) and meteringinaccuracies (under-registration,stopped or defective meters, and datahandling errors).

Illegal connections: Can only bereduced if water providers have inplace policies to convert them into legalconnections. An efficient tariff regimewith appropriately targeted subsidies isvital to address special cases ofhouseholds lacking the resources topay for the service. As a last resort, aservice provider should have crediblemeasures for disconnecting those whodeliberately do not want to regularizetheir situation despite the opportunitiesgiven to them. Illegal connections canbe tackled through a careful analysis ofthe customer database and comparingit to municipal registers or, if existing,combining geographic informationsystem data with that used by themunicipality or other service providers.In Bangalore, the Water Board hasa disconnection policy wherebyconnections are initially clamped offat the street level, after which legalnotices are sent. In the event that stillno action is taken, the Board resortsto disconnection.

In many cases, poor settlements resortto illegal connections because theservice provider refuses to connectthem due to many reasons (lack of landownership titles, uncertainty on whetherthey will become regular customersand pay their bills, and so on.) The bestsolution for this problem would be tofind ways to connect poor people andget them to pay their bills. Again, asdemonstrated in the case of Bangalore(Box 8), the water utility mobilizedslums and successfully helped themconnect to the network. Thesecustomers continue to be served withwater, receive bills, and makepayments. Earlier, the Water Board’spolicy was that a connection couldnot be approved unless the residentprovided a proof of residence. Ruleswere subsequently eased so that the

13

requirement for formal tenure proofwas replaced with a simple proof ofresidence such as a ration card,election identity card or even anidentity card.

Sometimes, existing legal customersusing large volumes of water installillegal, or bypass, connections toreduce their water bill. Analyzing thecustomer database, billing history, andvisiting the customers are ways totackle this problem. Meter readers canplay a very important role detecting thiskind of situation. In Bangalore, meterreaders and utility staff played animportant role in improving commercialefficiencies. The Water Boardincentivized utility staff and meterreaders through healthy competition toundertake better billing practices, notonly for checking illegal or bypassconnections, but also of connectionswith meter inaccuracies (as discussedlater). Monthly billing targets were setand reviewed, after which the bestperforming zone for revenue targetswas rewarded.

Metering inaccuracies can be dividedinto two main categories:

Meter problems: These could berelated to procuring meters ofinadequate class or poor quality,premature wear or corrosion due toaggressive water, low flows andcustomers tampering with the meters,among other reasons. Careful attentionon selecting the meters to procure isthe best way to avoid the first threecauses mentioned, while dedicatedcustomer service and educationpolicies could help to deal with thefourth one. Ultimately, a sanctionspolicy could be implemented to dealwith the most difficult cases. As Box 9indicates, Bangalore resorted to using

only high quality meters that were morereliable, accurate, and durable.

Data handling problems: For providerswho do bill according to actualconsumption, taking readings,recording and then handling data bymeter readers is another potentialsource of nonrevenue water.

There is a wide range of options forrecording readings and producing the

bills, from manual recording in a tablecontaining only the account numberand address, to remote centralizedautomatic reading. Deciding on oneoption for handling the data will dependon the particular circumstance for eachservice provider.

However, as a practical consideration,it is advisable to avoid copying the datafrom one table to another as muchas possible.

14

Water utilities will need to incentivize their staff to undertake better billingpractices, not only checking for illegal or bypass connections but also forconnections with meter inaccuracies.



Recording readings through a handhelddata entry device could be especiallyhelpful as preloading previous readingscould warn the meter-reader of anydata entry mistake beyond a certainthreshold so that he can correct itimmediately.

Some utilities in India, in Hyderabad forinstance, have been using handhelddata entry devices for recording meterreadings. The Board has witnessedimproved billing efficiencies as a resultof these initiatives. Initially the Boardoutsourced this function to aspecialized firm that had relevantexperience in this function. It issimultaneously also training its ownmeter readers so that they can takeover this function in due course.15

If these devices cannot be providedto the meter readers and they have torecord the readings manually, thenincluding the previous reading and anindication of the expected new readingcould be helpful. However, the lastoption would also induce the meterreader to record an ‘educated guess’of the new readings instead of theactual ones.

In summary, inadequate/nomeasurement facilities, inadequatecalibration of bulk meters,under-registration, poor quality andinaccuracy of customer meters, faultyand stopped meters, inadequate metermaintenance and replacement policy,inadequate meter reading practicesand under-estimation of water for freesupplies and operational use areproblems that can be corrected only ifthe water service provider has a robustmetering policy in place for its own

Box 9: Metering Practices (Bangalore)

Bangalore is one of the few cities in India that has all its 0.4-millionconnections metered, with the metering program being in practice since1966. Previously all household connections were installed with ‘Class A’meters that had an approximate life of three to five years. Costing US$11.2(as of 2004), the meter dials were prone to moisture and air and wereless accurate.

Under the nonrevenue water project 15,000 meters were replaced with‘Class B’ meters that are more accurate and durable. Costing US$30, thesemeters are multi-jet meters, tamper-proof, and robust in nature, having anapproximate life span of seven to 10 years. They are air resistant—theymeasure air to a lesser extent, since the magnetic drive inside the meter getsdisengaged once air passes through the meter. These meters also capturevery low water flows in case of low pressure supply.

The utility has also been actively encouraging regular testing and repair ofconsumer meters. Initially it was itself undertaking all meter testing and repairwork, repairing and testing approximately 1,500 meters on average in amonth. In October 2003, it outsourced meter repairs work to a privateagency. It charges its consumers a meter maintenance fee for undertakingthese repairs.

facilities and all the water connectionsthat it serves. This will mean checkingbulk meters, replacing customer metersat the end of their useful life and alsoactively checking them through regulartests and repairs. Monthly meterreading and billing, together

15 More details on this initiative are found in Series 2, on billingand collection practices.

15

with other measures, may helpmonitor and prevent customers fromtampering with their meters. Aproactive metering policy is alsoa prerequisite for the effectiveimplementation of the District MeteredArea approach.

Addressing theInstitutional,Organizational,and RegulatoryArrangementsfor Creatinga SuitableEnvironment forControlling NRWOther policies for keeping NRW at targetlevels relate to the internal organizationalstructure, regulatory framework fortriggering NRW control, staff capacitiesand training, as well as customer careand information campaigns. These areimportant because they ensure thatNRW management programs become asustainable practice.

NRW reduction will only be successfulif there is an adequate enablingenvironment and a proper incentivestructure for performanceimprovements. The current inherentinstitutional deficiencies in the Indiancontext would need to be resolvedthrough appropriate institutionalarrangements that incentivize serviceproviders to be more accountable and,at the same time, clarify roles andresponsibilities of all stakeholders in thewater supply and sanitation sector sothat the appropriate stakeholder can beheld accountable for performing itsspecific functions properly.16

Any NRW management programshould hence ensure that the incentivesare duly aligned with the objectives of

developing an efficient and reliableservice provider that meets the needsof all consumers. As demonstrated inBox 10, institutional arrangements inthe Vietnamese city of Haiphong weredesigned such that NRW managementwas targeted at the ward level. Thecreation of strong local units deliveringdecentralized services at the wardlevel demonstrated an innovative wayto reorganize the institutionalarrangements for increasedaccountability. Similarly, appropriateemployee incentive schemes were setup in both Bangalore and Haiphong toreward staff for improved performancethrough bonuses, rewards, certificates,and recognition. Such initiativesincentivized staff sufficiently such that itresulted in improved services.

Internal Organizational Structure:Regarding internal organizationalstructure, traditionally, serviceproviders have—besides the regularadministrative, finance, and humanresources departments—technical andcommercial departments.

The technical department is usuallyresponsible for water production,pumping, storage and distribution,network operation and repair, and mayinclude the engineering departmentthat would deal with planning,designing and constructionof new assets. The commercialdepartment would produce anddistribute bills, collect payments, andreceive customer complaints. However,this kind of organizational structuredoes not always lead to a strongawareness of NRW levels and makeseffective NRW reduction andmanagement more difficult.

An alternative way of internalorganization takes into consideration

the different functions or processes thattake place in what one could call thevertical flow within the service provider.In this sense one could identifyproduction, transmission (involvingmain pumping stations, reservoirs andtrunk mains) and retail distribution andselling to the end user.

This organizational structure resemblesa clear supplier-customer chain whereeach department produces or receiveswater and has to deliver it to itscustomer. The number of productionand transmission departments isdetermined by the number of facilitiesand the way the service provider wantsto bundle them, while the number ofretail departments would be ideallydetermined by the number of DistrictMetered Areas. In this type oforganization it is very easy to focus onNRW levels and management as themanager of each department isresponsible for its own NRW,considering that the production andtransmission departments are sellingthe water they receive to their customerdownstream the vertical flow within theservice provider.

The Haiphong case demonstrateshow institutional arrangements couldbe rearranged such that each wardwas made responsible for its ownwater supply functions.

Technical and administrativeunbundling proved to be an innovativeway for ensuring that service deliveryimprovements were indeed recorded atthe individual ward level by makingevery local staff more accountable forevery unit of water produced.Employees felt increased jobresponsibility because they were boundby trust to the community and, at thesame time, monitored by them.

16 The issue of improving upon the current institutionaldeficiencies is beyond the scope of this paper and has beenaddressed in the Performance Improvement note, Series I.

16

Reducing nonrevenue water is achievable but it remains challengingbecause the approach to reducing and managing NRW goes beyondtechnical solutions.



Box 10: Decentralizing Service Delivery (Haiphong, Vietnam)

The nonrevenue water (NRW) program in Haiphongtargeted the reduction of water losses throughrehabilitation and reconstruction work of the waternetwork at the ‘phuong’ or ward level. 17 The aim wasto redistribute recovered ‘lost water’ to areas of poorsupply through a two-phase program. The first phasewas a pilot in one phuong—Lam Son ward—thatwould demonstrate to consumers how improvementsin water supply services were going to beimplemented subsequently in the remaining37 phuongs. Lessons from the pilot would then beexpanded during Phase II to remaining phuongsstarting in 1994, with all 38 phuongs achieving NRWimprovements by 2000. The model would workthrough unbundling of the water supply network forboth technical and administrative aspects (refer Box atright) such that each ward unit could perform its ownwater supply services. It cost approximately $1 millionper ward to fully improve the ward’s water supplydelivery. While ward-level infrastructure improvementswere paid for by the Haiphong Water Supply Companythrough various means,18 the customer had to pay forthe service connection and for the meter, total cost forwhich ranged between $30 and $60 per household,depending on the length of the service pipe.

Ward-level work began by improving all distribution,branch, and household pipeworks in the ward,complete metering of the wards’ consumers, and implementing a reliable monthly billing and collection system. The company alsoput in place a disconnection policy for disconnecting consumers if bill payments were delayed by more than 20 days. An officewith employees selected from ward residents was also set up and made responsible for all operation and maintenance of thephuong network. This helped provide constant and immediate communication between the utility and consumer, facilitating quickresponses to consumer needs and tailoring services according to consumer expectations. At the same time, the employees wereprivy to information on residents, being familiar with characteristics of the households in their ward.

These local staff were incentivized to not cheat the system19 through an employee incentive system where ward level staff salarieswere tied to a variety of yearly performance targets as well as a system of rewards (certificates, awards, and financial bonuses) andpunishment (written and verbal warnings and salary deductions). For example, if NRW achieved was 1 percent lower than the setperformance rate then a bonus of VND 20 million (approximately US$1,200)20 was awarded, 40 percent of which was given to theemployee directly responsible and 60 percent for the remaining functions of the utility. Employee performance was evaluated at thedepartment level on a quarterly basis and sent to management for a review. A contract was also signed between the HaiphongWater Company and the ward for defining ward level obligations (consumer level and overall Haiphong Water Company).Consumers were also made aware of the ward’s responsibilities through a billboard, displayed outside the ward office, whichdetailed both consumer and ward office obligations and responsibilities. Monthly ward People’s Committee meetings were alsoheld for reporting and addressing community water supply service delivery needs as well as to provide feedback on utilityperformance given the implementation of the ward model.

Technical and Administrative Unbundling

Technical Unbundling involved enforcing good technical standardsat the ward level. The utility’s ward offices connected all customerswithin their ward to distribution lines through branch connections,installing a meter on every connection, and signing a customercontract that detailed the company’s and consumers’ rights andresponsibilities. Metering customers all at the same time helpedprevent illegal connections. In addition, the practice of installing amaster meter in each ward helped in reconciling ward billing with theamount of water the ward received from the water treatment plant.This helped monitor the points for distributional losses and itsestimate and hence water bill collection efficiency. Improved serviceswere not going to result only from construction of moreinfrastructure or replacement of water mains, but from ward-by-ward diffusion of distribution network improvements and fromcareful network management.

Administrative Unbundling involved the creation of ward level localHaiphong Water Supply Company consumer service sub-offices,staffed with five or seven members of the community, including amanager, two water meter readers, two bill collectors, and atechnician for minor repairs of pipes and meters and for overalloperation and maintenance of the phuong. These offices undertookmeter reading, maintenance and revenue collection once thenetwork was rehabilitated and established.

17 The ward or phuong is the smallest local unit of socialist government in Haiphong, with 1,500 to 3,000 households (10,000 to 16,000 people).18 In Haiphong this cost was covered not only by company revenues, but also through a Finish International Development Agency grant and a credit from the World Bank for the Socialist Republic ofVietnam’s Water Supply Project.19 By drawing local staff into ward offices, the risk existed that there could be collusion between the ward employees and the ward consumers. This was countered sufficiently.20 US$1 = VND 16,216 (as of April 19, 2008). Conversion rates are from www.xe.com; all conversions in the text are approximations.

17

Incentives for improved accountability:Within an improved institutional andoperating structure, water serviceproviders also need to be encouragedto demonstrate improved accountabilityof functions and responsibilities.

In some international good practices aswell, as detailed in Box 11, serviceproviders have used performancecontracts for incentivizing staff tobetter services.

This is also demonstrated in both theBangalore and Haiphong cases whereemployee incentive schemes rewardedstaff for improved performance throughbonuses, rewards, certificates, andrecognition. Employees were alsogiven disincentives to under-perform,since they were penalized whenperformance was not up to the mark.Both cases also set up clearly definedperformance standards that had tobe achieved by employees if theywanted to be rewarded forperformance improvements.

In Haiphong, a particularly transparentincentive scheme encouragedperformance of the line managers atthe ward level, where staff andcompany shared benefits of anyefficiency gains achieved.

Regulatory framework: The regulatoryframework and, especially, the tariffregime in place are also key issues,particularly when it comes to fulfillingthe regulatory objectives of setting,monitoring, and enforcing principles ofefficient tariffs and improved servicequality and standards.

Enforcing NRW programs thateventually meet these objectives wouldneed to put in place regulatorymechanisms, institutionalresponsibilities, and accountability

frameworks involving suitable incentivesand sanctions, and also clearlydefine the assignment of roles andresponsibilities at various levels ofgovernment. Such arrangementswould also need to deal with wastageand proper assessment of investmentprojects by providing referenceregarding the proportion of subsidy(if any) included in the tariff and allowa correct allocation of costs andbenefits from eventual projects tobe implemented.

For instance, regulatory frameworkscan be designed such that the impactof an NRW program can be monitored.The service provider’s NRWperformance can be reviewed througha memorandum of understanding thatclearly establishes the various servicedelivery related regulatory objectivesto be achieved over a specifiedtimeframe, with related incentives ordisincentives for their achievement or

non-achievement. The actualperformance of the service provideragainst these targets could beevaluated through third partyassessments, financial audits, andbeneficiary analysis through consumerfeedback to ensure transparency. Insome cases the provider could also bemade to comply with disclosurerequirements that would enhanceaccountability in service delivery.

Staff training and capacity building:Equally important for sustainability ofNRW measures is training of staff innew skills and techniques. Any effectiveNRW management program wouldrequire a range of skills at all levels,starting from managers andprofessional engineers to techniciansand plumbers.

Staff members need to be motivatedand skills in techniques and technologyneed to be transferred to them for

Box 11: Incentive-Based Performance Contracts

In some countries water service providers have used performance contractsfor incentivizing staff to undertake sustainable service delivery improvements.For instance, in Senegal, the contract included specific and time-boundtargets covering areas of operational efficiency including reduced waterlosses and improved billing and collection. The contract effectivelyincentivized the private operator to achieve these set targets, by rewardingthe private operator not for the water that it produced but for water sold andpaid for.

Another example can be found in Uganda’s National Water and SewerageCorporation, which signed performance contracts with its respectiveoperation areas, using an internal incentive contracting mechanism withfinancial incentives for encouraging managerial efficiencies. A detailed set ofperformance targets for each area were devised and included, amongothers, improvements in billing and collection and reduced water losses.The performance contracts incentivized staff to meet these performancetargets through bonuses over and above their regular salaries.A set of penalties were also imposed in case of poor performance.

18

Any nonrevenue water management program should ensure that incentives arealigned with the objectives of developing an efficient and reliable service providerthat meets the needs of all consumers.



effective leakage management andsystem operation and maintenance.

The NRW management plan wouldneed to be accepted along with anunderstanding of its challenges andadvantages through improvedawareness and motivation of staff toadopt such strategies.

Staff members need to be trainedadequately so that they have hands-onexperience with such strategies andcan sustain the efforts of keepingNRW at acceptable levels in thelong run.

This was kept in mind while designingthe NRW reduction programs inBangalore and Haiphong. In Bangalore,particularly, Phase II of the projectfocused in part on training utility staff forspecializing in NRW management suchthat they could control the system oncethe pilot project was over.

Asset management: Although it is not aspecific action related to NRWreduction programs, an adequate assetmanagement plan would also need tobe maintained by staff as it is a veryuseful tool to help maintain NRW atthe desired level by triggering all therehabilitation and renewal actionsrequired to keep the assets in normalworking condition.

Such a plan would also help in makingsure that network rehabilitation is beingundertaken in an economical manner.This is important since many times thepiped network may be very old anddamaged and may also not bedesigned to take into accountmanagement aspects and revenuecollection practices for the provider,especially in the absence of meteringpractices, missing network maps,and so on.

19

MonitoringTargetNRW LevelsThe first step of any follow-up strategyis to implement a periodic andsystematic monitoring procedure. Thiscomprises reading of all bulk, district,and customer meters at least once amonth and calculating the NRW foreach zone in which the system isdivided (ideally District Metered Areas).In all three case studies explored inthis note, regular monitoring wasundertaken to ensure that performancestandards and targets were being metwithin the specified timeframe. This is

20

Component

Nonrevenuewater

Real losses

Real losses

Apparent losses

Water losses

Type

Financial

Waterresources

Systemoperational

Operational

Basic PerformanceIndicator

Volume of NRW as % ofsystem input volume

Volume of real losses as %of system input volume

Liters/service connection/day[For systems with 20 or moreservices/km of mains]

Use m3/km mains/day[For systems with fewer than20 services/km of mains]

Detailed PerformanceIndicator

Value of NRW as % ofcost of running system

The InfrastructureLeakage Index = ratioof Current Annual RealLosses/UnavoidableAnnual Real Losses(see Box 12 for details)

m3/service connection/year

Source: Manual of Best Practice (PIs for Water Supply Services), IWA.

Table 3: Performance Indicators for Water Loss Management

A periodic and systematic monitoring procedure must be in place toensure that performance standards and targets are being met within thespecified timeframes.

Operational m3/service connection/year



Box 12: The Infrastructure Leakage Index

The Infrastructure Leakage Index21

measures how effectively a utility orwater service provider is managingits distribution network forcontrolling real losses at currentoperating pressure levels. However,this does not imply that pressuremanagement is optimal, and it maybe possible to reduce the volume ofreal losses (but not the InfrastructureLeakage Index) by improved activepressure management. It is ameasure of how well repairs, pipelines and asset management and activeleakage control are controlled, as indicated in the diagram alongside.

Infrastructure Leakage Index is defined as the ratio of Current Annual volume ofReal Losses (CARL) to Unavoidable Annual Real Losses (UARL). While this is anindicator with no units and thus facilitates comparisons between providers usingdifferent measurement units, it requires attention when calculating UnavoidableAnnual Real Losses, computed as:

UARL (liters/day) = (18 x Lm + 0.8 x Nc + 25 x Lp) x P

where Lm = mains length (km); Nc = number of service connections;Lp = total length of private pipe, property boundary to customer meter (km); P= average pressure (m).

The length of mains and number of service connections are usually known,but the distance between the property line and the meter may not beavailable. Customer meters are usually located close to the property line andhence ‘Lp’ is 0; for the remaining cases, the average and total length ofunderground pipe from the property line to customer meters could beestimated by inspecting a small random sample of connections.

specially demonstrated in the Haiphongcase where monthly ward levelmeetings were held for reporting andaddressing community water supplyservice delivery needs as well as forproviding feedback on the ward’sperformance as per the specific servicestandards that were set. Providersneed to find appropriate NRWperformance indicators (see Table 3and Box 12) that could be monitoredand evaluated on a regular basis.Suitable indicators need to be drawnup for monitoring all components ofwater loss. The use of variousperformance indicators as opposed tousing only one percentage indicator forNRW allows water providers to properlybaseline, track and target, variouscomponents of water loss includingfinancial, operational and waterresources aspects.

The second step would be to prepare atable for comparing the different zonesor District Metered Areas andprioritizing future actions. Many serviceproviders that have been successful inreducing NRW have a team lookingafter service quality, NRW, billing, andcollection for each District MeteredArea. Usually, there would be a monthlymeeting where management and teamleaders would discuss the success andfailure stories, derive lessons and bestpractices, and award prices forsuccessful teams.

These practices create a soundcompetitive environment and lead allteams to come up with their own ideasabout what to do in their own DistrictMetered Area to improve NRW levels,how much supply they need and whattheir target NRW could be. Theproposed actions would be assessedusing a cost benefit approach,prioritized and presented as the plan for

that District Metered Area. Onceapproved by management, it will beimplemented and, later on, it will beassessed regarding its actual cost andeffectiveness. In broader terms, thepolicies and processes of producingand delivering water should bereviewed and rethought about, in casethere is scope for improvement. Some

policies to be reviewed periodicallyrelate to network management, pumps,reservoirs and valve operations,construction materials andspecifications, rehabilitation andmaintenance practices and connectionof new customers, especially those thatmay have difficulties paying the full costof the new connection.

21 This index is one of the most recent performance indicators for NRW and was designed at the IWA LeakageConference in Cyprus, Liemberger, 2002.

21

Source: Best Management Practice 3: ‘System Water Audits and Leak Detection’, Review andRecommendations for Change, California Urban Water Conservation Council, Thornton International,April 2005.

UARLCARL

PressureManagement

Speed andQuality of

Repair

ActiveLeakageControl

PipelinesAssets

Management

ConclusionFor developing an efficient strategy forNRW management, it is hence key togain a better understanding of thereasons for NRW (as analyzed througha water balance) and the factors thatinfluence its various components. Thentechniques and procedures can bedeveloped to suit specific localcharacteristics of the network, theconsumers and the provider, toaddress and control each of the losscomponents. As discussed through thecourse of the field note, planning andimplementing an NRW reductionstrategy requires a diagnostic approachto understand and estimate how muchis being lost, where and why is ithappening, how to reduce losses andimprove performance and how tosustain such a strategy (as detailedin Table 4).

All three case studies demonstrate thatsuccessful leakage reduction is not a

result of sporadic and irregulartechnical exercises to reduce physicallosses. Such programs should be acontinuous effort and must controlcommercial losses since the latterhelps improve the revenue stream ofthe water company almost immediately.In the longer term, the effectivenessof NRW programs is a functionof managerial efficiency andinstitutional accountability.

Technical solutions are generally noteasy to implement, since the networkmay be very old and damaged, withmany illegal connections and high NRWlevels. The network may also not bedesigned to take into accountmanagement aspects and revenuecollection practices of the providersuch as an absence of metering (bulkor consumption), missing networkmaps, and so on. Hence, whiletechnical solutions are essential andmay be time consuming and not easilyresult in quick successes, their long

22

Key Questions Techniques

How much is being lost? Water audit: Measure components, check production or consumption, recalculateMeasure components water balance, review records, operating procedures, skills

Where is it happening? Pilot studies: Quantify total losses, that is, how much the leakage is in the distributionQuantify leakage network, transmission mains and reservoirs and how much is non-physicalQuantify apparent losses loss; refine the water balance calculation

Why is water being lost? Review of network operating practices: Investigate reasons for loss, that is,old/corroded pipes and fittings, poor practices, poor quality assurance, localinfluences as well as cultural, financial, social, and political factors

How can performance be improved? Development of an action plan: Update records systems, introduce zoning/DistrictDesign a strategy and action plan Metered Areas; monitor water leakage; prioritize areas; address nonphysical

losses; detect and locate leaks; initiate repair/rehabilitation policy, hydraulic modeling,develop geographic information systems

How can the strategy be maintained? Training/awareness: Improve awareness and increase motivation; train staff withhands-on experience; monitor the loss management program; implement assetmanagement systems

Table 4: Designing a Nonrevenue Water Management Strategy

Source::::: Adapted from Leakage Management and Control: A Best Practice Training Manual, 2001. WHO.

Any successful nonrevenue water management program will need to addresstechnical issues within the overall institutional, organizational, and managerialenvironment, if it is to remain sustainable in the long run.



23

ReferencesBangalore Water Supply and SewerageBoard and Jamshedpur Utilities andServices Company Limited. Variousinternal reports and documents.

Coffee, Joyce Elena. April 21, 1999.Innovations in Municipal ServiceDelivery: The Case of Vietnam’sHaiphong Water Supply Company.Department of Urban Studies andPlanning, Massachusetts Instituteof Technology.

Farley, M., and R. Liemberger.Developing a Nonrevenue WaterReduction Strategy; Part 1:Investigating and AssessingWater Losses.

Farley, M., and R. Liemberger.Developing a Nonrevenue WaterReduction Strategy; Part 2: Planningand Implementing the Strategy.

Haiphong Water Supply Company.Modes of Engagement with PublicSector in WSS in Developing Countries:Case Study. Haiphong, Vietnam.

Kingdom, Bill, Roland Liemberger, andPhilippe Marin. December 2006. TheChallenge of Reducing Non-RevenueWater (NRW) in Developing CountriesHow the Private Sector Can Help:A Look at Performance-Based ServiceContracting. Water Supply AndSanitation Sector Board DiscussionPaper Series, Paper No. 8.

Luy, D. X. April 2000. Working withCommunities to Improve Water SupplyServices in Haiphong, Vietnam.Presentation at the WB-Urban Forum.

Thu, Tran Minh Anh, and Dam XuanLuy. 1998. Delivering Water Services atthe Local Level in Vietnam. 24th WEDCConference, Sanitation and Water forAll, Islamabad, Pakistan.

term sustainability depends on effectiveinstitutions and finances. Technicalsolutions, thus, ultimately have to becommercially feasible and should beplanned for, such as proactive androbust metering plans, checking forunauthorized connections, unbilledquantities, and volumes of free water.

Any NRW management program willtherefore need to address technicalissues and become part of the culturewithin the overall institutional,organizational, and managerialenvironment if it is to remainsustainable in the long run. It will needto also build relevant capacity so thatlocal staff can keep the program goingand, at the same time, develop aneffective monitoring system that

enables tracking whether target NRWlevels are being maintained or not.Selecting the most appropriateapproach for NRW management willdepend on the level of leakage,the cost of leakage, and thecost-effectiveness of each method.

Reducing NRW is achievable but itremains challenging because theapproach to reducing and managingNRW goes beyond technical solutions.Although it is not feasible to eliminate allNRW in a water utility, reducing it toefficient levels will be required to bringabout some quick performanceimprovements in terms of bothincreased cash flows and improvedservices, that is, more water availableto serve consumers.

Water and Sanitation Program-South AsiaWorld Bank55 Lodi EstateNew Delhi 110 003India

Phone: (91-11) 24690488, 24690489Fax: (91-11) 24628250E-mail: [email protected] site: www.wsp.org

April 2008

WSP MISSION:To help the poor gain sustained accessto water and sanitation services.

WSP FUNDING PARTNERS:The Governments of Australia, Austria,Belgium, Canada, Denmark, France, Ireland,Luxembourg, the Netherlands, Norway,Sweden, Switzerland, the United Kingdom, theUnited States of America; the United NationsDevelopment Programme, The World Bank,and the Bill and Melinda Gates Foundation.

AusAID provides WSP-SAprogrammatic support.

ACKNOWLEDGMENTS:This field note was peer reviewed byMario Alejandro Suardi, Cyril Wear,Chris Heymans, Mathew Kurian,

and Vandana Mehra

TASK MANAGER AND AUTHOR:Pronita Chakrabarti Agrawal

BACKGROUND RESEARCH:Pronita Chakrabarti Agrawal

(research carried out in 2005 and 2006)

Editor: Anjali Sen GuptaPictures by: Sajid DarokhanCreated by: Write MediaPrinted at: Thomson Press (India) Ltd

The findings, interpretations, and conclusions expressed are entirely those of the author andshould not be attributed in any manner to The World Bank, to its affiliated organizations, or tomembers of its Board of Executive Directors or the companies they represent.

Trung, Do Quang, R. Snow, L. Doukas,N. Thanh, and N. Trung. 1998. Water-Loss Reduction Program in Vietnam.24th WEDC Conference, Sanitation andWater for All, Islamabad, Pakistan.

Water Loss Group: IWA Task Force.Best Practice Performance Indicatorsfor Nonrevenue Water andWater Loss Components: APractical Approach.

World Bank. June 2002. AnalyticalAdvisory Activities, Benchmarking theUrban Water Sector. Vietnam, Hanoi.

World Bank. Report No. 16299-VN.May 20, 1997. Staff Appraisal ReportSocialist Republic of Vietnam, WaterSupply Project. Infrastructure OperationsDivision Country Department I, East Asiaand Pacific Region.

World Bank. Report No: 29128-VN.November 22, 2004. Project AppraisalDocument on a Proposed Credit in theAmount of SDR 75.20 Million(US$112.64 Million Equivalent) to theSocialist Republic of Vietnam for theUrban Water Supply DevelopmentProject. Urban Sector DevelopmentUnit, East Asia and Pacific Region.

ABOUT THE SERIES

WSP Field Notes describe andanalyze projects and activities inwater and sanitation that providelessons for sector leaders,administrators, and individualstackling the water and sanitationchallenges in urban and rural areas.The criteria for selection of storiesincluded in this series are large-scaleimpact, demonstrable sustainability,good cost recovery, replicableconditions, and leadership.

performance improvement planning designing an effective ......1 tata steel (formerly the tata iron...

Documents