:COPY _

:11397 Type: LLINGNESS TO PAY FOR WATER I TAF, MIR ANJUM Dept.:ANJUM J byShYMir AnjiUmAltaf, ;HarOOnJamal, -md Dale Whittington. ,

U~~~~~~~~~~Ž

,-- ~N X -~~,

(0

ca

ID coQ -DU.

| g os~~~~ 0 1992 The Interntonal Bank for Recostrwucdonand Developmenhe World Bank 1818 H Street, NW Washington,DC 20433 USA

Cover photo: Mir Anjum Altaf

Manufactwredin the United ates of Amedica

MmeUNDP-Word Bank Water and SanitationPogrm was oaized as a joint endeavor of te United Nations DevelopmentP_gramme and the World Bank and bas been one of the primary playes in woddwide efforts to meet the challengeof providingbasic water supply and sanitationservices to those most in need in the developlnewodd. Pawem in this venture a the developingcountres themselves and the muldlakral and bitera agencies that fund the Progam's actvies. Me UNDP-WoddBank Wster and Sanition Progfam producesits own publicationsand also reeas woks nder the auspices of the Wotid Bank. lhe Prgas publicatos ane divided into two seris, a Water and Sanitation Prgram Repast seies and a Wat and Saniatn Dicsion Paper series. Ihe PogmamReport sedes pets formaldiscussions of the Program'sopeations and researchactivities as well as exminations of relevat projects and tds withi the water and saniaton setor. Program Reportsare subject to dgorous extenal review by independet authorties from apprprate fields. The Discussion Paper series is a less foml meas of communic timely and topical o De,vation, findings,and opinionsconcening Progam activitiesand sector issues.

nhe docunoutha bas pnpd and pied by do UNDPWaddBank W S o Pg Waterand Sanitatoe Divhon, Iaftrauim aUd Uaa Dswlopmea puaumn, th WoWddBEL Copies maybe obined t Wser ad antda DM Mathe maybe quotd with pre atlubutlm s findelp, interpedon% and c_du,o expaed In tb papera tdWy toe of th anto and shoud n be atibuted any wa to he UNDP-WoddB Waer and San Prgm, the UnitedNatos D opa Progrmme, t Wodd ank or any afiad oMy mps tha ampany th. text havbeen piepa solely for theonvenienceof the medete designaton presetati of marl in themdo not Implythe ep on of anyopinionaoe n thepat of the UNDP-WoddBak Wat and Santtion Progm, the United ionsdDe roann the Wodd Bnk, or any alaed onlo Table of Contents

Table of Contents .... ii

A.clwledgnients .... via Exeutiw Summary.x 1. Introducion ...... 1 Objectives . . .1...... Policy Ove.ew... I lstitutional Responsibilities. . . 2 Levels of Service and System Costs. . . 2 Tariffs ... 3 ut Plas ...... 4 2. Study Design Methodology, and SiteSelection ...... , Study Depsi S..5 Methodology.. 5 Site Selection.. 6 SamplingWiJ Vg.es 7...7 General Description of the Study Areas .. 7 Sheikhmupra,Fasaabad and Rawalpid Ditrcs 8. .8 3. Analytical Overview .11 Principal Unes.of nquiry .. 11 A.ctalChoice Behavior ...... -.-.-.-.--. 11 HypotheticalChoice Behavior ...... 11 Economic Viability of LInvestmentln Rual Water Supplies ...... 12 Description of Vadables Used in the Analysi ...... 12 Explanationnd Rationale of ndependent Varabks ...... 14 EstimationTechniques ...... 18

4. Sweet Water Zone: Overviewand Field Procedur. .21 Policy Issues .. 21 Water inthe Sweet Water Zone .. 21 Selecion of theStudy Viages .. 22 The Questic,aire and Bidding Games. 23 Conductingthe Survey .. 25

5. Fmdings in the Sweet Water Zone: Analysis of Actual Choice Behavlir .27 Upgraing Service Levels. . . 27 Who Instas Electic Mbotoa?.. 27 Who CDnnectsto Piped WaterSystem? . . . 28 Electric Motor and HouseholdConnection as Alteatdve Choices. 29

Program Reo Series iiJ 6. Findings in the Sweet Water Zone: Analysis of Willingpess-to-PayBids ...... 31 HypotheticalChoice Behavior ...... 31 Service Options Offered ...... 31 Response of Householdsin Villages without an Operationa!Water System ...... 32 Responseof Householdsin Villages with an OperationalPiped Water System ...... 33

7. Estimated Revenues and Cost Recovery Potential ...... 37 Provisionof a StandardPiped WaterSystem in Villages Familiarwith Such Systems ...... 37 Provision of a Standard Piped Water System in Vllages Unfamiliar with Su Systems ...... 39 Provisionof an ImprovedSystem in Villages with an Existing Piped Water System ...... 40 Experimentswith Altemative Financing for Household Connections ...... 41

& Bracldsh Water Zone: Overviewand Field Pcedures ...... 43 Selection of Study Villages ...... 44 Description of the Questionnaireand Bidding Games ...... 44 Conductingthe Survey ...... 45 9. imdings in the Braclsh Water Zone: Analysis of Actual Choice Behavior ...... 47 Who InstallsElectric Motors? ...... 47 WhoCo nnus to Piped Water Systems? ...... 48 Electric Motors and HouseholdConnectios ...... 49 10. Fndings in the Brackish Water Zone: Analysis of Willingn, )-Pay Bids ...... 51 Wfillingnessto Pay for Connectionto a Stands._-_ystem ...... 51 Willingnessto Pay for Connectionto an Improved System ...... 52 Response of Households in Villages with and without Operational Piped Water Systems ...... 53 Response of UnconnectedHouseholds in Type A Villages...... 54

11. Estimated Revenues and Cost Recovery Potential ...... 55.SS Costs of Piped Water Systems ...... 55 Pwvision of a Standa Piped Water System ...... 55 Provision of an mptoved Piped WaterSystem .... 5...... 55 Comparisonof Standard and ImprovedPiped Water Systems ...... 58

12. Arid Zon: Overvew and Field Procedures ...... 61 Selection of Study Villages ...... 62 Description of the Questionnaireand Bidding Games ...... 63 Conductingthe Survey ...... 64 13. Fmdings in the Arid Zone: Analysis of Willingoess-toW-PayBids ...... 65 ServiceOptions Evaluated ...... 65 Response of Householdsin Vi-llges without Plans for Istaation of Public WaterSystems ...... 65

iv UNDP-WorldBank Water and Saiation Ptogran Willingnessto Pay for Connectionto a Standard PHED System with House Connections ...... 66 Wlingness to Pay for an ImprovedPiped Water System with House ConnCtios ...... 66 Multivarate Analysis of Willingwess-to-PayBids for Piped Water Systems...... 67

14. Esimated Revenues and Cost Recovery Potential ...... 71 Costs of Piped Water Systems ...... 71 Provision of a Standard Piped Water System ...... 71 Provision of an Impoved Piped Water System ...... 73 Provision of a Piped Water System Based on Public Taps .74 15. Concusions and Policy Suggesions .79 The Nature of Demand for Water in Rural Areas .79 Private Upgrnding af Service in Central Punjab .79 Ihe Ecnomics of VMage Water Supply Options ...... 81 The Arid Zone .82 Discussion.83

References...... 87 Appenidi8es. AppendixA .. 91 A1 VillagePofiles ...... 91 A.2 Rage Pofiles ...... 92 A3 Profilesby Vilage and Household Type...... 93 A.4 Patem of Household Choice over Available Service Optionsand Approximate Private Costs ...... 94 A.5 Reasons Mentioned by Households for Obtaining a Domestic Connection to Piped Water System ...... 95 A.6 Reasons Mentioned by Households for not Obtning an Domestic Connection-Type A Village ...... 95 A.7 Relationshipbetween Economic Standing and Choice of Service Level-Type A Village ...... 96 A 8 Level of Satsfaction with Piped Water Supply System-Type A Village .... 97 A.9 Major Shortomings of Piped SupplySystem Mentioned by Households-Type A Vi-lage ...... 97 A.10 Most Important Improvement in Piped Supply System Desired by Households-Type AVillage ...... 97 A1ll Household's Infonmation/BeliefsRegarding Characteristicsof Piped Water and Altendve SupplySystems .98 A12 Number of Years HouseholdsWould Keep HandpumpsOperational If a Piped Water Supply System Eisted in Their Viadge.99 A13 HouseholdPreference Regrding Responsibilityfor Operation and Management of Piped Water Supply Systems.99

ProgramReport Seria v Appedi B ...... 100 B.1 Decisionto Installan ElectricMotor. Results of a LogitM o ...... del 101 B.2 Decisionto Connectto a Piped Water Supply System:Results of a Logit M%odel ...... 102 B3 StatiscAnalysis of Willingnessto Pay for a StandardPiped Water Supply System:Resuts of Oibm Last Squas Regr-asionModel ...... 103 B.4 StatisticalAnalysis of Willingess t Pay for an ImprovedPiped Water Supply Systemby Householdsin Villageswith an ExistingPiped Water Supply System: Resultsof an Ordiny Lest SquaresReL,:ession Model ...... 104

Appendix C ...... 1...... l0S el Distributonof Wilingnss-to-wPayBids for a StandardPiped Water System in ViDageswithout an OperationdPiped WaterSupply ...... 105 C.2 Distributionof WTPBids for a StandardPiped Water Systemwith a Flexible Financing Arrangementin Villages without an OperationalPiped Water Supply .106 C3 Distibutionof W7? Bidsfor an ImprovedPiped Water Supply in Villageswith an OperationalPiped WaterSystem ...... 107 CA4 ConectionFequcies and EsimatedRevenues: Provision of a StandardPiped WaterSystem in a Type B Village ...... 108 C5 e Frequenciesand EsimatedRevenues: Provision of a StandardPiped WaterSupply System in a Type B2 Village ...... 108 C6 Frequenciesand EstimatedRevenaes: Prvision of an Improved PipedWater Systomin aTypeA Village ...... 109 C:7 on Frequeniesand Estimated Revenues: Responses to DifferentOptions by Unconneted Householdsin a Viuage with an OperationalPiped Water Supply ...... 109 AppeadixD ...... 110 D.1 Vilage Pioffies ...... 110 D2 Vfbg Prfles ...... 111 D3 Profies by V ageand HouseholdType ...... 112 I.A Patern of HouseholdChoe overAvailable Service Options and Approximate PrivateCosts ...... 113 D.5 ReasonsMenticaed by Householdsfor Obtaininga DomesticConnection-Type A Vage ...... 114 D.6 Reaso Mentioned by Households for Not Obtaining a Domestic Connection-TypeA Villge ...... 114 D.7 Relationshipbetwen EconomicStanding and Choiceof ServiceLevel-Type A Vma e...... 115 D.8 Levelof Satisfacionwith PipedWater Supply System-Type A Village .... 115 D.9 Major Sorcomis of Piped Water Supply System Mentioned by Houshods-Type A Village...... 116 D.10 Extentof Informion Regrding Parametersof PipedWater Systems.117 D.11 CoStof Piped WaterSystes ...... 118 D.12 PaRceptionsRegadingOuaityof Water ...... 118 D.13 HouseboldPefence Regardi Responsibilityfor Operations and Maintenance of Piped Water Sysems ...... 119

w UNDP-Warld Bank Water and Saton Pnun Appedix E ...... * 120 E.1 Deson to lns lland Blectic Mor Resultsof a lsogtM iodel . 120 E.2 Decisionto Connc to a Piped Water Supply System:Resuts of a Logit Mlodel ...... 121 E3 Statisical Analysisof Wilingle to Pay for a StandardPiped Water System: ResultsofOrda IrastSquaresRegressioModel.... 122 EA Satitcal Anays of W1nges to Pay foran Ipoved PipedWater System: Resulth )rdinarye Lat SquasRegression Mo .. 123 E5 Statisti Analysb of Wilingnes to Pay fot a Stad Piped Water System-4ype A Vilage: Results of Ordnay LAst Squares Regression kodl ...... 124 EA S"tisdc Analysi of WMilige to Pay for an Impwved Piped Water System-Type A VWage:Results of OWiny Last Squas Regresion Miodel .. 125 E7 Statsticad Analyss of Wihlgns to Pay for a Sandad Piped Water System-Type B Vilag: Resuls of Ordinay Last Squams Regssdon Model .126 Appdi F .127 F.1 Distrbutionof WTPBisforaSSmndadPiped W. System.127 P.2 Distributl ofWBi dsforanImprvWd Pped WatrSystm .128 F3 Disribuo of WTP Bib for a Sandad Piped Water System with Altenive 1naclngArngsanaitn:A2 Househds. .129 F.4 Distributionof WIP Bids for Sandud and Impred PipedWat Systemsin Vilags withOpetatin,Piped WaterSysm. .130 F5 ConnectonFrequewcs and EstmatedRevenues: Pomiion of a SandardPiped WaterSystm .131 P.6 Cnntion Frequenciesand Esmated Reenws: Prvision of an Improved Piped WaterSysm .. 131 F.7 CFrequednces and Esnaed Revenues:Provision of Opis to A2 Households..132 Appdi 0 ...... 133 0.1 ViasgeProfles .. .133 0.2 VflUgePoi. s...... 134 03 Profiks by VdilageType . . .135 0.4 Lev of Satisfactionwith PipedWatr SupplySystem-Type A Villpag 136 G5 Major Shortcoming of Piped Supply Systm Mentionedby Conneed Housekolds-.TypeA Vilg.s,.. . . 136 0.6 Cost ofPiped Water SystemsintheArdZone . . .137 0.7 ReasonsMentond by Househbol for Obmininga Domesic Connecion-Type A VIag ...... 137 0. Extent of Infomadon Regading Pametem of Piped Water System.138 0.9 Hueold Pefence Rprding Reqpnblt for Opetion sad Mainte of PipedWatwSyftm ...... 139

Appeai H...... 140

Progm Repot Series *i' H.2 Statistical Anl8ysis of Willingness to Pay for a Standard Piped Water System-Type B2 Villages:Results of OrdinaryLeast Squas Regression MIOdel... 141 H.3 StatisticalAnalysis of Willingpessto Pay for a Standard Piped Water System-Type Bl Villages:Results of OrdinaryLeast SquaresRegression MIodel.142 HA Statistia Analysis of Willingmss to Pay for an ImprovedPiped Water System-Typo B1 Villages:Results of OrdinaryLeast SquaresRegression Model... 143 H.5 StatisticalAnalysis of Willingnessto Pay for a Standard Piped Water System-Type A Villages: Results of OrdinaryLeast SquaresRegression Model .. 144 H.6 Statistcal Analysis of Wingess to Pay for an Improved Piped Water System-Type A Villages: Results of Ordnary Least Squares Regression Model .. 145

AppendixI .. 146 I1 Distributionof WTP Bids for a Public Tap System and a StandardSystem with House Connectionsin Viliages without Plans for Instaliation of Public Water Supplies... 146 I2 Distributionof WTP Bids for a Standard Piped Water System with House Connectionsin TypesA, BI, and B2 Vilages .147 13 Distributionof WTP Bids for an ImprovedPiped Water Systemwith House Connectionsin TypesA and B1 Viflages.147 14 ComectionFrequencies and EstimatedRevenues: Provision of StandaWdPiped Water Systemwith HouseConections in TypesA, Bi, and B2 Villages 148 I5 ConmectionFrequencies and EstimatedRevenues: Provision of Improved Piped Water Systms with House Connections in Types A and B1 ViUlages.148 1.6 ConnectionFrequencies and EstimatedRevenues: Provikon of Public Taps and Standard House Connectionsin Type B2 Villages ...... 149 Anppdix J ...... 15O J.1 ComparativeSize of Rural Localities in the Three EnvironmentalZones by Pooulaten Size (1981) .15O 3.2 Sources of Water and Light in the Tbree EnvironmentalZones (Percentageof Housing Unitswith Access,1981) ...... 151 J3 OccupationalProfile of the Three EnvironmentalZones ...... 151 34 Financial Cost (Rs) to an Average Household for Different Serice Options (Brekish Water Zone) .152 J5 Households'Choice of Sevice Level by Soieconoric Characterstics (Villages with PipedWata Supply) .153 1.6 Mean WTP Bids for House Connectionsas a Percent of HouseholdIncome . . 154 J.7 Costs of Village-LevelWater Supply Options (Rs) ...... 155 3.8 Cbsts of Vlllage-Lvel Water Supply Options (Rs) ...... 156 J.9 CompamtiverFeatures of the ThreeEnvironmentalZones ...... 157 3.10 ComparativeWater-Related Characteristics in the hree EnvironmentalZones . 158 3.11 Mean Willingness-to-PayBids for MonthlyTariff of Piped Water Systemswith HouseCoCDecdonsin the Three Zones ...... 159

vli UINP-Worid Bw* Water and Sati Progrm J.12 Actual and HypotheticalFrequency of Connectionto Piped Water Syswmswith House Connectionsin dheThree Zones ...... 160 J.13 Households'Willingness to Pay for connecion to Piped Water Systemby SocioeconomicChacmtetistics (Villages without Piped Water Supply) ...... 161 Maps ...... 163 SheikahupaDistrict .. 163 F5saiabad District ...... 16. RawalpindiDistrct ...... 1

Progrm ReportSeris ix Acknowledgments

We Wold like to thank a the officials of the Public Health Engineeting Dqprlment governmentof the Punab for their cooperationin the executionof this study. Our fieldworkwas made possible by a number of village facilitaos who ovexendW themselvesin familiarizingus withboth the isues and the peoplein the varous fieldsites. We would like to recordour appreciationfor the enthusiasmand courtesyof the followingIndividuals: Mr. Shafl Shad, Haji Sarda Mohamimad,Mr. MohammadAsghar, Mr. Nazir Ahmed Wattoo, Chaudhry MohammadAfial, MasterMohammad Aishad, Mr. Mohamme Nawaz,Chaudhry Mansab Dad, Mr. AbdulRahim, Chaudhry Selman, and Dr. MobammadRaza. We wish to expressour thanhalo the followingsupervisorsenumerators for the haul work and the manyideas they contnbuted:Mr. ShafiAhmed, Mr. AmjadAli, Mr. AsifHashm i, Mr. Mohammad Ramzan,Mr. Mohnmad Akbar,Mr. MoquetAhsan, Mr. ObaidUllah, Mr. ShaukatAlh, Mr. Kashif Haya4Mr. ShujaatYaqoob, Mr. Aqeeltirmir , Mr. SherazZafar Butt, and Mr. AmjadTikm. Mr. Arif Hasanfmiliaized us with the issuesin the rual watersupply sector and advisedIn the site selection;data entry was perforned by Shafi Ahmad;research assistance was providedby ObaidUBah and MohammadAkbar, the reportwas writtenby AnjumAltaf and typedby Mohammad Shafique and S. Qaisar AR AnjuaL Fmally,we would lile to acknowledgethe help, encouragement,and ideas offeredby our researchcounters in the UnitedStates, in paticular, JohnBriscoe, Xlnming Mu, and KerrySmith

Praon Report Series xd Executive Summary

Durig 1988 and 1989 a team frm the Applied EconomicsResearch Centre (AERC), Umiversityof Karac cied out a sty of households'wilfnpuss to pay for imprved water servicesin rural Punjab. Uhestudy was caried out In the groundwaterzones: the sweetwater zone whero&good qualitywater is easilyaccessible and where the officialpolicy is not to providepublic facilities;the brackishwater zone wheregroundwater is freelyavailable but of pooequality and where piped supply systems with householdconnections are sanctionedfor villages with populatons exceeding5000, and the arid zrne wherethe groundwateris of good qualitybut is difficultto access. The objectivesof the studywere to determine: * The wilingnessof householdsto pay for imprved servicelevels; * The determinantsof the willingnessto pay for improvedservice levels; * The preferencesof householdsregarding the managementof water deliverysystems; and * The appropriatenessof existinggovemment policy on the provisionof water in rural areas. To deternine whether householdswere actuallyprepared to pay for the costs of runmng improvedsystems, the ABRC team used a willingness-to-piy(WIM c 'contingent6 valuattoan methodology.WTP studiesare simplyhousehold surveys in which a memberof the householdis asked a series of structuredqutions designedto determinethe maximumamount of money the householdis willng to pay for a good or service. The three groundwaterzones differedfrom each other economically.With rence to the averagenural household income in Punjab,the sweet water area (SheikhupuraDistrict) was more affluent,the brackishwater area (FaisalabadDistrict) represented the averagelevel, and the arid area (RawalpindiDistict) was less affluent.The villagesselected were close to the districtheadquarters becauseit was felt that such locadonswould pesent immediatepolicy challenges as they evolvedinto periuban areas. Becauseof the officialpolicy of providingimproved services (piped systems with houseconnections) only in vilages withpopulations exceedig 5000,the vilages selectedin the sweet and brackishgroundwater areas were lage in size. In the arid zone, where the policy is not being enfored, smallersized villageswere selected. The AERCteam caied out 261 householdinteiews in five vilages in the sweetwater zone,495 interviewsin six villagesin the brwackishwater zone, and 401 interviewsin nine villagesin the arid zone. As expected,the percentageof householdsconnected to pipedsystems at the currenttariff was highestin the arid zone (96 percent),lower in the brackishwater zone (75 percent),and lowestin the sweet water zone (55 percent). The mean monthlytariffs that householdsin villageswithout piped water were willing to pay for piped systemswere Rs S0, Rs 40, and Rs 20 in the three zones respectively. The principalfinding of the studyis that in Ceral Punjab(sweet and brackishwater zones) there has been a quaitative change in the nature of householddemand, from one for water to one for water-basedamenities like indoorplumbing, showers, and flush tolets. In the abseoceof piped systems,households (all of which alreadyhave privatehandpumps) are providingthe highersevice

PNogramReport Sews xdii levei on an indivdual basis, using lectic pumps an. overhoadtank. Sixq percent of householdsin the brackish water zone and 30 pecent of those in the sweet water zone have installed such systems It wa estimated that in the brackish water zone in a typical village of 5000 people without piped water, houshds had already invested one million rupees In capital and were spendingapproximstely Rs 10,000 per mouth in opeation and maintenancecosts.

Ibis level of actual expenditureis of the same magnitud as the total cost of a public piped water system seving one hundred percent of the households. 'he aggregate willingness to pay for piped systems is also of the same magnitude The survey indicatesthat cost recovey of piped systems i8 possibie in C.ntml Punjsb. Piped systemsare conidered a substitutefor electric pumps. In villages without such systems in the brackish and sweet water zones, the pecentage of households with electric pumps drops to 33 percent and 11 percent respecvely. However, a considerablenumber of households(29 percent and 7 percent in the two zones respectively)continue to invest in multiple systems because of the poor reliability of piped systems. As expected, it is the Ticherand more educated householdsthat demand and are willig to pay for reliability. However, the demonstrationeffect of the choices of the elite is a clear indicator of the trend for the future. In tho arid zone, demand is still lagely for water for personaluse, and householdsare satisfied with a lower rliability of service, which is stiUlqualitatively superior to the alternativesof public wells and surface water. However, although the wilinges to pay is high, the small village sizes mean that tie capital costs cannot be recoveredbecause of the absence of economies of scale. The transiton to higher service on an individualbasis calls for a review of the policy of not povdig piped systems in the sweet water zone. ITis is necessary because the former option is socially inefficient, the monthly operation and maintenancecosts being almost two-and-a-halftimes those for piped sysms. The level of actual expenditure also indicates that piped systems need not be subsidzed in CentralPu4iab. A collectivesystem is futher necessitatedby the dainge Problemthat results direculy from incrsed water comumption fllowng upgrded service. This poblem is not amenable to individual soluto It is suggested that instituional mechams be explored to facilitate the prvate constrution and managementof collective water supply systems. Piped systems wil not become economicallyviable or be able to compete effectively with prvate opdons unless their .liability is improved. This is not possiblewithout metering because the demand for water at zero margi cost (namely, unmetevedconnecions) in a dry agricultural area such as the Punjab is immense. Without metering, water must be rationed by reducing reliability; people secure other sources; their whllingnessto pay for piped systemsdecreases; and the utility cannot collect sufficient resources to run the system efficiently. A majority of the sampled householdswere in favor of metering ra were willing to pay higher tariff that those who favored flat rates. IIn the small villages of the arid zone, more acceptable and manageablcviants of systems with public standpipes need to be explored. Contrauyto general opinion, ho-'seholdswere willing to pay a reasonableamount of money for such systems. The mean williness-to-pay bid was Rs 35 per month, and 84 percent of the households indicated that they would subscnbe to such a semice at a tariff of Rs 15 per month

xiv UNDP-WorldBank Water and Sntaion Program V%A

to~:

11111t

{ ~~~~~~~~~~~~~iLii}A 1. Introduction

1.1 Ihe World Bank has been placing increasing emphasis on rual areas in its lending in the water sector. Consideing all Woid Bank expentes on water supply, the proportionof fund allocated to uroalprojects has ineased from an average of 8 prcent between 1974 and 1980 to an average of 14 percent frm 1981 to 1985. However,a Bank review of projects in this sector concluded that oveall perfomce was disaonting. Te review suggestedthat technologyper se did not appear to be a major problem. It was concludedthat the desig of rual water supply projects had been ovedy supply orlated and that crucial demand aspects had been neglected. In particular, it recommendedan emphais on undeding (a) what people want and (b) what they are willing to pay for.1

12 In light of the above conclusions a research study was approved to find ways to improve the financial and economic performance of water sector projects by developing improwd information on households' wiliness to pay for upgraded services in rual areas. Ihis reseah project, titled 'Willingness to Pay for Water in Rural Areas," involved field studies in six countries: Brazil, Nigera, Tanzania, Zimbabwe,India, and Palkstan Tis documentreports on the results obtained from the field study carried out in Palistan. Objecvs

13 The objectives of the study were to determine the following:

e The willingnessof households to pay for improved service levels; * The deteminants of the willingmessto pay for improved service levels; * The preferencesof householdsregading the magement of water deliverysystems; and e The apprpriateness of the existing governmentpolicy on the provision of water in rural areas.

1.4 An additional objective was to test and validate a rapid appaisal method (the contingent valuation method to be descnrbedlater) as a potential tool for planners in designing rua water supply systems. It was hoped that the use of this method and the informationderived from it would help the water authonties improve their decion-making on appropriatelevels of service, cost- recovery polices, and water pricing in rural areas.

Pory Oveviw

1.5 Coverage,Targets, and Alocons. At the start of the Sixth Five Year Plan (1983) 22 pecent of the rural population of Paldstan was considered to have access to water of acceptable

1Wodd Bank: Water for Rural Cmmunitks: He4dng People to Hel ThaehePIS Policy and Research Division, Water and Urban DevelopmentDepartment4 25 pages, May 1987.

Program Report Series I quality. Identfig rurl watersupply as a neglectedsector, th pl announceda doublingof ual cover to 44 pret by the end of 1988&By the mid-poit of the SixthPlan, however, pgress was wel below btrg withonly 28 pecent deemedto be coverd.

1.6 Und the 5-Point Prgrmme announcd on December31, 1985 rural water supply was agin identif as a pdioity sector,and a ta was set of incrasing coveragefrm 28 peret to 66 percentby 1989.While the mtron asated with the 5-PointPrgamme is no loger in office, receot poy po emts have indicatedan even greatercommitment to rua water supply.Th tentatie tat beingset for the end of the SeventhPlan in 1993 is 75 pcent.

1.7 Virtuallyall resourcesfor the nura water supplysector flow fromthe deral to the povinci govenments.Accoigy, the fedea govrnment is the criticalpolicy-making level of governmt, with the prvincl agencies having responsibilityfor Implementingthe policiesset at the fedel level. 18 At the federal level the Minty for Phang and Development(Plning and DevelopmentDision, Physical Planning and Housing Section) is the key policy-makingbody with majorrponsiity for icorporatingrurl water supplyinto oveall deveopmentplans.

1.9 The Minsty of Housing and Worf (Envnent and Urban AffiiBmDivision) has the major responsbility fo the technical aspects of rual water supply. TMetechnical agenciesat the proinciallevel must rport to this ministry.

1.10 At the pmvcil level the Public HealthEngineering Deparment PED) is the tebnica agency which has responsibilityfor the consuction and itnia opertion of maot rur piped water supply systems Te local authoities (Dtistct and Unio Cowuni), under the Depatmentof local Govenmentand Rurl Deveopment,ane supposedto be responsiblefor the operationand _ainweof publicwater suy systems In fact, becauseneithr of these bodies dischges this responsiy i, in most instces, it is assumedby the PHED.

1.11 Ihe major consquence of the dominanceof the rral water supplysector by the PHED is the over-emphasisof the engineeringaspects and the relative neglectof te socia and economicaspects of sectoal issues.

Leek of Sevice and SyskemCots

1.12 DecisIons on the lvel of sevice to be prvided by government ral water supply progams ave made entiely on techncalladmnistrativegrounds. Ihe efficiencyof these policies from a stricty economic pempecive has yet to be established. In Puaab, for ince, the curent policy regading sevice evel i as iolows:

2 UNDP-World1 Waferand Sadon Pogrm Sweet water areas: * Public supplies are no longer bilt. Brackishareas (with canal irrigation and add arm): e Standpipes for villages with populaton below 5000. * House connectionsfor villages with populationabove 5000.

1.13 With re exceptions,the govment bews 100 percent of te capital costs of public water supply poets. In Punjab and Sind typical capital costs for piped sstems providing houe conecdons are as follows:

Areas wih swee water (supplied by tubewell): Rs 450 per capita in Punjab and Rs 620 in Sind. Imgaied areas with brackish water (supplied with filtered canal water): Rs 600 per capita in Punjab and in Sind Rs 900 (where perena canal water is available)to Rs 1300 (where canal supply is nonpenial). Arid areas Rs 120041800per capita in Punjab.

1.14 Based on experiencein Punjab, standpipe systems genaly cost the PHED only 10 pert to 20 perent less per capita. In parts of Sird, where settlementsare more scattered, the unit cost differences between house connectionsand standpipes are greater. The costs to the householdfor a domestic connectionare typically Rs 80 to the PHED as fee and about Rs 500 for the matedis asd hbor to effiectthe connection from the distributionline.

1.15 Ihe cost to the PHED of operating and manning piped water systems (based on an aveaagehouehold size of 6.5) is estimated to be between Rs 20 to Rs 30 per householdper mont In Puniab and about Rs 35 per householdper month in Sind.'

1.16 In principle, after a two-year delmonstraton period," District or Union COcIls should tkel over the running of piped systems. In fact this does not happen, and the PIlED codnues to bear the resynnsibilityand the expenses.

1.17 Public tap usedsare supposed to pay Rs 5 per fimily per month. However, plaly because the users are dissatisfied with such systems (they dsire house conncrtions) and parily

'These cost figures of Rs 450 and Rs 600 for Punjab are reported by Briscoe (1987). The figues reported to us were Rs 300 and Rs 50, respecdvely.These are also the figures used by the PliED in preparig cost estimates (US$1 * apprximately Rs 20 dwing 1987-1989.) 'Again, the figures for Punjab reported by Briscoe seem to be on the high side. Ihe PHED uses a figure of between 3% and 5% of capital costs as the annual operation and manenance (O& estimatm. These generally work out to be lower than the figure mentioned above.

P hograRot Sries 3 becauseof the difficulidesinhe_ t in collectig for standpipesystems, essealy no attemp.is made to collectthe user fee.

1.18 For systemspoviding house connectos, the monbly taiff is betweenRs 10 and Rs 25 per conneton. This s a flat talftf, sincehousehold water ppliesin Pakistanare not metered. Collectionis unevaenin some areas most who are connectedpay, in other ares compliane is less. In Punjabte PHIEDbas no data on ovll compliancebut bimevesthat about 60 peret of those connecd pay.Ihose who do not pay anethreatened with disconncon, but in factthe threatis rely carriedout. Future Plan

1.19 Costrecovey is beginningto emergeas an importantisue in the financingof rural water supply projects One proposedsolution is the proposal for the establishmentof village developmentcommittees. ITis is seen by planneisas a necessarystep in the assumptionof local responsibilityfor the efficientoperation and maintenanceof vilage water supplies.

4 UNDP-WorldBank Waoerarsd Saato Praun 2. Study Design, Methodology, and Site Selection

SWdyDealg 2.1 To acieve the objectv of the study, the proposedreseah design underlook fieldworkin threedifferent regins chosento covera uge of economicand enonmental conditions. In each goon two types of sites (Ipe WA'and Type 'B*) wae to be chosn Type A site were to be locatedin an area wheoea fmctoningimpoved wat supplysystm wasbing usedby between 30 percet and 70 pert of the populton. Type B sites wee similar ones neatby at which an impovedwer sopplysystm was not yet avilable. It washoped to conduct200 houeholdinteviws at each she yielding an overall sample size of 1200households. 4

2.2 SinceType A siteswould include both types of housholds (connectedand volunwtaly unconnectedto an availableimproved wate source)It wouldbe possble to use an indiect apprah (on based on obsevation of actl cboies) to unesm householdbehavior. Thus, it would be possle to ess the effeas of differt dcharacteristisof impwved and altnatie soures (pre, dstnce to source,quality, level of servce, elbt, timesie Itallation, etc.)and users(economic, sca and demogmaphicacois) on the liklihloodof a household's using an imprwed source. 2.3 Since no observationson actualchoice would be availablein type B sites, a direct approach(the wilingness to pay [WTP),or contigent valuationmethod) would be used to eicit ho Dseholdpefrence for impoved service. 2.4 WV studiesare simplyhousehold sweys in whicha memberof the householdis asked a stuctured series of questiowsthat are desiged 1o deterin the maximumamount of moey te homseholdis wlling to pay for a good or service.Whea WIP stdies an coonductdto assistwith water sectorpolicy or planning,the specifiedgood or servicecould be a houseconnection to a piped distibutionsystem, accoss to a handpumpor standpostKor prvision of householdsanitation facilities. WI? studiesare also termed'contingent valuation' studies because the spondet is ased aboutwhat he or she woulddo in a hypothetical(or contingent)situation.S

4Wodd Bank, "Willigness to Pay for Water in Rual Areas,' Research Prposal, Water and Urbn Delopment Deprtment, March 1987. SFordetails of the methodology,see Ronald 0. Cummings,David S. Brookshire,and William D. Schulze (editofs),Valking Envwmenta Good: An Assemn of the ContingentVabaonl Mewd Totowa, New Jersey: Rowman and Allanbeld, 1986.

Program Report Series 5 2.5 Comparsns betweeanthe results obtained from Type A and B sites (e.g, coinecdon freqency at given taff sates) shoud help to validate the reliilit of the WTP method as a tool for rpid appisal of paning cboices and allematives.

Silo Selecaon

2.6 The overall research frameworkproposed for the mulu-countrystudy was tailored to the needs of the specific conditons in Pakistan Deliberaions by the Pakistan team foflowed by consultatio with the counterpt staff at the World Bank led to the following decisions:

a. Restction of the studyarea to Punjab. Since the WI? surveys involved techniques that were unfailia to the enumeratorsand that invol%edexusive interviewin it would be adviable for team leades and interviewsupervisors to be familiarwith the languag of the respondents. IThs effecdvely restricted the sample area to Punjab. In em of poplaton, Punjab is the largest province in Palistan with S7 percent of the rual population.

b Within Punjab, three eviromental zones were identfied based on groundwater chaactristics: the sweet water zone where good quality water is easily accessible; the brackish water zone where water is easily accessible but of poor quality; and the arid zone where water is not easily accessble though of good quality.

cv Restriclion of impwved source to piped distribution systems with household connrions. The researh desig deemed it important that householdsbe paying a money price for the improved water supply. Tis limited the choice to piped distributionsystems with householdconnections since tariff collectionfor handpumps or public standpostswas virtually nonexistent

d. Selectdon of lre villages. Partly as a result of the restriction ntdoned in paragraph 1.12 Ci Punjab householdconnections are supposedto be prvided only to villges with populationsexceeding 5000) the sample villages had to be large in size. It was also felt that fom a planning perspective the many lare villages close to major titieswould present challengingand immediate policy problemsas they would evolve into peduban tDwnswithin the next five to ten years. Anaysis of such locationswould be pudculary useful for the planningauthorities. However,

where possible, as in the arid zone, smaller vilages were included -n the sample.

e. Disaggregationof Type B villages It was decided to include two kinds of Type B villages - BI where an improvedsupply was scheduled to be installedin the near fuue, and B2 where no such supply was scheduled. Tbis was consiered to be helpful in testingwhether any strategicbias vas incorporatedin the WI? bids. Such a bias would occur if householdsin Bi villages, already assured of a water supply, systematicaUyude-eported their bids in order to convey the impreion that they would not cect if the tariff were raised. Householdsin B2 vlages, with no prospects of ccess toa piped supply in the foreseeablefuture, on the other hand,

6 [DP-Wrd Bank Wter and Saitato PrWo mightsystematically over4ate their bids in the hopeof influencingpolicy-makers to includetheir viae in thasemare for the stlation of piped watersupplies.

2.7 Basedon tie aboveconsidetions v s wereselected and surveyedin the foRowing arem:

Sweet water zone - ShelkhupuraDistdct Brackishwater zone - Faisabad Distict Arid zDne - Rwalpind District

2.8 Ihe loation are depictedin mapsat the backof the book. Accodingto information obainedfrom the PIE (1986 a nately half the rual populationof Punjabresides in the sweet wat zone ad a quartereach in the brudksh and arid zones.

Sampingwhin V es

2.9 Since no householdlists were available,the followingprdure was adoptedfor amplingwithin the selectedvillages. Roughmaps of the selectedvllages wereavailable fomm visits pior to the mainsurvey. Villageswer dividedinto our a xately equalquadrants, each quadrant cmprsing a knwn number of lan Two eaumeaors wre assigned to each quadrant,one rsposie for houseson the left side of a lane ad the other for houe on the right Mhetotal n_mberof hous i the villagecould be estimatedfrom cemus recods Since the samplesize had been broadlyspecified in the resh design for any pardcularvillage the numberof houes to be sampledwas known. Me first houseon each side of a lane, for every hne, was chosenrndomly and te every nth housewas sampledbased on the samplingratio. Householdswhere the appropdate esponde was absentwere not roplacedby nearbyhouseholds. Lre villages,which rquired more than one vit, weresureyed bothin the monngs and in the evengs. Smallervillages, vhich we visited only once, were surveyedin the evenings. Iis was meant to ensurethat no occupeaona category (e.., fmers or wage employees)were symatically missedin the survey.

Genrl Dusrpiln of tm StU Areas

2.10 Pnjab. Pujab, thelagest pvice in PaIistan,contais 57 pct of the tota rua popuaton and 56 pent of the totalanmber of rml localitiesin the couny accordingto the 1981 census. The size distributionof ra locities in Puzjab virtuallyidentical to that c uII the countryas a whole(see TableJ-1). In 1981,about 3 percentof the vigs exoeodeda popuiation sizeof mo, 22 pemrntexceeded a size of 2000,and 46 percnt exceededa size of 1000.' However, iu terms of the prporton of total nuralpopulation, 17 peroentresided in viages exceedn a size of 5000, 57 percent resided in village exceedig a size of 2000, and 83 peet resided in vilages exceedinga size of 1000inhabitt

'At the 3% populationgrwth rate per annumused by the PHIED,the prpotin of largervMages wouldbave _ad. considerablysince 1981.

PosaM Repot Series 7 2.11 The level of economic affluence in rural Punjab is ako very similar to that of rural Pakdstan. In 198441985the monthly rul household income per capita7was Rs 243 in Punjab and Rs 234 in Palitan as a whole.

2.12 In Punjab Province,21 percent of ral locaities (15 percet of rual houswholds)had accs to electricity in 1981. By 19861987 the percentageof rural locaities with access to electicty had ineased to 49 percet Ts is impotnt since conetion to the electric grid faciitates the economic provWionof piped water.

2.13 Accordingto the 1980 Housing Census, 45 perct of rmal households had acces to water inside their homes. Of these, 3 percent reLiedon piped water, 37 percent oo handpumps,and S percet on wells. The emaig 55 pent of householdsobtained water fom owide their bomes Of these, 3 pert telied on piped sources, 15 percent on handpumps,16 percent on wells, 4 perct on ponds, and 17 pert on springs, rivers, strams, etc. (see Table 1-2). Thus by 1980 6 perct of rual houseolds had access to piped water, 3 pecent to domestic connecions, and 3 percent to public taps. Access to piped water is stated to have doubled by 1986.

FababbpuaFmad Rawalidl Diticts

2.14 Within Punjab, Sheikhupura Distict (sweet water area) and Faisalabad District (-ckish water area) lie i the entral canal-irrigatedplain whils RawalpindiDistrict (arid area) lies in the norther ain-irigated region.

2.15 The size distribudonof nu caitdes is shown in Table J-1. It can be seen that the proporton of larger-sized villages is much higher in Sheikhupura and FisabaWd Districts than in Iawalpindi Distict Thus 26 percent of the urallocalities of SheikhupuraDistrict and 60 percent of Faisalabad Disrict exceed a population size of 2000 inhabitants. The compartive figure for RawalpindiDistrict is 11 percnt l-is is probably due to the fact that the available water resources in the arid zone cannot support lrge-sized setlements.

2.16 Data for Sheikhupura,Faisalabad, and RawalpindlSub4istricts are also provided in Table J-1, since the samplevges are locatedmostly wihia the sub-ditricts. ITe data show that the first two sub-districtshave a higher popwdon of larger sized villages than the districtsas a wholea

2.17 The monthly rual household incomes per capita for Pakistan and Punjab in 1984- 1985 were Rs 234 and Rs 243, respectively. The cemparative incomes for Sheikhupmua,Fas _abad, and RawalpindiDisteits are Rs 307, Rs 247 and Rs 217, respectively. The age in Punjab is fom Rs 175 to Rs 320. 'Ihus Sheikhupurais among the more affluent districts, Faisalsbad represents the average level of affluence,and Rawalpindiis among the less affluent disticts in Punjab.

7'Te montddyrua household income is obtaed from the Household Income and Expenditure S-waey, 19841985. The average rura householdsize Is obtined frm the Popuation Census, 1981. Pab DevelopmentStatistics, Bureau of Statistis, Govemmentof the Punjab, LahI;e 1988.

8 UNDP-WorldBank Waterand Sanito Pogam 2.18 Ihe occupatonl stctu in the three districts is shown in Table J-3. Sheikhupura and Faisalabad Distrcts have vey similar pefils wbile Rawalpindi District has a much lower proporton of the working populationenged in industryand a higher proportionengaged in services. The proportion of the woring population engaged in agriculture is virtually the same in all three districts and accounts for just over half of the totl working populatio

2.19 Ihe comparativesituation with respect to acces to water is shown in Table J-2. In Sheikhupur and FaisalabadDistricts, handpumps inside the house are the prmary source of water: 72 percent and 64 percentof total households,respectively, rely on such sources. The second main source Is handpumpsoutside the house. In Rawalpin Distict, on the other hand, 62 percent of households rely on wells outside the house and another 25 percent on springs, rivets, and streams. Ihe modeling of houv"flds' wate source ies thus much more complex in the arid zone especially because many sources dry up during th summr months and are not accessible.

2.20 Zones with accessble groundwater have a lower proportion of households supplied with piped water than the average for the prvince as a whole. his situadon is the result of current goverment policy, which is based on need and not on demand.

Pro.n ReportSeries 9 3. Analytical Overview

3.1 mheanalysis hall be dietd wwa te investato of the foowing broadiss in each evironme zone:

* Actual choice behvior, * Hypotheticalcic bear, and * The ecanmic viabilityof investmentsin rad water supplies Ibe detailsof to Issuesto be stigted and the techniquesto be used fr tat ppe are discussed IAte folowig sections.

AMsa Chaim Baw

3.2 In each evonmental zone householdsbave a number of bhocesia deciig upon the type of water soure to use. The study would identifyfte set of choiesand infer fkm the obevadtionsthe effect of dift riablesao te liiood of a houOls makn a panricul chice. chbBeh H atetica Choic Behavi

3.3 In mny villages aticu servic opon, e.g, hous to a plped system, may aot exisL Such an optionis not pa of the choie set of households e valution of sch a hice is obtined by offeing te oeholds a well-descred hypothedtcaldio in a big gme forma the respones, in tm of wllingpes-to-paybids, are e sod t9 estime the valueplacoe on the pulr opt offered.

3.4 The desaipion of the hyWhe choice Inludes peentat of all the welevant chrctists of a spedfic piped senice opton. hse iclude hours of py, qualityof water, eped p ue and eiaty, and te pricesunder which service would bo avable. In ?Pmb the price stcue for piped sevice includesa on-time churge(an ofl co o fee patyble to the pls the cost of nectg the house to the dui_o le, wic is bone by the housold ) and a fiat onty taf In this study,the amout of the otime MosMare includedas pa of the dscripto C ilages wherethe d buiones have not yet bee Id, the bod are give an a e cost figue based oanthe s In villageswih pipd wer systems). The hols ae then asked whetherthey wouldcon to the systemdescribed at a given monthly ti Dqpndi upon the response,the tadff is maisedor loweredadd the hoolfd is asked to respondagai. hs pat is folowed consetly in fte study: the o m cos are always Indicatedto the hou ld and the biddingis on the levd of the monhly ta

?Serm 11edl 3.5 'Me willingness-paybids are used for thu dift pupos:

* To detemie th popoin of householdsthat would accepta partcular sevice optionat any giventulZ whbichallows caculation of the possible revenueyield; * To studythe detma of t willingess-t-py bids andtherby to identify posible teds as soil and ecoomiccondto chan in the futu; and * To compare,wher possible,the pattn indcatedby the hypothcal choices(eg., comecdon atio at a givea tadff level) with the patter refectng actal choice behavior. Such a eoLpaison would serve as one test of the validity of the contigentvaluatio methodology.

EcomomicViab_ty o s In Rua Wate Stuppu

3.6 Ike wiingstopay bids for any pardcularopdon wouldyield an estimte of the numberof householdsthat wouldchoose the opdan at a given pdre. In the case of water souc hese responseswould be intpreted in term of the choice to onnect given a paricular conection cst and tauiffleveL

3.7 Ibe aboveinfomadon on the pecentageof householdsconnecting to a servicelevel at varin taff rstes is sufficientto esdmatethe monthlyreveues that could be genated if that service wer actuay offered. This estimatecould dte be compar with actualdata on maintenance andcapita casts to Jetmnine the degree to which cost recovey is possible in any specificsituation.

Dsiptn .o Vales Used In the Antals

3.8 In expainig householdbehav, either in tms of actualchoies or willingnessto pay for hypothtcal choices, sponeswould needto be relatd to a set of indpendentvariables rpeseng bothsoce and bouseholdc tuics. The followingIs a complt lstn, descripti andrationae of the idepdent variablesused in the nultvrate analyses(Not all vaablesmay be presentin any specifc analysis.)

3.9 We Ident a set of broad factos that are relevat to an analysisof water-related behavior.To captue pacular aspec of thesefactors, a numberof vadablesane speced withi each e Te factrs, vaiables, and their indcaton are gi below.

Factor Vale S _eclflcatias

Need for water - Householdsize Numberof householdmembers - Cnsumpdoinof water liters per capitaper day - Ownershipof animls Numberof cattle ownedand kept insideor just outsidethe house

12 UNDP-Wl BAv*Water and Saadan PmVa Factor Varbe SIMfe,

Availble r - Adult women PorcentIn household for odlecng water - Chldroe Peroen in household * ddlad as a b_oubd semba

Abiity to pay * Householdexpenditure Rupeespr capitaper month - COnrstnctionvolue of Rupee house -Ownershipof land or Dummyvariable: 1 If household property owns lnd or other prperty, 0 otrwise

Exting arrgem - Privatesource Dummyvariable: 1 if householdhas for water a privatesource, 0 otherwse - Vendoms Dummyvariable: 1 if household uses vendors,0 othewise - Time Minute of bouseholdime (onewa s spentper week in obaing waterfrm outide the house

Qualityof water - Quality Dummy vaiable: 1 if respondent cuntly used conside water safe for health, o otewIse

Pesonal I - Age AV of household head of householdmembers - Education Numberof yew of educationof mast educatedhousebold member - Occupatol Dummyvarble: 1 if occupaionof householdhead is reated to farmig o othewise Sex Dummyvariable: 1 if male, sponden, O othewise

Household - Externalexposure Dummyvaiable: 1 if any male attitudes member has had expue to RUb outsidevillge ftr a periodexceig six months,0 otherwie - Aware_s of piped Dummyvariable: 1 if respondent systems has visiteda villagewith a piped water System,0 othewise

PF08ramReport SWei 13 Facow Vrhe S

- Meteing of water Dummyvariable: I if respondent eels Suppies domestic house connection ought to be metered, 0 otherwise . Respons_lty for Dummy variable: 1 if respondent prson of water believes water ought to be pwvided free by the state, 0 otewise - Satisfactionwith Dummy varable: I if respndent exaisg water In vilage with a piped system is System satisfied with its management, Ootherwise

Ote fcors - Dace of house from Yards distribution sle - nce of iae from KomeAes proical headquarters - Prximity of viae to Dummy variable: 1 if village is perenal source of close to perennialsoure of water water, 0 otherwise

Control variables - Village dummies Identication of specfic villagles - Houebold type dummies Identfication of household tpes Al, A2, B1, or B2 - Strtg point dmmy Identifieswhether the bidding game was srted with a low or high valu.

awd Ratkookof luspelndat Varies

3.10 Need for Water. Ibe pref for an impwved soure ought to be directiy related to the need for water. Ihe most obviousdeterminant of need Is househo! size. The water needs of anfims, espeally cate, an also slgificant. In the hot summersof the Punjab, catte have to be kept cool by being allowed to wallow in pools of water or by being hosed down. Because this entails comidembl expenditu of householdtime, the comnenenceof having running water avaflable widtin the house ought to be an attuactWepr ipositn. (Infmation on the ownership of anim was not obtained in the sweet wate zon.)

3.11 TMetotal bouehold need for water is also rlated to the level of consumptionof water per capta per day. n householdswhere the prmary sourceof water supply is idtherhandpomps inside the ouse or tp it is very diffcult to esdmate he totl quantity of water consumed. The task is somewhat moe mangeable In vllages (as in the arid zne) where most of the watr is eidtherfetewhd

14 UNP-Woradl*k Water and Satation NW= fbm oide the home or deliveredby vendo. We haveexpeemted withesimating the minimum amountof wats a oumboldwould normally need ftr its elsentialuse. M&sesmate is detmined by psingd the t with a hypothtcal situationin which the household'sprimy sowrcewas out of adr or uder por (a situationwith whichall householdswere familia). lhe respondentwas akd to ndlo ttheveel wh1h waterwould be fetchedfhom outside th housein sucha situadon and IDgive tho numberof vess that wooMbe refuiredto meet the ssntial needsof the houe (it s resoable to assum hat households would cuti nonestal usage in such situations). Ile capecityof e vssel was estmated by asking t respondentshow much milk it could hold, a judgmet whih rpoadnads made with great ease. In the arid zone, where householdsare hmilisr with the vwedg of water,rsponden wer simplyasked to state howmuch water they would acqu fm a vndor for meetig essenal householdneeds on a representativeday.

3.12 Of comae, the vaiadon in essentialcoumption would be much less than the variton in toal consumpto However,under the cirumstances this was the only quantitative measue of wate comption that it was posible to estimatewith any degreeof accuracy. 3.13 In householdswith accessto impoved water sources,the consumptionof water per capt per day, ev fr essntd usage,might be higherthan that of comparablehouseholds without improvedsources. Suh etdogeity wouldmake It difficultto interpretthe vaiable.

3.14 Awalabi LaborSwpply. Whoreverwater is primarilyfetched from outside tho house, as in the aid zone, e avilabblelabor supply (which, in the contextof fetchingwater in the Punjab, y omen and chiden) assumescddcal significac Householdswith scarcehbor ought to be man desro of imprved water soures.

3.15 mhmchoice of the proportionof womenand childrenas idependent vaiables means tht tho xcludedcaegy i the proportionof men. If an increasein the pportion of womenat the expne of men makesoly a marginaldifference to the need forwater, the pteferencefor an improved sourcewould be lesend bmse of the increasein the availablelabor supply. However,if a woman needs mor wat than a man (iuabilityto batheat naturd soucs could be one reason),the overall mpactcould be ambigou An incre in hildrenat the expenseof men,on the other hand,could be expecedto both deee e need for water and increasethe avaiable labor supply. Therefore, an incre the prpoion of childrenshould have an unambiguousnegative impact on the preferencefor expeniv impoved sourCeSof water.

3.16 AbNi to Pay. A greaterability to pay could nomally beepected to be direcdy nrlatedto a greate pernce for an impwvedsource of water. However,if moreaffluent housetolds bave already made significat investmentsin expensiveprivate water supply argements (eg, Installatonof an dectric motoron a handpump),the impactcould be ambiguous. 3.17 Mm eation of rral hoaseholdincome is genealy very difficultbecause of seasonalvarao and noncashcomponent (eg, own production).Te effort requred for accurate estmation is not feasie in suveys where the prmay objectiv is not the estmation of income, as, for example,in willingness-to-paystudies. W6 have used monthlycash householdexpenditure per capitaas a proy for bouseld inwomeper capita. M variableis subjectto limitaos sice it is

Prqga ReportSeries 15 unableto corWectfor the possiblelower cash epedit of faing houselwds. (Infomationon this variablewas not obtainedin the sweetwater zone.)

3.18 Te estimationof householdweath is also a dMcult proposidon.We have usedtwo proxyvariables The ownerhp of land or other popet is faly stragtfo The construcdon valueof the housein whichthe fmily rendes could be more controversial.It has been hfequenly documentedthat the highestptiority In the use of discretionaryfunds is given to housing. This is especay true for rual householdsi Pakistanwhere housingis a very visible Indicatorof socidal status and achievement The pattem of utilizationof remittuces fom the Middle East by rurd recipientshas establishedtbis beyonddoubt. Thus, housingquality could be a reasonableindicator of householdwealtL We have measured t by askingthe amountof moneythat wouldbe required to consuct a house similarto the one occupiedby the hosehd if the constrction were to be cared out at the time of the survey. Ihe latter conditiontakes care of the problemsinherent in the possiblevaluation at diffetentpoints in ime.

3.19 One frequentlyused alternativeis to enquireabout the ownershipof householdassets (usuallydurable goods like TVs,motorcycles, etc.) and to couct an indexof asset value. We fel that the valuationproblems associated with this procedure(when wer the goodspurchased? at what price? new or used? are they in working order? do they really exidst?etc.) result in an estimate whose reliabilityis not likelyto be any betterthan that of the indicatorwe used. Thisis especiallytrue since e residece canbe consideredto be the pmry and most valuableasset of the household,dominating assetslike TVs, radios,etc, and is also visibleto the enmerator. Thebenefits of relyingon a single indicatorare savings in terms of time and less apprehensionon the part of respondentswho are genely suspiciousof enquiriesinto ownevihipof assets (MVs,radios, and motorcyclesrequire licenses)and reluctantto admit to socialdeptivation 3.20 4sidngArrang _nenfir Water. The impactof the ownershipof a privatesource of uvsteron the preferencefor a pipedsupply can be quitecompl If the souce involvedhigh recent capital expenditureand is an acceptablesubstitute for piped water the impact cwld be negative. Hfowevr,if it is not considereda substitute(e.&, does not providethe same qualityof water,as is the case in the brckish zoe) the impactcould wel be positie, being an india of the desirefor imprved sevice. At the same time, if the runningcosts ae hlger tan the tariff for piped water and the capitalcosts are recoverable(e.g, by the sale of used equipment)the impactcould againbe positive.

321 Householdsusing vendors could be expectedto expess a clearpreference for piped water,since vendedwater is uwuchmore expensiwe than the flat montly rates at which piped water is suppliedin the Punjab. However,since existingpiped water systemsare genally unrelableand requireinitial eco costs,the choicebetween uSingvendors and obaining a privateconnection may be less clear cut thnmit woud first apper. Also, puschang water fiom vendorsallows houeholds to have gsater controlover their casb flow.

3.22 Ibe excluded households,sccording to the dummy variable, are those which exclusivelyuse their own laborsupply to fetchwater fromotside the house Such householdsexist only in the arid zone, since in the other two ones virtuallyevery householdhas at least a private

16 (NDP-WorldBan* Waterand Sanitat PmV handpumpinide the house. ITe importantimplication is that in the arid zone the amountof time spent fetchingwater is highlycorelated with the ownershipof a privatesource of water or the use of water vendors. 3.23 In the sweet and braclkshwater zones, householdscan be distinguishedfiom each other by whetheror not they own an impoved privatesource of water whichconsists of an electric motorinstalled on the handpump.However, since the qualityof the groundwaterin the brackishzoe is genally poor mosthouseholds spend some time in fetchingwater for drinkng and cookingfrom outsidethe house. I: can be hypothesizedtdat householdshaving to expendmore time wouldexpress a greaterprerence for a connectionto a piped water supply.

3.24 Qualy of WaterBeing UsedA Households that consider the watercurrently being used to be unsafefor healthcould be expectedto demonstratea greaterpreference for a pipedwater supply. This is not an objecdveassessment of water quality,rather it dependson a certaindegree of health consciousnessand a judgmentabout the qualityof the availablewater source.

3.25 Peonad Characis. It is generay believedthat youngerand more educated householdswould express a greaterpence for modemconveniences like piped water supplies. 3.26 Householdsinvolved in farming landlords,tenants, and owner-cum-tenants)are libely to expess less interestin piped water supplies for a numberof possiblereasons: flexiblework schedules,the locationof work close to water sources,and the ability to combinewater-reatefd activideswith work (e.g.,bathing and wateringof animals). The generallypresumed conservatism of faiminghouseholds could be a contributingfactor.

3.27 It is generallyassumed that womenwould express a greaterpreference for pipedwater sources tman men since the former are primarily responsible for the arduous task of fetching water. Respv-sesfrom both sexesare availablein the arid zone to enablethis assumptionto be tested. 328 Atitdes of Respondents.Households which include members who have had exposure to life outsidethe villagemight express a lower or higherpreference for the servicebeing offered dependingupon how it compareswith what they have seen elsewhere. Thus, the reactionto public taps or limitedserice hourscould well be negative. 3.29 Respondentswho have seen operationalpiped water systemsthemselves could be expectedto have a positiveattitude towards the provisionof a similarfacility in their own village. The attitudeof spondent in villageswith pipedwater could be affectedby their levelof satisfaction with the managementof the system.

3.30 Respondentsin favorof the meteringof waier suppliescould be taken to represent ose who are comfortablewith the notionthat a commoditylike water shouldbe chargedfor. They are, therefore,ikely to be willingto pay more for piped water than thosewho considerthe prvision of water to be an obligationof the stat

Program Report Seri 17 331 Ower Factors. Houholds have to bear the costs of conetn g their houes to the ditribuion line It I. thus reasonableto expectthat bouseholdsfurther away fromthe lae wouldbe less likly to conect to a piped system. Probablythe cost would be higherand becau beyond somecaitical distance the altemnativeof installinga privateimproved soure mightbecome economically more attaive.

332 Vaglevel cha i could be expectedto have someimpact on tho level of prefece expressedfor piped systems. For example,te distace of a vfllagefom the distict headqugrimcould be a proy for tho generallevel of affluenceand awarenessin the village.

3.33 Simiarly,in the arid zone, the proxmity of a vflbl to a perennialsouce of waer could be epeted to lower tho felt need for improvedsources comparedto other vilages. Such vilage-levelinfluences are testedfor in the arid zone.

3.34 Con ol mariables.Wherever it is felt that specificvillage characteristics could be affectng the results,the suppositonis tested by using a village identificationdummy. Similarly, diffrenes in the responsesof vaious typesof households(Al, AZ BI, B2) anetested by the use of householdidentification dummies.

3.35 Somebidding ames are used to test for the prsence of a satig point bias. Such a bias wil be mifested if there are systematicdifferences between the wlingess-to-pay bids of respodents who were randomly assigned a high ar a low sarti g point. A stg point dummy is used to test for the presece of such a bias.

Estlua Teebuijue

3.36 Ana(ysi of Actl ChoiceBehaio. The analysisof actualchoice behavior is based on the esmation of a logit model. The dependentvariable is binary, indicadnagthe choice or othe of *.parlcua servic level. The modelis used to assessthe impactof differentficos on the lkeHhoodof a household'smaig the paicular choice.

3.37 Anapsl of Jypodhecal Choice Behaior. The analysis of hypotheticalcdoice behavioris based on the estimationof an OrdinaryLeast Squaresregression model. Since a bidding gameis used to obtain the wilinges-to-pay estimates,the observeddepden varable is not the maximumamount the householdwould be wlling to pay,but ratheran intervalwithin which the "tueW wilums to pay falls. iear gresion is actally not an apprpriate techniquefor dealingwith such an ordinal dependentvaiable; in this situationthe coect appmrchis to use an odered probit model' However,the use of the mid-pointsof the intemls as a dependentvariable in an ornary

9Whitingto, Dale,John Bdsooe,XUming Mu, and Wim Bamo. Ssimatng the Willingness to Pay for WaterServices in DevelopingCountres: A CaseStudy of the Use of ContingentValuation Surveysin SouthemnHaiti" Eonoi De loWt and CQwl Change38:2 (January1990): 293- 312.

18 UNDP-WoarldBnk W and hSaiao PmV= least squaresmodel seems to yield esult which awecosiste withkthose obtainedfrom an ordered probitmodel, and the paametersare much asier to interpet

338 In this study the intrvas are quite small(eg, Rs 10-20,20-30, 3040, 40-50,50- 60, 60-70,70-100 in the arid zone). In addition,zero bids are cleady identified,and bids beyondthe end-pointsof the range(Rs 10 and Rs 100)are elicitedand recordedas actualvalues (the sweet wate zne is an excepton. See AppendixB-1). he use of actualvalues b d the end-pointsand mid- ponts wid the range shouldyield close enoughapproximations to th true"Wbids to makeordinary loastsquar esimatd an acceptablefEist level of analysis.

339 P Matatiof theReudts of dteMitaiate Asraysa. The resultsare presentedby wportg two models Ihe fist modelincludes all the relevantviables from the list presentedin Section3.9. In the analysisof hypotheticalchoice behavior step-wise deltion of variabls is used to idetify the estimationwith the highestadjusted Rsquawed. This the secondmodel reported. Such a prsentationhelps to ascetain the stabilityof the estimatedpuameters as insignificantvariables are deletedfrom the model.

3.40 Meanvalues of all the variablesused in individualrgressis are reported.Variables whichare statistcallysiflc.t at the 90 pementand 95 percentlevels are identifiedby one and two respectively.

fWhittington,Dale, Mark Mujwahuzi,Geard McMshon,and Kyeone ho Willin to Pay for Water in Newala DimtVw Tanaa Strategiesjbr CoSt Reowy. USAID Water and Saition for HealthPoect, Field ReportNo. 246, Washigto DC Jue 1989,pp. 205.

Prograw Rot Series 19 4. Sweet Water Zone: Overview and Field Procedures

4.1 Supura Distict is locaed in the cet canal-irrigatedpt of the Punjab.The districtcapital is ShelkhupuraCity with a populationof 141,168according to the 1981census. It is situated35 kilometetsfrom Lahore,the provincialcapital, and 105 kilometersfrom Faisalabad,the center of the textile industryand the third largest city in Paldst Agricultureis mechanized,and industry is located along both the principal axes, the Labore-Sheilhupuraroad and the Sheikhupum- Faisalabadroad. 4.2 Threefourths of the area in SheikhupuraDistrict lies in the swoetgroundwater zone, and water can be tapped at an averagedepth of 25 to 30 feet. Waterquality is almostunaivsally perceivedto be good.

Porqy iues

43 The currentpolicy of the I unjabPublic Health Engineering Department is that public piped distnbutionsystems are not to be built in the sweet water areas. hs policyis the result not of a demandbasedbut of a need-tasedrauionale. The PHEDhas beenentrusted withe responsbility of poviding acceptablequality water to the lagest umber of peoplepossible. Giventhat there are many areas withoutsuch access,the scarceresources of the PHEDare deployedaccordingly. Within such a frameworkthe low priorityaccorded to the sweetwater areas is understandable. 4.4 The above policy has also been supponed because the private sector is rightly consideredto be more efficientthan the publicsector. In his recommendationsto the WorldBnk Brie (1987)states that w(I]nthose areas whe government-fudedwater supply prgams are not undertaken(such as the largeareas of the Punjabin whichgood qualitygroundwater is available),the privatesecor has a majorrole. Individualfamilies contract private drillers for sinkinga well, and purchasea handpumpon the open manke ... Whilethere are probablysome imprvementswhich could be madein handpumpdesign, govemment policy in tis area - namelyto leave it to families and the privatesector to resolve- is appropiate.3

4.5 Thus,one of themajor objectives in includinga sweetwer zon was to i_VIStipte te extentto whichthe abovepolicy was justified. The other objectvesincluded those tha. formthe core of the researcheffoutr i.e., to ddemine thoewillingness of peopleto pay for impwvedservices.

Waer In the Swee Water Zone

4.6 Historically,wells were the prmay sourceof dnking water in the vrillagesfllidg in the sweet waterzone. Indeed,the centerpoint of a villagewas identifiedby a well,and morewer locatedat other convenientpoints, especialy as the villagegrew in size.

Pwran ReportSeies 21 4.7 Howeaer, as househlds became more affluent, wells were replaced by bandpumps inside the house. Now vitully evey household has installed a private handpump. Almost all the wells, except thoso in some mosqus, have been bdcked over. ITis process of improvementin the level of service took place between 15 and 20 years ago. Ihe excess of the privately bome costs of opation ane -9itenance of the handpumpsover whatever contrbution mast have been requird to keep the wells operationd prowvidesa baseline esimate of the value placed on the convenience of having a water source inside the house. Ms conveniece was almost entiely in the form of time and effort saved in fetching water from an outside soure Given the high summer temperatumeson the Punjab plains, the value of the convenienceis easily understood.

4.8 Our survey rvealed that a process of fiuther improvementin the level of sevice was under way. Perhaps as a result of incrased affluence, almost 20 percent of the households in the vilages surveyed had installed small electric motors ono the handpumps. These motors could pump water Into an elevated tank for distributionto various parts of the house through indoor plumbingand also could help to operate a flush toileL Thus the complete system associatedwith a piped supply was being replicated pivately at the indvidual level. This was an important finding because it revealed that householdswere willing to spend on improved service levels. Numerous economic implications resulted fiom the spread of this process The shal be discussed later.

Selecton of e Stud VMiaes

4.9 Since it is govenment policy not to insta piped supplies in sweet water areas, there are very few villages with public supplies. The choice of Type A villages (villageswith operational piped water systems)was thus vey tricted. We selected the only two vilages that lay within the maximum distance from the distict headquarter that we wanted to consider. Public supplies had been insalled in ese vilages as an exceptio to the geneal policy, most probably due to the influence of local politics.

4.10 Similarly it was not possible to locate a Type B1 village (village where a pubLic system was due to be inaled in te near future). We selected a village in which a piped housebold supply had been operational, but which, for various reasons, had fallen into disuse over fiv years sgo Curendy an effort was under way to restore the sstem. The selection of such a village offered the opportunityto see how poor performancein the past had affected the willingess to pay in the fure of those householdswhich had been comected to the system.

4.11 The location of Type B2 villages (villages with no plan to install a public water system) posed no problem, and we selected two convenientvillages close to the distrct headquarte.

4.12 he following are te partieul of the five selected villages.

* Jandiala Sher Khan: Type A Miza Vrka * Kharianwala, Type Bi * Ghazi Minua : Type B2 Bhadd troMiar

22 UND-WordBn WatradSntto ga While the populatons of the types A and B1 villages exceeded5000 by necessity, the two B2 villages were selected from the size categoy 2000 to 4000. The poplations of Glazi Min and Bhaddroo Minara were 4514 and 2173, respecively, in 1981. A more detailed profile of the selected vilages is povided In Table A-1 in Appedix A."

The and Bidibig Gams

4.13 The core quesdonnaireconsisted of five broad modules.The filst dealt with the socio- demographienature of the household.ITe second dealt with the household's water use practices. ITe third consisted of bidding gmes designed to eLcit the household's willness to pay for improved service levels. The fourth obtained infrmation on the household's attitude towards Issues conected with the provisio of ddinkg water. Th fifth module aimed to obtain information on household assets

4.14 To captre infonmationspecific to different types of households, five quesdonire shedules were used. AU the schedules contned the core questionaire in addition to the supplementay questions required. The key details of the various schedules are as follows:

Schedule Al * For households connected to an operational Type A water swpplysystem; -vilage Schedule A2 * For households voluntarily not connected to an operational water supply system;

Schedule Bll * For households pviously connected to an operation system (system currdy in- Type BI operatonal; resttion under conid- vilage eration);

Schedule B12 * For households not connected to the above system when it was operational;

Type B2 Schedule B2 * For households in a vilge without a piped village water systm and not expecing onw to be insled in the near future.

4.15 Some biddg games included a built-in test to determine whether the srting point of the bidding game affected the wfllngs-to-pay bids Ihis was achived by varying the startin point. For selected bidding games, half the questes contaned high stang pos while the ather half contained low starting points.

Stadstical tables and analyses generated by the suvey, grouped by subject, appear in the Appendixes.

P?qvigaRqxwt Seria 23 4.16 The followingbidding games were included in the questionnaires:

Ouestionnaire Biddiuf Games Startinn Point Schedule

No. Descrition

Al 1. WTP for an improved system High, Low (4 hours of additional supply). A2 2. WTP for a standard system High, Low with a more attrctive financng option (all connection charges to be bome by the water authority).'2 3. Same as bidding game number 1. High Bil 4. WTP for a standard system. Higi' B12 5. Same as bidding game number 4. Higb, Low 6. Same as bidding game number 2. High B2 7. Same as bidding game number 4. High & Same as bidding game number 2. High

4.17 The bidding games were desigaed to evaluate hree options:

• Mhe standard PHED system. * The standard system with a more atactive financing option (all connection costs bone by the water authorities). * The standard system with an additional 4 hous of water supply.

4.18 For analysis, the WI? bids from the following games could be pooled.

Bidding game Standard system Households(B11+B12+B2) numbers 4,5,7

Bidding game Standard system with Households (A2+B12+B2) numbers 26,8 financn option

Bidding game Improved standard Households(Al+A2) numbers 1,3 system

nExistngpolicy requires the household to bear ali the cost of connection from the distribution hle into the house.

24 UNDP-WorldBan Water and Sankdn Pogram 4.19 In addition to the above bidding games, evevy respondent was asked to descnibe the characteistics of his preferred water supply system.He was then asked via a direct question to indicate the maximum montblytariff he would be wg to pay for such a system.

Cor luciag the Surey

4.20 The survey was carried out In the five villages over a one-week period (Mlah 6-13, 1988), and 261 interviews were completed in the five study villages:

Jandiala Sher Khan : 99 : TypeA :144 Miza Virkan : 45

Kharan Wala : 58 : Type B: 58

obazi Minara : 28 : Type B2: 59 Bhaddroo Minaa : 31

421 Ihe number of interviews by schedule type was as folows:

Al : 79 A2 65 Bil 40 B12 : 18 B2 : 59

Tree visits were made to Jandiala Sher Khan, two to (hazi Minara, and one each to Bhaddroo Minara, Mm Virkan, and Kharianwala The interviews were conducted in Punjabi and the questionaires were administered to an adult mae, usually the head of the household. Some basic socio-economicchad cteristics of sample householdsand their attitudes towards water- related issues are presented in Tables A-2 and A-3.

ProgramRepor Seria 25 5. Findings in the Sweet Water Zone: Analysis of Actual Choice Behavior

5.1 Wateris fieely availableIn the sweetwater zone. The qualityof water from tradidonal sources(wells, handpumps) is consideredto be satisactory and as good as water from piped water systems.These two acts are important:water is not a scarcegood and there is litde qualitydiffetia betweenwater obtainedfrom taditionalsources and that obtainedfrom improved sources lTerefore, if houseolds are spendingmoney on higherseice levelsit is eitherfor considerationsof convenience, status, or some indiret benfits not reated to needs of water for human cnumption.

Upgra&dg Sevie LevelI

52 ThUtransition from wels as the pimary sourceof waserto handpumpsinside the house has long been complete. Householdshave inured the capital cst of between Rs 60 and Rs 1,000 in curret value for the couvence of not havingto go outsidethe house to fetch water.This i8 so even thouh in the sweetwater zone publicwells werecomn ently locatedand never very far fhom any household.Indeed, the village center was mated by a well in a majoity of vilages.

53 A second transitionis now in progess. Our survey shows that 20 prcent of the householdsin the studyarea haveisaed electricmotors onto the handpumps.T1h capitalexpenditure of betweenRs 1,000and Rs ,000 in curent pries, dependingon the qualityof the motor,obviates the need for manualpumpin. This s being followedby the i3stallationof ovhead tans (cast Rs 400 to Rs 500m,Idoor plumbing (cost: apaximately Rs 400 for a 20 foot network usully enougb to operate a bathroom with shower), and flush toiles in some cases.

5.4 At considerableexpense, households are puttingin placemodem indoor plumbing systems with their own privatesource of water.ITe monfthlymaintenance cost alone accordingto the survey is Rs 6 for the handpumpand Rs 16 for the motor. To this one must add the electricitycharges incured to operatethe motor.Ihe tol monthlyepenses are far in excessof the flat monthlytadff of Rs 10 for a houseconnecion to a publicpiped water supply system where it is available.However, evenin villageswhere the latteroption is available,11 percentof the householdshave instald electric motom. This could signl a demandfor more reliableservice. (Table A-4 shows the patternof householdchoice over availablesevic opto along with their approximatecosts.) WhoInstalls Ebctie Motsiu? 55 The statisical analysis(for deti see Table B-1) shows that the better educated, hie prpertied, and the more economicly affluenthouseholds have a significantlyhigher probability of Isaing electricmotors. Theseresults accordwell with the theoreticalexectation that bett off householdswould be morewiling to investin improvedand/or more reliablelevels of servie.

Progm Report Serks 27 5.6 Since handpumps quiremanual effort for drawingwater it could be hypothesizedthat overcomingthis inconveniewemight be a considerationin installingan electricmotor. However,the statisticalanalysis shows that neitherthe variablesrelated to the need for water (householdsize and per capitaconsumption) nor thoserelated to avaiabletabor supply (proportion of womenand children in the household)are significant.iLs suggeststhat upgradedservice levels ax the main motivadon for the illation of electic motors(ioor plumbingand flush toilet are possiblewith motor- operatedhandpumps but not with manualhandpumps).

5.7 Tis conclusionis importantbecause village elites have historicallyexerted a vey strong demonstrationeffect on the rest of the rral population. Theirchoices can thereforebe takenas a pointerto the potentialdemand for reliable,modem piped watersystems in rural areas. 5.8 In this contextthe analysishighlights another importantfact: in t*e subset of villages wherea publicpiped water scheme exists, households are significantlyless likelyto installan electric motor. Indeed,in such vilages only 11 percentof the householdshave electricmotors, whereas in villageswithout piped watersystems the correspondingfigure is 30 percent. Ihis indicatesthat house connectios to a pipedwater system are conideredto be, and havethe potentialof being,an attracive altemativefor consumers.

5.9 However,it was also a fact that 7 pecent of householdsthat had domestc connections also had electic motowmTis is a strong signal that public piped water systemshave to be of comparble reliabilityto competeeffectively with privatealtematives. Wo Cowucts to Piped WaterSystems?

5.10 In two of the samplevillages, Jandiala Sher Kban and MiczaViran, hcoseholdshave had the altemativeof connecingto a publicpiped water supplysystem with householdconnections. In the sample55 peroentof the householdshad availedthemselves of this alternative.

5.11 Thestatistical analysis (Table B-2) indicatesthat householdsthat needmore water (larger householdsize and higherdaily per capitaconsumption 1 ) are morelikely to connectto a pipedsystem. At the householdlevel there is no real limit to the amountof water that can be drawn up with a handpump. However,the physicaleffort requiredincreases in poportion. Iherefore, one could concludethat for householdswith a need for substantialamounts of water the conveence of access associatedwith a piped connecionis a factorin optingfor that choice.

5.12 It is also quite probablethat in areas with good qualitygroundwater households with handpumpshave not investedin storagedevices. Everyonedraws up water for his or her own use and withoutwater storageit is not possiblefor womenand childrento substitutetheir labor for that of the adult malesin the house. Perhapsthis is the reasonthat householdswith a higherproportion of womenand chldren are not less likelyto connectto pipedwater systems (note that the respondents

I' he data suggest that essentialdaily per capitaconsumption of water is not an endogenous variable,i.e., it is not higher for householdsbecause they have connectedto piped systems. Table A-3 indicatesthat the valueof the varable is actully higherin villageswithout piped watersupplies.

28 UNDP-WorldBapk Watr and Swhaiion Pmran weoe all adult males). On t cony, householdswitb a higher proportionof children in the howeholdane more likely to connectto a piped system.

5.13 The othae importantcondusion suggested by the statistica anysi is that wedth is not quite as sigdflcant ademlnt of fthedecisio to onmet to a plped system as it is for the Installationof electricmotors. is , perhap, becaue the fonner is a much less expensivechoice (the capitaland monthlyoperation and maintenancecosts are Rs SO and Rs 10 as againstRs 3,S00 and Rs 22 for an electricmotor - see TableA-4) and thereforegenerally affordable in an area where the monthlyhousehold income was of the order of Rs ZOOOin 1964-1985.

5.14 However,cost considerationsdo enter into the deciSionto connect. 'Me variable representingdistance of the housefrom the distributionline is negativea64 significaLt. The costs of connectionthat have to be borne by the householdare directlyrelated to this distance(approximately Rs 10 per runningfoot). The increasein cost withdistance is likelyto act as a detert to connecion not only becauseof the expensebut also, perhaps,because beyond a certainpoint the cost becomes comparableto that of an electricmotor. The latter is lIkeLyto be a preferredalternative in such situadtonsbecause of its gmeaterreliability. The issueof reliabilityis importantand shall be discused agin later.

5.15 The analysisalso shows that householdsengaged in farmingare less likely to connect to piped systems while householdswith better educationare more likelyto do so. Householdswhich have electricmotors are not significantlyless likelyto connectto the pipedsupply system.

5.16 TMereasons mentioned by the respondentsfor eitherconnecting or not connectingto the piped water supplysystem are listedi TablesA-S and A-6. TIese were obtned as responsewto open-endedquestions.

Electri Motorand UEhodmMCwnwctou AlternatIeCboices

5.17 Ithe ft that househodsare significantlyless likelyto have electricmotors in villages wherea pipedwater supplyis availablecleady suggeststhat the two are consideredto be alternative improvementsin servicelevel.

5.18 At the exdstig, subsidized,rates, a domesticconnection to a piped supplysystem is a much cheaperoption. Both the averageprivately bome capitalcosts and the monthlyoperation and maintenancecosts are less than half for a domesticconnection compared to a motor-operatedpump. As a result;many more householdscan afford to upgradeto a higherservice level in villageswhere piped water suppilesare available(59 percentas aginst 30 percentin villageswithout piped supplies - see Table A-4).

5.19 It is also of interestto note the relationshipbetween the choiceof servicelevel and the economic status of the household in vilages where a piped supply option is avadable (Table A-7) Motorsare installedby householdsat the upper end of the economicscale. On the other hand, domesticconnecions are affordableto householdsonly slightly above ite averagevalue of the economicindicator used. The mast afiluent householdshave both electric motors and piped

ProgramReport Seies 29 connections.Given the poor performanceof pubHcwater supply systems, this probablyreflects the wllgns to pay for reliableback-up service by hoseholds that can afford the expense.

5.20 If the exsting subsidyon the capitalcosts of piped sysems (approximatelyRs 300 per capita with a householdsize betwe 7 and 10) is removed, the capital Costs of the two options becomecompuarble wbile the opetationand maintenancecost of the pipd syst remainlower. However,in such a situationthe much greater reliablty of the privateoption would make it a dominantchoice, andconnecton rates to publicpiped systems would be likely to fall very steeply.

30 UNDP-WorldBank Water and Sanian Pngru 6. Findings in the Sweet Water Zone: Analysis of Willingness-to-PayBids

alytU BehvW 6.1 In te prios capter the _aysis was basedon the acual, obsevable choicesmade by Isntis chapterth aalysis b basedon the spoasesproffered to hypotheticaloptions preseatedto the respondents._ T sponse yield the maximummonhy taiff that a householdis wiling to pay for the putcular opdon described.

6.2 To the extn tat hohold utand al the changes that will sut fom the acquition of the opo presentd, te amount they say they will pay, their bid, can sere as a mesur of its benefitto thm WIP bids my ncludevaluation of aesthetic, healb, and oter difcult- -meau benefitsof wate. If WTP bids e an accuratemeasure of iiduals' preres, the summatio of the WIP bids for all householdsserved by a projectcould semve as an estimateof the tota beefits of the prjecL It can be compaed with the cost of th projectto decide whetherthe Investmet is jused.

Servic Opt Offeted

6.3 I11 folowingsere pons were p ntd to differentgroups of respondets

3 A hoseold concto to a sadad pipedwater system of the typeexisting in type A ville; * -e same, with a mom affdble financingarngement and - A hoold concio to an improvedpiped water system. 6.4 I the Punjabthe sandad systemis designe to providea maximumof 15 aOns of water per capitaper day with the senice availablefor eight hous a day at the most However,actual perform is very uv (40 peret of the connectedhouseholds exprsse dissatisacdon with the system)with problems of reliability, suficient pessue, sad srice for less than the specifiedperiod (See TablesA8, A-9, A-1) For this reasn, all householdsbmiliar with existingsystems do not perceiveh samepwduct when evaluating te 'stadad' system,espe y becauseprssure vaaotins oocr almostfiom ae to bla 6.5 Mmbopt withan alteratie financig arrangementwas offeredto detemine if the low conncton rate (5 peet of boueoldb) to piped water system was related to the stuctre of the costs involved. Obtag a connecton requires two types of payments: a one-time payment made up of an offici coanecon fee and the costs of connectg the bose to the distributo Hle, which have to be bomreby the houseold; and a flat monthlytri Mm dominantcomponet is the cast of connectingthe boe to the dbua le The co_nnectiofae is of the oraer of Rs 100, the connectioncosts are of the order of Rs 500, and the mothly tuiff is Rs 10.

prt,5TU ReporSeries 31 6.6 Cshb flow problems(i.e., the inabilityto bear the one-timecosts) may be acting as a deteret to conmectionfor some households.To test this hypothesis,an optionwas offeredin which the water authoritieswould betr the connecdoncosts in exchage for a highermontldy tarf The objectivewas to test if comnectionratios would go up with the avaUabiltyof such a fcing arnement and the extentto whichtariffs could be raised.

6.7 The impmvedsystem was offeredto householdsIn villageawhere a systemwas already operadonal.The improvementoffered was an increaseof four hous in the daily supplyperiod and the willingnessto pay for the imprved systemwas elicited.

Responseof HousehokdIn Vbges wthout an Operatin Water System 6.8 Wilngness to Pay for Connetion to a Stndard syste At presentno householdin either T¶ypeB1 or B2 viUagehas the optionof connectingto a standardpiped water system. Ihe differencebetween B1 and B2 viUagesis that there are no pls to instll a systemin the latter category. TheBE villagehad a systemin the past whichhas beenout of operationfor overfive yewas and for whoserestoration efforts are now underway. Withinthe Bl villagewe can distinguishthose householdsthat had obtaineda connectionwhen the systemwas operational(Bl1) and thosethat had chosento remainunconnected (B12). When the standardsystem option was offeredin the Bi village she WTPbids were conditionalon the desgned performancelevel being guaranteed.This was made necessaryby the extent of dissatisfactionwith the performanceof the standardpiped systemwhen it was operational. 6.9 The distnrbutionof the WIP bids is shown in Table C-1. In calculatingthe connection ratio we can assumethat all householdswhose WTP bids are less thanRs 10 wouldnot connectsince the existingtariff rate of Rs 10 per month is not likely to be reduced. We also treat the No Responseeas prtest bids and add themto the groupof householdsnot likelyto connectto the service offered. 6.10 At the existdngtadff of Rs 10 per month,the percentageof connectionfrequencies for the BlI, B12,and E2 householdswould be 70,833, and 74.6,respectively. The meanbids (computed over the valid responses)for the same groupsare Rs 17, Rs 17, and Rs 21, respecively. The mean bids of the subsetof householdslikely to connectare Rs 21, Rs 19, and Rs 25, respectively.

6.11 A numberof observationscan be madebased on the aboveesimates: * Thirtypercent of the Bl1 households(all of which wereconnected when the piped systemwas operational)would not reconnecteven if the performanceof the system were guaranteed.This reflectsa loss of cdbility in the system. * Connectionratios resulting at the existingtadff rate are quite high (74 percentin Dl and 75 percentin 32).These compare favorably with the actualconnection ftequency of 69 pecent;,which obied in the BEvMage when its pipedsystem was operationaL This corespondencesupports the walidityof the WIP methodology.

32 UNDP-WorldBant Waterand Santation Program * The meanbid of the householdslikely to connectin the Bl vitlageis approximately Rs 20 per month. Only 4 out of the 57 rsponses were for bids over Rs 25 per month. Tbis i reassuring.The monthlymaintenance cost of a motor-operatedpump is Rs 25, and becauseof its greaterreliability, it shouldprovide an upperbound on the monthlytariff of a pipedwater supply. (The one-timecosts of the motor-driven pump are, however,higher. See TableA-4.) * The meanbids are appoximatelyRs 4 per monthhigher in the B2 village.This could be due to overbiddingin B2, underbiddingin Bl, or both. However,overbidding seems the more likelyexpanation since nearly19 percentof respondentsbid Rs 50 or more in B2 compsted to 5 pecent in Bl. Such high bids could be intepreted as evidenceof stategic behavior.

6.12 Type of Houseold LAely to Connect to a Standd System. A multvaiate analysis of the WTPbids presentedin Table B-3 indicatesthat youngr, more educatedand properiedhouseholds are more likelyto connect.Farming households are less likely to connectcompared with nonfarming households. Householdsthat have a motor-opertedhandpump system are willingto pay Rs 7 more than othes Tiis could be because,compard to their expenseson the motor system, a piped onnecion would be a cheaper alternativeeven at the higher end of the feasible tariff range. Householdsthat favorthe meteringof domesticconections are also willingto pay more, perhapsa rlecftin of theirappreciation of wateras a valuableresource. (The complete results of the multivariate ansalysisare presentedin Table B-3.)

6.13 Wiingness to Pay for a Connecdonto a Stndard System with a More Affordable Finaning Arrangement.Types B12 and B2 householdsthat had neverpreviously connected to a piped systemwere offeed a more attactive financingoption (one-time connection costs to be borneby the water authorities)to test if affrdaility was a factor in the decisionto remainunconnected. The distibutionof the WTPbids is shownin Table C-2. Tese can be comparedto the WTPbids for the standardoption presented in Table C-1.

6.14 Table C-1 indicatesthat 17 out of 76 households(B12+B2) would not connectwhen offeredthe standardoption (WIP bids less than Rs 10 per month). The comparablenumbers from Table C-2 are 17 out of 74. It is clearthat the moreaffordable financial arrangement does not result in inceasingthe connectionratio. However,the meanbid over the B12 and B2 householdsincreases by approximatelyRs 3 per month. is suggestssome preferencefor the moreaffordable anangement amongeven the householdsthat wouldotherwise connect to the standardsystem.

Responseof HouseholdsIn Viaps with an Operatial Piped Water Sysem

6.15 Wdlingnessto Pay for Connecon to an Imprvd System Type A villages are characterI by the existenceof an opeationd piped water supply systemwhich is workingbelow its design specfications.Type A viges includetwo types of households:Al, whichhave domestic comections;and A2, which have voluntarilyremained unconected to the system.

6.16 All householdsin the Type A villageswere offerd the choice of connectig to an improvedsystem, the improvementbeing an additionaldaily supply of 4 hours of watr from the

ProgramReport Series 33 existingsysem The distributionof the Wr? bids is shownin TableC>3. A numberof obertion can be madebased on the figus

* The numberof "No Reponse or protet bids is muchhighe han TypesB1 and B2 villages(over 10 prct as comparedto a maximumof 2.5 pecent). 1s may be a reflectionof the lack of credibilitythat houeholdsfamii withthe peforance of piped systemsplc in any prmim of impovemet in the exitg systm. * Ihe proporion of houseioldsbidding more than the existng ariff of Rs 10 per monthis about 50 prcent (51 percentfor Al and 45 pert for A2 households) Thus,only half the householdsthat ae curently connectedto the systemae wing to pay more for an improvedsystem, while half the households that am curntly unconnectedwould connectto an improvedsystem * The meanWVW bids are low, Rs 14 for Al householdsand Rs 12 for A2 households. If it is assumedthat none of the Al householdsconnected to the standad system woulddisconnect if the serviceIs imprved x.e.,the potest bidsare ignored)and wil contine payingthe asing taiff of Rs 10 per month,thei mean WIPbid would dse to Rs 15 per month. * The mean Wi? bids computedover the subset of housholdiHely to connectare Rs 20 and Rs 19 for the Al and A2 households,respectvely.

6.17 Typeof HousehoWsLAe to Connectto an ImprovedSystem. The multivauatesatisicd analysisof the WTPbids (for deils see TableB-4) indicatesthat wealdtierhouseholds are wlling to pay more for an improvedsystem. Householdsthat favormetered connecons ae willingto pay Rs 6 more tha othets. Householdswith motorsagain bid Rs 5 per month more than thosewithout motom.

6.18 Householdswith a higherproportion of womenand childrenare wlling to pay more for an ineased supplyof water. Thissuggs that labor supplyconside an not relevat in aeas wherewater does not have to be fetchedfrom outside the house.

6.19 A vey interestig aspectrevealed by the multivaiateanalysis is the comparave bdeavior of faumingand nonfarminghousebolds. In vilages withoutpiped water the formerbid Rs 4 per month ess for a comnectionto a piped water system. However,in villageswhere a pipedsptem has been in operation,farming households bid Rs 4 per month more than nonfarminghouseholds for an ineased supplyof water.Ths could be attributedto leaned bdeaviorwhereby fuming households have come to recognie some previously(prior to the instlation of a piped system)unforeseen advantas of domesticconecdons. Our field obserationsrvealed that householdswith aimas wer vey keen on a domesti connecto becauseit madethe task of washingthem much more convient. Buffalo whichneed to be keptcool, cold be hoseddown at home, aher han be ta to tho cn or the vila pond,tu savingon tim and squvio coss FPer, water for dnkig by anals coud also be povided tOugh the domestic connection. Although this hypotis w not tesed sttstcaly, as we did n collectinfrmatio on animalownershlp, if faring houehols own more animls than nonrming houseods the expcna wouldbe quite plausible

34 UNDP-WorldBak Water ad Sani'_n Progra 6.20 If the explanationis true, it would also have sgificant impication for the dedsg of rural wat supply system Me PHED design cuiteir at prsent take into account human needs only and use a consumptionfu of 15 gallonsper capitaper day. However,if householdsare actuly sin the water to caer to animalneeds also, includingwater-Intesive ones likmwas the desig estiate coud be easilyexceeded. Perap this sug_sts one explanation for the ubiquitous poble of low prsu which plgues ual wate supply systm In PMt 621 Wilhgm to Pay for a Coecdon to a Stdvrd Syte vAh a More Aordable FinwwhsgOptin To expe heweasons for the nouconnectionof A2 housolds to the existng system, ty were offered the same financig option mentioned elier in the case of type B villages. Tbe dibton of WIP bids is included in Table C-3. The data suggest that the aiabilty of the flexible financing option in tie type A villages would rwaiseth, onnection ratio from 55 petrent (79/144)to 79 peet (113114A)if the existig tariff continuesto be chawed. Appximtdy half the previouly unonnctd A2 househlds wouldconnect Their meanbid for the monthlytarf is ls 21.

622 Howeve, TableC-3 asboshows that if the stadd systemis improved(an extra 4 houm wer supply per day ovet te existingsupply period)She same resullscould be obtained. Sligtly less than half the A2 houeolds wouldconect at the existn ta Ther meanbid for the monthy iff is Rs 19.

623 While it is possiblethat there may be very litde overap betwe the householdstht decide to connectin the two cases (poorerhouseholds might be the ones that connectunder the flxble financing aptn while households dissatisfiedwith system pefom emigt be the oe to coinn to the imprved system) the net result fiom the point of view of connecton rates is quite similar. Therfre, if the objecdve was solely to achieve high connection rates at minimum cost any choic betw the two option woud need to be basedon comparive costs.

PW RepoAtSwim 35 7. Estimated Revenues and Cost Recovery Potential

7.1 The wllWess-topay bids can be usedto estimate the likelihood of connecdon to nd the m ues geated fom the pvsion of various upgraded services. Such a computation helps determine whether the provisionof such services would be economicallysustainable.

Pa1ision of a Stkdard Piped Water Systm In Villages Falr with Such Systems

72 The Type BI village is particularlyattracive frm the analytical point of view. Type B1 households have had first-hand experience with a piped water system but do not have access to the same at present. Therefore, their willingness-to-paybids are for a commodity with which they awe quite familiar and for that reason could be expected to be more reliable than if the commodity had been a hypotheticalone.

73 CanwecdonFrequces. The connectionfrequencies and revenue estimates are shown in Table C-4 and are plotted in Figure 7.1. At the existing monthly tadiff of Rs 10 the connection frequency would be 84 permet, if the sevice level wv juaranteed. This is to be comparedwith the actual connection frequencyof 69 percent (at Rs 10 per month), which prevafledwhen the piped water systemwas opeatnl over five yeawsa. There are numerousindications that the system was poorly manged and it eventually fell into disuse. The impact of that history is demonstrted by the fact tbat, even with a performance guatee, only 83 perceat of the households that had previously been connected indicated a willingness to restore tbeir connectionsat the previous tariff rate of Rs 10 per month.

7.4 The plot of connectionfrquency against monthlytariff rates lends credibility to the bids elicited thrugh the willingness-to-paymethod. Tree reference marers could be used to interet the plot. Below Rs 7.5 per month the connecton frequency is very high (95 percent and above). This is as it should be since Rs 7.5 is the approximatemonthly expenditore at the minimumacceptable service level, the manually operated prnvate handpump. At Rs 10 per month the connection frequency is approximately84 percent which compares favorably with the actual frequency (69 percent) which pevied at that rate. As mentioned earlier, the increase in comecdon frequency could be attributed to the fact that a performanceguarantee was part of the package offered to respondents. Increased economic grwth during the past five years could also be a contributingfactor. At Rs 22.5 per month the connection frequency drops to 21 percent. Thi corresponds well with the response of the 26 percent households that have installed electric motors in the Type BI village; monthly expenditureon electric motor-operatedsystems is around Rs 25 per month. At Rs 27.5 the connection frequency becomes negligible at 7 percent, indicating that the ectric motor is the preferred option at this tariff rate.

7.5 Monty Rewue. The plot of revenue against montblytariff (Figure 7.1) indicates that revenues would be mximazed at a tariff rate of Rs 17.5 per month Revenue yield would be Rs 1026

ProgramReport Series 37 1,250

25 2r0

da0 Revesonutehla81e 0 05 10 if 20 25 3 5 40 45 50

F~~~~uws d andyrvemads (pe h0eusslioli): PtiwMaaof a stmmiuPW wa sr pply systm In a type Bt age per 100hmchd of wih59 percent b cnnctd At ariff rate of Rs 125Sper month die figus wouldbe Rs 926 and 74 percn (fhes fiues are to be comparedI situationwbich existed when the pipedsystem was opetona whenthe connectionfequency 69 peret an the reenue Rs 690 per 100 householdsat a taiff rate of Rs 10 per month) T)eSfore, any tarff in the tinge of Rs 12.5 to Rs 175 per month shud achieve the dua objectves of a reasonablyhigh cnnetio frequencyaud highcost recoveq. 7.6 Op_wdonamd Mai ane Cost. To whatextet cantho nge of revenuesmentiond abovepay for the opeton and maintenancecoss of a pipedwater systm in the sweet war ZOn? An examinaon of the a cost data used by the PHED for the scme isalled in the Type Bl vlg would be usf in ths coteL

7.7 ITe scheme was iidated dung 1973-1974and competed In 1976.Ihe systemwas dediged for a popdation of 6800 at 1S gallons per capita per day and cost Rs 333,080 at adprximatelyRs 50 per capita The annualoperadon and maintenan costs were ompted to be Rs 10,S16as fows (anmnalO&M costs wok out to be 3.2 pent of capitalcosts, at the low end of the 3-5 peroet rag uosedby the PHED):

38 WWDP-WolBak Water and Saio Pmora Personnel(Operator, PlumberValveman, Rs 4930 (47 perent) Watchman) OperationCost (Elecical & Mechanical)Rs 2717 (26 peret) AnnualRepair Rs 351 (3 percent) Contingenies(at 5 percentof b+c) Rs 153 (1 percent) Dwercation Rs 2365 (23 percet)

Total: Rs 10,516(100 percent)

7.8 Usingthe figre for householdsize mentionedin the 1981 Census(6.7), the O&Moost (Includig deprecaon) worksout to Rs 1 per householdper monthat the 100 percentconnection rate assomedby the PHED.Even if a connectionfrequency of 50 peret is assumed,a tariff rate of Rs 2 per monthwud meetO&M costs The tariffrate was set at Rs 6 per householdper monthin 1976. 7.9 Suice 1976,costs have escalatedrapidly. The capitalcosts per capitacurrently used for lhe desg of tubewellsystems are Rs 300. CaculatingO&M costs at the upper-endrate of 5 perent of capia costs fr a populatinaof 10,000(the appmate populationof the Type B1 illage at present)would yield a figureof Rs 125 per householdper mouth(the householdsize at presentis 1), mumng uivm coverage.If the connectionfrequecy is 75 perwe the tadff rate that wold enmuefull recover of O&Mcosts wouldrise to Rs 16.7 per householdper month.

7.10 Ihe esdmatesobtained above are upperbounds. A look at the brakdown of O&Mcasts enumeratedabove shows that only 26 percentof the costsare due to the actualrunning of the system. he t are fixed costs of which the largest proporton is due to peromel costs (47 percent). Theefoe, with increasngpopulation te O&Mcosts per householdshould decrease. If we compute anual O&Mcosts at the lowervWalue of 3 pewcentof capitalcosts, the corr dng tariff rate for fuln reovey of O&M costs at a 75 perent connectionfrequency would be Rs 10 per householdper month. 7.11 Rs 10 per householdper monthis the exisng tariff rate in the sweet water zone. In gera, it seems the tariff rates imposedby the PHED,which range from Re 10 to Rs 25 per monthin the Punjab,are calculatedon the basis of recovemyof O&Mcosts. In the light of the above, th tarff rates, connecdonfrequenes, and revenueestmates revealedby the WT? analysisfor the Type Bi vilage suggestthat full recoveryof operadonand maintenancecosts is possible.

Psiwkb of a S_dad !pkd WaterSystem In Vilges Unfaia with Such Sysems

7.12 Results from the Type B2 villages, where the bidding game could be cosidered hypoteical ae easonablysimilar. Table C-S and Fure 7.2 show the connectionfrequies and estimd rves at vaious possibletaiff rtes. one significantdifference in compaion to the Type Bi vMageis the high connecdonfrequeies (around20 percent)at tariff rates beyondRs 27.5 per montdL is is due to the disprpotonate numberof highbids indicatedin Type B2 villagesas compad to both the Type 31 and the Type A villages. These could be due to the pence of staeic bias in the bidding either overidding in the type B2 vilages or, less likely, uerbdg in the other two types of viages.

ProgramRepor Sria 39 1,000

I 00:_ 750

]75 _- /\\\\ \ 7

\ Connections 500

250 25--

0 5 10 15 20 25 30 35 40 45 50 Monthlytariff (Rs)

FIge 7.2 Conuetlan frequsuclesand mont revene (per 100 householks):Provisn of a stadard piped water supply systm In a type B2 wiage

7.13 Revenueswould be maximizedat a tariff of Rs 12.5per houehold per monthyielding a rvenueof Rs 933 per 100 houeholdsof which75 percentwould be connectedto the svstem(Figure 7.2). ITis figurecompares favorably with the taiff requiredfor full recoveryof O&Mcosts. Assuming a villagepopulation of 5000,a householdsize of 9, Rs 300 per capitacapital costs, 3 percentof capital costs as O&Mcosts per annumand a connectionfrequency of 75 petcent,the tariff requiredfor full recoveryof O&Mcosts works out to be Rs 9 per householdper month.

Provi !n of an ImpmovedSystem In VMlagswih an Exdsng Piped Water System 7.14 Type A vilages, wherea pipedwater supplywas in operation,were offeredthe choice of an additional4 hours of water per day from the existingsystem. Table C-6 and Figue 73 show the connectionfrequencies and revenuesfrom various tariff rates, as revealedby the WTPbids.

7.15 The reenues wouldbe maximizedat a tariff of Rs 17.5 per monthwith a yield of Rs 693 per 100 householdsof which 40 percent would be connected.At a tariff of Rs 12.5 the correspondingfigures would be Rs 599 and 48 perct (the existingrevenue potential is Rs 550 per 100 households,the connectionfrequency being 55 percentat a taiff of Rs 10 per month).

40 CJNP-World Ba* Water ad Saiatio Poram 750

100

500

£50 250

25 Connections \

0 5 10 is 20 25 30 35 40 4S 50 Monthlytariff (Rs) Fue 73 Cectiou, Freqnces and mot revenu (per 100 households): PIrovison of aS Imprved piped water su*ppl sysem in a type A viue

7.16 Ihe connectionfrequencies, and thereforethe rvue yields, are lower than at equivalent tariffs in the Types Bl and B2 vllages&This pObably reflecs dissatisfactionwith the performanceof the existing system and the fact that an extra 4 houwsof water from a poorly rA system suffering from low pressure in the pipes is not very atrave to the respondents.This suggests that the emphasis ought to be on improvingsystem performanCe.

EpeIments with Atwnative Finnig for Household ConnCto

7.17 The alternaive finacing option was descibed eadier. Table C-5 presentsthe comparative connecdon frequenciesand reveaue esimates if either the stadard opion or the flexNbleoption alone are offered to respondentsin a Type B2 vilage.

7.18 Suppose the PHED seeks to recover the extra capital expenditureof Rs 500 per household incufred under the flexible financingoption over a period of 5 years at an interest rate of 10 percent The additional monthly payment would amount to Rs 10.5. Ihus the effective tariff would be approximatelyRs 21 per month.

ProgramRepo Series 41 7.19 he two optioucan nowbe compared. A the oaxltg tff of Rs 10,85 p of the householdswould connect to the stan6rd system yielding a montlyf r Bueof Rs 850 per 1000 households.At a taiff of Rs 21 per monthunder the flexibleopdon, 54 perceutof the houslds wouldconect yieldinga monthlyrevenue of RS 1134por 100 households Ilus, thermue yield wouldimpove but the con1ectionfrequency would decine. Te objecdveof the poicy would not be served.

7.20 A varant of the policy would be to offer the fexgiblefinancing optio only to th householdsthat havenot connaectedto the systemat the existng iff rate. Thecontion &quenc and revenueestimates resulting fom such an offer to A2 hosehold in a Type A ve are shown In Table C-7. At a taiff of Rs 21 per montb, 24 peent ofth oushlds would coaect tote system aiing the overl connectionrate to 66 pert (55 peet huehol coanectedat the existng tariff plus 24 percentof unconnectedhouseholds conneing uder e fleaibe financing anrangement).Thus, under the cost recovey conditionsstipulated, the poicy of oeng both the standad and the flexibleoptio in the same villagewould succeed in aising the connecdon ate to some extent. 7.21 The monty revenueyield per 100 householdsunder the abovescenado woud amount to Rs 781 (ts 10 x SS + Rs 21 x 11). I can be compe to the evenue yild f theexlstg systemis impoved in the Type A village. Table C-6 shows the connectionfrequencies and the rvenue estimates Whilethe exact costs of prviding an additional4 hour of waterper day ar not known,suppose that the tariff would have to be raisedto Rs 17.5 per mor b. At this tadff the coneton frequencywould be 40 percentand the monfflyrevenue yield per 100 eholdswould be Rs 700. Thus,the connectionrate wouldbe considemblylowe and the mothly eue would be onlymargnly less compaed to the flexiblefinancing option.

42 UNDP-WorldBoe Waterand Satatim Pgr.p 8. Brackish Water Zone: Overview and Field Procedures

8.1 FaisalabadDistrct is alo loct in te co l cnIrrigUed par of the Punjab.Ihe districtcapital* Faiabad Mty, had a populationof 1,104,209accoilng to the 1981 cenus. It is situated145 kIlometess from Iahor, the capitalof tho PUmjab,and is the thirdlargest city in Pakistan. Fslabad is an Intenationallyrecognized center of the teffile industry. 8.2 Almt half the area in Fasalsbd Disutictlies in the brackishgroundwater zone. Water is availableat accssble deptbs but its qualityis generallyperceived to be poor, althoughthere are vations, often witin vllages 83 Villagesin whatis now the bckish waterzone formerlyrelied on wellsas theirprimary soue of water.In keepingwith he tansition in the sweet water zone almostevey householdhas now replacodthat sourcewith a pdvate handpumpinside the house.

4 However,because of deelopments attributedto water-loggingand salinity,the quality of ounmdwathas deteioratedovertdme so that by now dissatision with its taste and its impact on healthis fairlywidespread. Of the householdswithout accs to piped conntions, 54 peroentrely on sourcesother than thei pive handpumpsfor water used for ddarnkgand cooking Amongthese alterative souce an canals and public handpmps alongsidewater-courses channeling water frm cna to agrculturallands where the seepagewater is of betterquality. Water from such sourcesis either fethed by hosehold memberswho devoteappoximately 35 minutesper day to this task or deliered for a charWeby water caiers. 85 In such a situationpiped supplysystems provide the only conenient source of good qualitywater. Neither privatohandpumps nor motoized pumpscan providewater of similarquality. Terefore, it is not surprisingthat connectionfrequeciwes to piped water supplysystems are on the high side, awund 75 percent

8.6 What is surpising is the high percentageof samplehouseholds (50 percnt) that have ntheless inted motorizedhanbpmps. Tis is patly attibutable to the fact tbat the motorized pump can yield somwbat better quaity water than the handpumpby goingdeeper into the subsoiL It alo unde_ores the demandfor water-laed convenienceslike indoorplumbing, ush toilets,and showess,which are not acoessiblewith handpumps. 8.7 Te abovefac suggestthat omesticconnections to a reliablepiped watersupply system shoud be prefered to motoized pumpsbecause they can prvwidea better qualityof water fit for dindng and cookingin additionto the water-relatedconveniences mentioned eulier. Besides,domestc conecdons ae a less expensivechoice for the consumer.

Poram ReportSeries 43 S8b0km of Sft* VW""

88 The selction of TypeA vlag was randerd difficultby the need to ensurea rasotle poportion of householdsthat d remined unconnectedto a piped suWly system by choice; the conmtecionfrequencies wero generallyon the bigh side. Two vfge Manawalaand Sudhar,were selectedwher the conctdon frequenciesweae aund 75 pcnt. Manawala,a large villagenow incopo withi he limits of the Faisalabadmunicipaty, has a piped supplysystem more than 10 yeas old. Sudha, on the othe had, has a piped supplysystm which was only commissioned in Januar 1988.

8.9 Selectionof TypesB1 and B2 villageswas slghtf ¢ Therear manyvilt where piped suppies an scbeduledto be installed.The selected illaes of the type, Bbaiwalaand Akhlga, are at the sae wherethe disiution malu havebeen laid but domesticcomnctons have yet to be sanioned. The systemin Bbaiwalab basedon a tubewell,while that in Akalgarhis based on filered canalwater. The two Typ B2vile are both lcated close to the disrict headquarters. 810 The owingare the paicula of the six selectedvilgs:

• bfiaewala :TypeBA * Bhalwala Type.El

Santpuwm Type B2 Gatwala

All the scted villagesexceeded a popu size of 5000. A more detailedprofie of the select vla is providedin Table D-I."

eso te Qos_,r and B8ding Gam

811 The quionnaires are the same as used in the sweet water zone with only minor modificatonsdue to siwseci ndens and the exeenw gained in the is ound of interiewing. Mm one majdrd is that there is only a singloschedule pertining to the Type Bi viae since, nie the weet wate zone, there was no need to ditnguih among householdsin the Bl vmage

"4ta6tic tablesand analysesgenerated by the survey, grouped by subject, appear in Appendixes D, A and F.

44 UNDP-Wold B* Watgr ad Sa&iaion Prgram &12 The biddig games wern desied to evsluate the following opidons:

The sandard PHED piped water supply system of th kind exiting in Tye A villages. i option was offered to all households. T WTI of already connected huseholds (Al) w icitedby bidding up fram th exisng tarff. Ihe WI? of the vluitaly unconnetd bouseholds (A2) was elcitod by bidding up from a tawiff rate of Rs 0 per montL * An impved piped water supply system which would supply clean and safe water continuouslywith adequate pressure and rliability. Tis opon was offeredto eveay household. * The standard PhD system with altemradvefa arngm Only A2 households wore offered the folowing two alternatives: (a) 50 poe t of the conanon costs to be bore by the water authorities; and (b) 100 peroent of te connecton costs to be bone by the water authoriies.

8.13 Ihe design of the bidding games ensured tht potest bids and genuine zaro bids would be clearly identified. When a respondent indicated that he would not be willing to connect to the piped water sytem at the lowest monthly tarff included in the bidding game, he was asked whether he would connect if the manthly tariff was zero (Le., the water was free). If th respondentindicated that he would stll choose not to connec, the eumator asked him to explain the reason(s) why. The lowkad hh stng pointswed weres 20 andRs 40 pe moth as compued to Rs 15 (or Rs S in some cases) and Rs 50 in the sweet water zome.In the a a of the startin points households eitbhr had low stag point or high st poits for both the games offred to them (for existing and impoved systems).

Condctg the Smrvey

814 Ihe survey was caried out in the 6 vilages over a 1O.daypedod (August 18-28, 1988). Some basic socioeconomic chatist of sample households and their attitdes towards water- rdated issues are prseted in Tables D-2 and D-3. A total of 495 interviews were compileted,as follows:

Manawala : 106 : Type A: 202 Sudha : 96 :

Akalgh : 106 Tpe B1: 200 Bhaiwala : 94 :

Sanwtpum : 52 : Type B2: 93 Gatwala : 41

Pogrma ReportSer 45 The numberof interviewsby scheduletype was as folows:

Al: 151 A2: 51 BD: 200 B2: 93

Two visitseach were madeto Manawal, Sudbar,Ak blh and Bhaiwaaand one each to Santpu and Gatwaa.

46 UNDP-World BW* r and Santat Prqram 9. Findings in ihe Brackish Water Zone: Analysis of Actual Choice Behavior

9.1 WaW is fiely availble in the bdackishwater zone. Thequality of water fromtraditional sources(well hmndpmps)Is generallyconsidered to be poor for drinkng and cookingexcept from handpmps slond water chanels which pick up seepage water from canals.

92 The tUnsito fm well to handpumpsas the pmay source of water is complete. Most householdshave privte haodpumpsinside the houseto cat to their water needs. However, for dinkin ad cooking many householdsfetch water from ouside (public handpumpsalong wat clanels at the villge periphey or canals)or have water deliveredto their homes. 93 About 50 pet of the householdsin the sampledarea have installedelectric motors on thei pnvse handpump.Te motodredpump, because it can Lft water froma greaterdepth than the manualpmp, yieldssomewhat beter quaity waterthat householdsnot wiling to expendthe time or efbot quiredto c watoefom ouide use for dnkng and cookingas well.The electic motors also provd th m upgaded sevices tat were mentionedin the case of the sweet water zone. Hower, th quity of pipe water is clealy considd supeeior,and, whereverthe option of connetg to such a systemis avaiable, the conction fequenciesare high. Table D-4 showsthe patte of houseld choiceovr vailale serviceoptons along withtheir apprximate costs

Who lDab EZsetric Metw.?

9.4 Onoespin, the saisdcal analysis(for detailssee TableEB1) confirms expectations based on econc theory. Beer educated,economically affluent and propertiedhouseholds are morelikely to inall electricmotrs Houseol wth moreanimals kept insideor just outsidethe houseare less lHey to install motors Nhi sugget that animalneeds are not a factorin the decisionto install electricmotor Perhapswater obtined from this source is consideredtoo expensivefor such a

95 One impota differen fromthe pattemin the sweetwater zone is that vadablesrelated to householdsie sad tme spet in feng water are quite unambiguouslypositive. This would makesene in the brackishzoe wbere motorsare also used to imprve upon the qualityof water availabl fom handpump. lius, motorswould serve both the objectiveof making water-related setvicesavaib and of prwvidingbetter qualitywater for drinkingand cooldnginside the house. This is one reasonfor tho greaterinidence of motor-operatedpumps in the bracish zone (62 percent of in Type B villgs ainst 30 percentin the sweet water zone).

9.6 As in the sweet waer zone, householdswith access to piped water systems are signfty less likelyto isa elctric motors(33 percentof householdsin Type A villagesagainst 62 prcent in Type B villages). Th Is because,ideally, piped coxnnions providebetter quality water as wRlL

PrqgramReou Series 47 9.7 However, 33 pert of households In Type A vilges continue to have motor-opeated pumps (29 percent have a piped conection as well as a motor). Ihe correspondingnumbers in the sweet water zone are 1 peroet an7 peret. Part of the explanationcod lie n the fact that one of the sample vfllages, Sudbar, has had a water supply for less thaDsix months. Householdsthat had installed motors prior to the provisionof piped water might continue to retain them till such time as major repairs are required (44 permet of hoseholds in Sudbar have motots against 23 pemrentin Manawala). Ite other reason could be the greate relability of electric motos in povidig water for geteral household needs besides drinking and cooldg.

9.8 Householdsthat consider water from their prmary source (handpumps)to be satisfictory for health are less likely to install motors, but the vadable is not statsticaly significant. This would suggest that the quality of water is not the primary motivationin installing electric motors.

Wbo Connects to Pipd Water Systems?

9.9 In two of the sample villages,Manawala and Sudhart,households have had the aleative of connecting to a public piped supy system with domesdicconnection Of the sampled households 75 percent had availed themselvesof this altetive.

9.10 Ihe statistical analys (able B-2) shows tbat better educated and economicallyaffluent households are more likely to connect to piped water systems Householdslocated further away from the distributio line are less likely to connect because of the increase in connection costs

9.11 Householdsthat consider water from their handpumps to be unsafe for health are more liky to connect to piped systems. The coefficient is statstica1ly significant sggesting that the improvement in water quality is an imporant factor in the decision to connect

9.12 Farming households and those with a higher propordon of children are less likely to connecL Timscould be, as argued earlier, both because of reduced needs ad the availabilityof labor to fetch water for drinking and cooking fom outside the house. Tis rinforces the conclusion that piped water serves much more as a substitute for water fetched frm outside the house than moor- operated pumps.

9.13 Coury to the pattn in the sweet water zone, householdswith motors are more likely to connect to piped systems(the coefficientis positivebut not significant). This suggests that the two options are not considered to be complete substitutes in the bmckish water zone.

9.14 Houeholds that favor the metering of wat supies have a greater likelihood of connecing to piped water systems.

9.15 One variable which needs explaion is per capita household expenditue Both the linear and quadrtic terms are dose to significance with the fomer having a negative and the latter a positivecoefficient One explanationcould be the poor conelation betw iome and expenditure at low values; an increae in expenditure und to incomes (e.g., increase in family size) would

48 UNDP-WorldB Water and SaniatiPron ram stain the budget which might be adjustedby giving up items for which cheaperalternatives ane availabie Water,which can be fetchedat zero out-of-pocketcosts, Is one such item. nn the other hand, at higher leve, Iraes in per capita expenditurecould be expectedto reflec. ucaeasesin income.

9.16 The reasonsmentioned by the respondentsfor eitherconnectig or not connectngto the piped water supplysystem are lsted in Tables 5 and D-6. 'Mes we obtaied as rspon to open-eded questions.

EbetricMotors ad HouseholdConnections 9.17 Ideally,household connections to piped sytems shouldbe a dominantchoice because they are not only less expive than electic mons but also prvide better qualitywater. Howver, in reality,piped systemshave not provedto be weliableenough to be acoeptableas a dependable source for water-relatedsevices like idoor plumbing,showers and flush toilets. Therefore,even in a village lile Manawalawhere the piped supply is over 10 years old, 21 pernt of households continueto have electic motorsalong with domesticconnections 9.18 hs lack of rliability has restrictedthe utility of piped systemsto providinglimited water of good qualityfor drinkingand cooking, In villageswithout piped supplies38 pecent of householdsrely on the baseservice level of handpumpsin villageswith a pipedsupply the percentage alis to 21 percet. Hov, householdsthat desirea reliableupgraded service levelcannot do so withoutinvestng in a pdvateelectric motr.

9.19 For compron with the sweet water zone, Table D-7 shows the rlationship between coice of sevice level and the economic status of the household in villages where a piped supply option is available. Once again, it is the most affluenthouseholds that have both a domestic connectionand an electricmotor.

Pogmm ReportSeries 49 10. Findings in the Brackish Water Zone: Analysis of Willingness-to-Pay Bids

10.1 The followingopdt were evaluatd throughbiddig games: * A houneholdconnection to a sandad pipedwater systemof the tpe existingi Type A villages; * Tie same, with more affordablefinancing an . This optionwas offered only to the unconnectedhouseholds in Type A villages;and * A householdconnection to an improvodpiped water system. 10.2 he perceptionsof householdswnih experience of pipedwaer sysems am presented in Tables D-8 and D-9. It can be noted that 66 permet of the connectd householdseprsed dissatisfcod with theoperation of the system,with tie prmaty complait reating to its reliiity. 10.3 To detemninethe acceptanceof a well-functioningsystem, willingness-to-pay bids were licitedfor an improvedsystem. ITe improvedsystem was sipulated to provideclean and safe water on a continuousbasis with an acceptablelevel of reliabilityand presure. 10.4 Howeholdsthat had _maineduncmnected to availablepiped watr supplysystem were offeredtwo alternativefinang packagesto detemine if the one-timeconnecton costs werea fctor in their remainingunonnected. In the packagesoffered the water authoritieswere to bear 50 perent or 100 percentof the connectioncosts in returnfor a highermonthly tariff W _o to Pay for Con(ecton to a Standard System 105 The distributionof WIP bids for connecion to a standairdpiped water systemis prsented in TableF-1. In the brackishwater zone bids for such a systemwere . jtained fromType A householdsas well as fromTypes BI and B2 houeholds. 10.6 A numberof observaionscan be madebased on the data presentedin the table: * Ihere is little differencebetween the bids obtined frm Type B1 and Type B2 household.The meanbids are Rs 41 and Rs 37, respecdvely;the percentagesof householdsbidding more than Rs 12 per month(and thereforelikely to connect at the pevailing tarif) are 97 petcentand 90 perent, respectively;and the mean bids of the latter householdsare Rs 43 and Rs 40, respectively. * he bds and conction ratios are slightlyhigher in the Type B1llage in comparisonto the Type B2 vfllage.Given that distibutn Hneshave aleady been laid in the formerand there are no plans for the installtion of a systemin the latter, one wouldhave expected strategic bias to lead to a convene pattern. One obviousexplanadon would emergei£ despitethe advancedstage of installation, TypeB1 householdshave no moreinformation about system parameters (conection fee,monthly tarif) than Type B2 households.This wouldalso consitute concete eviden of the lack of communitypaticipation in the pepaation and implementatonof village-levelproject The relevantdata are presentedin Table

Prgm Reot Seie 51 D-1o, which rvewas a very high level of nonawarenessregarding basic information about the pamneters of the piped water syslem. This nonawarenessis just as prevalent in Type BI villages as in Type B2 villages. The pattem in the Type A villaSe is quite different. The mean bid (Rs 16) and the connection atio (75 pent) are much lower than in either of the villages without operational systems. Even the mean bid of the connected households(Rs 21) is about half the mean bid of similar households in the other villages.

Ihere can be two possible reasons for the above patter Fist, the experence of Type A households with the piped supply in their village could be negatively affecting their bids; Types B1 and B2 households,not aware of the ctual performanceof such systems,are not only bidding higher but more householdsare indicating an intention to connect than in the Type A village. Second, there could be an anchoring or strategic bias effect because the existing tariff is known to the respondents. In this, the design of the bidding game for already connectedhouseholds, which involves bidding up from the exiStingtariff to a value at which they would disconnect,could be a contbuting factor. 10.7 2ypesof Households ith Higher Widlingnessto Pay for a StandardSystem. A multivariateanalysis (see Table E-3) shows that the WI? bids are systematicallyrelated to a number of explanatoryvariables. Householdsthat require more water (geater household size and per capita consumptionof water) and those that need to spend more time fetching it from outside the house bid higher than others.

10.8 The bids are positivelyand significantlyrelated to householdexpenditure per capita; an increaseof Rs 100 in the variable would raise the bid by approximatelyRs 2. Younger households, those which include members who have lived outside the ilage, and those favorag the meterng of water connectionsbid more than others.

10.9 Householdsthat consider the water from existing sources to be satisfactoryfor health bid Rs 3.5 per month less than others but the variable is not statistically significant. The onsigificance of the existenceof electic motors (50 percent of sampled householdsponessed motors) indicates that they are not consideredas substitutes for piped water. 10.10 As mentioned earlier, householdsin Type A villages bid significantly less compared to Type Bi households. There was no significant differ between the bids of Types Bi and B2 households.

Wilnss to Pay for Conecdon to an Imprved System 10.11 The distributionof WIP bids for connection to an improved piped water system is preented in Table F-2. It should be kept in mind that Types AZ B and B2 householdswere asked to assume that, for the improved system, connection costs were zeto. This was to maintain compatibilitywith Type Al households that already had a domestic connection and would not have to incur connecton expenses again for access to an improved system. 10.12 The following obsevations can be made based on the data presented in the table:

The connection ratios are very high and similar for al three types of villages: 95 percent, 99 percent and 97 percent for village Types A, B1, and B2, respectively.

52 UNDP-WorldBank Waterand Sanitat Progratm * The mean bids are again similar for the Types Bl snd B2 vilages (Rs 58 and 51 respectively),which, in tum, are higher than the mean bid in the Type A village (Rs 33). * The fact that Types B1 and B2 householdswere offered connectionsto improved systems at zero connecon cost (compared to connectionsto the standard system for which they had to bear the connection costs) does not seem tc have affected their bidding to any significantextent. Thus, the icrses in the mean WI? bids over the standard system for the Types A, B1, and B2 householdsare 98.5 percent; 40.9 percent, and 37.6 percent, respectively.The correspondingfigures computed only over those households likely to connect (those bidding more than Rs 12) are 64.9 pement; 37.7 perent, and 29.5 percent,respectively. In both cases the increase is much more for the Type A households.This could be due to the fact that having had expeience with the standard system, the Type A householdscould appreciate the improvementsmuch better than the Types BE and B2 households for whom the standard system itself would be a great imprvement. 10.13 a pes of Households with Higher Wngnss to Pay for an Improved ysteL The multviate analysis (see Table E-4) shows vitally the same patters as for the standard system; households that are willing to pay more for the standud system are also willing to pay more for the improvedsystem. Wealth,education and health considerationsemerge as additionalsignificant variables in the case of the improved system. Response of Hosehods In Viages wth and whout Opera_Ion Piped Water Systems 10.14 To develop an understandingof the factors that affect the willingness-to-paybids for various senvice levels it is useful to focus on the Type A and Type B villages separately. Since Type A households have had actwalexperience of existing piped water systems, their eWaluationof such a system would reflct concrete considerations. It would be of interest to see the overlap with the conserations that affect the evaluation of an hypotheticalimproved system. A similar analysis could be performed by comparing the evaluation of the standard system by Type A and Type B households since for the latter the standard system represents a hypotheticalchoice. 10.15 Respome of Type A Householk. The results of the multivariateanalyses of WTP bids offered by Type A householdsfor existingand improvedsystems are presented in Tables E-5 and EA6 It can be noted that the only significant considerationsin the case of the existng system are income (expenditure) and ownership of livestock which is kept inside or just outside the house. Householdswith such animals are willing to pay Rs 2 per animal per month more than others. he typical livestockowning householdpossesses two animals.The premiumof Rs 4 per month constitutes 25 perent of the mean value of the dependent variable. 10.16 Ths last considerationis important and it confirms the hypothesisthat was developed dunng the analysis of the sweet water zone. Piped water from the existing system is used for washing livestock and this suggests that the consumptionfigures used by the PHED in the design crtera ought to take this factor into account. 10.17 In the case of the improved system the significantconsiderations are income, wealth, household size, age, concems about health, and attitudes towards meteing of domestic connections. While ability to pay and animal needs dominated in the case of the existing system, modern attitudes and quality of life considerationsemerge as sigificant in the evaluation of the improved system.

ProgrwnReport Seris 53 Younger more affluent and more di ting households bid higher for a 'modern system regardless of whether they ownedanalor not. 10.18 Responseof TypeB Households.Me responseof TypeB householdsto the existig system(Table E7) shows that householdsize and per capita consumptionof water ate significant deteminantsof WTPbids indicatingthat the need for drinkingwater playsa pan in the evaluatin In addition,age and preferene for the merg of domestic onmectionsare other signficant determinansindicating the influenceof atfitudes. 10.19 The vadablesfor wealthor incomeae not significant,suggesting either that the bids are not anchoredby the abilityto pay or that the aiff rates are a very low proporionof monthly incomesand so consderedgenraly affrdable by mosthouseholds bidding in an hypotial madre Ike latterhypothesis is supportedby the fact that wealthdoes emergeas a sigificant vaal in the biddingfor the impwoedsystem. 10.20 Moreover,the coefficientfor the owneishipof animalsis negativeand not significan TIs is a strong suggeson that learnig has taken place in the Type A villages sultiag in the significantlyhigher bids by the owne=sof livtock. Reponse of Ilc nete Househods k Type A Vlages 10.21 To explorethe reasonsfor nonconnection,Type A2 householdsthey were offered two alternativefinacing affangementsfor sharingthe costs of connecdonto the existig system.Under the fist anangementthe waterauthorities would contribute 50 percentof the costswhile underthe second they wouldcontribute 100 percentof the costs; in exchane the tadff wouldbe raised. 1022 Table P-3 shows the distrbudon of the WIP bids. It can be seen that while no householdsate at presentcnnected at the prevailingtariff (Rs 12 per month)the connecto ras (at the same tarif) underthe two a ts wouldbe 47 percentand 63 percet, dspecy. The mean WI? bids of the ecting householdsare Rs 16 and Rs 21 per month,respecively. 10.23 TableF-4 showsthe distrution of the WI? bids offeredby A2 housolds for an improvedsysm At he existingtiff 82 percentof A2 housholds wouldconnect to the Imvd system. Teir meanWTP bid is Rs 27 per monthas tarff.

54 UNDP-WorldBak Waterand Saitaon Poga 11. Estimated Revenues and Cost Recovery Potential

cwos rsId Water Sestm 11.1 Two yp of piped wat distiat systems a used In th backish wat zw systm boed oa tbwlls i canas (as in Bbahvaw)ad sym basd on filtraionof cal waw itself (as in Akga). C and O&M cass of toe two systm suder vario condWos am shownin Th. D-11. The ait cos percapita of a tubowellb4edsyste b Rs 300, whle tha of a cau walwarsytem b Rs 500. Bsed ns f e toWlcdita cos of to system can be ompued fr dth type of larg vMas studie (ppopulation 1o_00). 112 TIe PHEDesimas anua O&Mcos to rang fo 35 peat of capitblcos. IQ t cas of te sweet watr zons, the boe bond Is more for lr villae s bocum of econmies of scale a ag co_mp t of OJM costs beig fiSedin nature.Usingdte avera h_oud size in the samped am (89) *t monty charges rquid to rcv t O&M cosls fy at varios connco fequenci hv ben computed. For tubwll-basd sysatms, en th ue boundestmate (at 5 pec of capitalcos) of Rs 14.83per houseold per mraoth(assumg the pale 75 percent _c c freqency) are wel witi the acivabl rebgionFor can water systems the rresponding vau s Rs 24.71. Howeve, at th more appopiae valu of 3 percn of cpital cos the blt gue drops to Rs 14.83,agi atet that shouldbe qut adhevablegive te existng udff of Rs 12. hrow o a Sla ud Pipd Water Sytm 113 The co_nneon fr nies and evu estima peaining to te povison of a sandad pipedwater systemat mo lty tariff uses bhn In Table F-S. Mh connecio beq_ado fo TypesA, B1, and 22 vlg a plottd in FI 11.1. It can be seen that the frequenie for th TypesBl and B2 vigs are verysimila and muchhWi ta te for the Typ A vIlage. Th reasonfor the posib bias in h espwmes of Typ A house_oldshas beenmentond eaier. The plot susges that ere is no furdher ed to ditnus betwe te Type Bl and B2 vllages. hemm pbt of etmated reue i shown in Figure 112 11.4 If the taet of 75 prcnt conn s bto be maintined, the mothy if cannot be Iresd beyond th exiting of Is 12 At tl tdff a tbew ed system is eonomilly viab at te low boundof OhM costs (s 8.9 per ouseholdper mont) but not at the upperboud of Rs 14.83.A and wate systmmb not economlcllyviable en at th lowe bod (s 1483). pa dof - ImprovedP4pd Wa_r Systm 11.5 Tbe conneci fiqu_cles and vmue estae peoig o the pwviso of an improvedpiped watr systemat diffeet monty taff es ae shownin Tabe F-6 h conneci frequenci we ploted in RPr 113. Againt can be soted tht ere is no need to ditngush betweenmte Typo B1 ad D vaes e corsndig plot of estme revene is shown in Figr 11.4.

?ira ReotS u 5 100

|75 t

8504

25 \A |

o~~~

0 10 20 30 40 50 60 70 80 90 100 Mondhlytariff (Rs) ,Igure11.1 ConnectIon encts P_ovid ot _anud p"d water systm in 4age types A, B1, and B2 2,500

2000

1,500 X

/10 -- ---

O 10 20 30 40 so 60 70 80 90 100 Monthy tariff(Rs) Ffge 1L2 Mef vom (er 10 lellDMlb)-. 1!oMo of a sbdd fpe wow eyt it As 1 100

150

25

0 10 20 30 40 50 60 70 80 90 100 Moutly tariff (Rs) F%we 113 Comio feqms. Pri*h of a Iupnwd piped waft sysre f hi vivl typesAtBB 1 t e B,2 3,000

2,500-

2,500

2,000

1,500

500

0 I I 0 10 20 30 40 50 60 70 80 90 100 Monthlytarff (Rs) Fuigr. 11 MoaftIy riemie (e 1*0 bom ): Pmvidsa ofn lm proved piped wafte system In vIllag types A, Blt aindB2 11.6 Agan, if a lowerbound of 75 percnt connecionsI to be maintained,tt can be en that the taffcnbeserange of Rs 17.5to s 25.0per month ina Type A vlagF }*m IIA showsthat revwues wold be matx edat Rs 25 ponmotL Ihlus,atdff of fo eample Rs 20 per monthwould achieve both high connectionfrequencies and highrnveue colections. If a Type B village is used as a reference,the feasiblerange for the monthlytadff coud exend to R 35. Whiaethe exactcosts of impvng the piped water systemane not known,it seems that at least the O&Mcosts could be fullyrecovered without causing households to discomect om thi systm becau of an utnwllgns to pay an increaod taiff. Co,parbon of Stadrd and Impoved Plped Watr Systems 11.7 Since the 'fype A villagesprovide the lower boundson connectionfquen and estimatedrves we can use the responsesof TypeA householdsto comparethe gmis rultig impovingthe existng pipedwater systems. Figures 115 and 11.6show the connection&quecies and estimatedrevenues in a TypeA ve resultig fromthe pwvisionof standrd and impwvd systm 11.8 The imprved systemcompletely dominates the standardsystm Ih moly tariff can be raisedfrom Rs 12 to Rs 20 withoutthe onnecidonhequency drpping below 7S prcent. The maximumestimated monthly revenues go up from Rs 935 per 100 bouseholdsto Rs 1693 per 100 households. 11.9 The Responseof UnconnectedHousdsold in Vidlageswids an Opetadetm Piped War System. TableF-7 showsthe responseof Type A2 householdsto the four opions offerd to them: the standamdsystem, the standad systemwith two financinga emes, ad the imprve systm Mheconnection feques and the estimatedrevues ae plottedin Figures11.7 and 11A 11.10 The impowvedsystem dominates the otheraltnatves Thusthere should be litte dout that impovementsin the existing system have a greaterpayoff than offeg special inentives to householdsthat have not connectedto exsting systemsin the badcsh waterzon

58 lNDP-WorldBank Water and Santa Pogram 100

50

25

0 _ J - - ~- - - .13* - 0 10 20 30 40 50 60 70 80 90 100 Monithlytariff (Rs) uwi lS COnP fiequsmeeqIn a type A v3.s: St _dadvwi Impovd plped VWe

2,000

1,00 ISOO/ \~~~~~~bnroe

1,000

500 NStandard

0 10 20 30 40 50 60 70 80 90 10O Monthlytaiff (Rs) F1g UA6 Month "mm (Per 100 l_ ) In a t A ADie S d Xr_d s 100

75 \ \ '\, Standard(water authoties bear50% of connectioncost) j50 StanddStandard (waterauthorities bear 100% of connectioncost)

o to 20 30 40 so 0 70 0 90 10 1|2SO~~~~~~.mp,-e Montly tariff (Rs) F%pm 11. Cemelma frequenciesunder dlffentd opdkuwpresened to type A2 h.mhdbo

1250~~~~~~~~~01

1,000 i \ Standard(water authoriies bear 50% of connectioncost) Standard(water authorites bear 100%of connectioncost)

O-

2500 / /q g' a \ '''''4>

20 4

0 10 20 30 40 50 60 70 80 90 100 Monthlytariff (Rs) men 1L.8 MatM rev ampr 100-neS) m_dir dffren opihu presetd to pe A_ hb 12. Arid Zone: Overview and Field Procedures

12.1 RawalpindiDisnict is locatd in the nrtffemn, ai-Iigated part of Punjab. The districtcapital, Rawalpindi, has a populationof 794,843accordn to the 1981census. It is situated 16 kilomets from the federalcapital, slabad, and 272 kiome m LahorI, the capita of Punjab. RawalpindiIs the ft largt city in w coouy ad is the headquartesof the armed forces. Not much lndustryis locatedin the distdct, .whichhas traditionillybeen an ameaof high recruitmentin the army. 12.2 Almostthe enti area in RawalpindiDistrct lien the ar zone oundwater i aviable at much greaX.rdepths than i the ohr district sted sadve this ddes up dri the peak summer period from April to July. Peraps becauseof the lack of water, the vies in the arid zone are much smaller,the tp villagepopultion lying between1000 to 2000 ihabitants 123 VIllagesin the arid wne rely on wells as their pdmary source of drinkingwater. However,unIe the other two zones studied,these welb are generallynot withineasy reachof the households.Becaue of the scarcityof agdcutual land mostvillages ae locatedon tockyformations, usually at an eleation above that of the agriculturallad. he undwaterIs often at a more accessibledepth near such and becauseof seepagefrom somewater source(the Soan Rive i three of the studyvillages) or the lower levadon. Thus,many wells, pivate as wel as public,are located at a dse frm the vllagev. In addition,the natual sourcesof water (river,rvne, ec) are also at a distance and often inlve a steep climb on the way bak

12.4 Because of the distance and the elevaion,water i8 fetched oldy for the most essenal uses. Many water-relatedactivities take pace at the sources, eg., betbing, laundy, and watering of animals. Durin the summer,households spe a mely 6.5 peronhous per day in trvelling to and fromwater sources(cropared to 4 hours per weekin te braksh waterzoe). Most of this time i8 spent by women and childn except for that rInved in the watering of animas, which is generallya task handledby men. 125 TMesummer months from Aprl to July are particularlydifficult becaue of the bwerng of the water table. Waterin wels becomesalmost inacnessible and waitingtime becomes exceedigy long. Ie need for waterbecomes an overwhelmingconcern durng this periodbefore the grundwater is rechargedby the monsoonrains. 12.6 Vendorsare presentin vllges withoutpiped water but onlya minorityof houeholds (5.7 perment)use them on any srt of guar basis. ds is perhapsdue to the cost whichcan reach betweenRs 100 to Rs l50 per monthif a vendoris used to delivr 30 litemof waterdaily. However, ndorsare routinelyused on socialoccasions like mariages and deaths. On such occasionsa vendor may be paid betweenRs 300 to Rs 500 to fetch as muchwater as may be neededover two or three days.

fIhe oomputadonsexclude queuing because infmation nly avaiable for mamximumqueng uim

hProgaMReport S,er 61 12.7 Becauseof tho desperateneed for watew,quite a few householdshave developeda privatesource (24.4 percent) or haveattempted (5.4 perceat)to do so. Whilesome have expmented with wells, otheashave tried a simplebore down which a tiuweinch diametermetal sleea can be loweredand dtawn up filledwith water. Mhecosts are quite hig Digginga well costs RB200 per yard throughsoft arth and Rs 400 to Rs 500 per yardthugh bard,stony, or mcky fboations. A bore cos Rs 75 peryard and oftn has to be bestwe 20 to 2S yads deep. Even so, it yieldsUtte water and invriably dries up duimg the summermonts.b. Co,msequeatly,efforts have icreased, begnning with the 5-PointPrgamme, to providepublic water suppls to vilages in this rgion Many villages have received,or are in the prcess of receiving,public water supplies. Vilae smaller than 50&0Inhabit have baen selcted to recve piped watr sym with house connections, contay toth stated polcy of the Punab PHED. No schemes based on public taps ar under oonsideraton. 12.8 Connectiontates were vezy high, and even in vilages suppliedfor less tau six months, almost all the householdshad conected. This was so despite the fact that the monthlytaiff of Rs 20 was much bigherta inithe other two zones studied. is underscoresthe great need for water in the arid zone. Seacton of Study Viages 129 Vlages with public watr suppliesin RawalpindiDistrict showedtwo import featues. Most of tee supplieswere relativelynew (two yearsold or less) and the comecdonrates were very high (closeto 100 percent). These cha are reflectedin the Type A villages selected. 12.10 Becauseof te highconnection rates there is no meangful disdtct betweenhouses With(Al) and without(A2) connections. heefore, a compaableanalysis of actualchoice behavor basedon connectonto the watersytem is not possiblein the add zone. Ihe selectionof B1 and B2 vas was staightd. All the sdected B1 villages had public water systems at an advanced stage of constuc 12.11 The smal populat of the viages in te arid zone made it nocessy to select tree villagesof each type.The followingare the partca of the nine seleced villages: Jawa Bands : TypeA Dhalla Papin Payal : TypeBi Gorakpur Dhuddia: Mohra : Type B2 Bodhial

MTheavrage capital cost of a successfulprivate source i Rs 7960 and the monthlymantenance expene are appxmatey Rs 40. TI uahl attemptshave cost Rs 5730 on the avege.

62 ENDP-WorldBak Water and Saiti Prgram 12.12 Tbreevillages, on of each ty (Dhalla, Gomkhpur, and Bodhiul) were located by the Soan River, ho main peenna iver in the area. The others were located by noupeal ravines. Th fllowing villages were located at a consideaableheight above ground level: Dhala, Gorakhpur, Bodbial, Papin, ad Banda. A detiled profile of the selected villages is provided in Table G-1L Dmxt*a of d I and BM Games

12.13 The questionires are essentiallythe same as those used in the brackish water zone. The oe 4mjoraddition peains to secdon addedto obtainmore detailedinformaton an al the waer sourC0Savaiabl and on the pattem of water usage. ITis additionwas made because, unlike the other two zo (whre handpmps inside the house were the prmary source of water), many more sources e uosedand diffent soucs were used for different purposes. It was felt that more infmation ought to be collected if a source choice model was to be constructedat a later stage and also to obtain estimte of te qutity of water fetched and time spent on water-rebltedactivities. 12.14 The specification of a soure dhoice model requies the precise identfication of sources. I the sitation where multiple facities of a gie type existed (e.g, more than one public hadpump in the vilg) the failities were tecorded in the questionnais by their local nmes (e.g, moique handpump ot viUage center handpump, etc.). At the time of data entry each facility was given a uique two digit code. Thus source 43 would indicate a public handpump (source type = 4) bcatd by the mosque dpump number = 3). Each code would denote the same facilty for all households. es Idecation codes would also be used to cross-referencethe data on source charcteristic and water usag. 12.15 Iormation on water usage patters dunng summer were obtainedfrom allhouseholds. Howee, Infomation for the winter monthswas collected only from a sub-sampleof households. This vas prmaiy to reduce Intview time and also because the limited objective was to obtain an average scalg factor to estimate the cosumpion of water during the winter months.

12.16 Ther was one additional bidding game In the arid zone. Ihe details of the bidding ames and the opdons they wera dedgned to evaluate are as follows:

A schme basod on public taps in which a public tap would be located at most 20 yards fom any house, Water of satis&ctoiy quality would be available for apprximately 4 hours per day. Ibs option was offered to households in B2 villages * The standardPHED piped water scheme with house connectionsof the klnd existing in Type A vilages. This option was offered to all householdsin the sample. The WIP of already connectedhouseholds (Type Al) was elicited by bidding up from tho existing waM The WIT of unconnectedhouseholds (Type A2) was elicited by bidding up from a taiff ate of Rs O per month. * An Improved piped water supply system with house connectons which would supply clean and safe water continuouslywith adequate pssure and reliability. T option was offered to every household in Types A and B1 villages.

t i tl ad aalyses generatedby the survey, grouped by subjec$ appear in AppendixesG, H and L

PgrWn Repo" Serie 63 1.17 As noted in the caso of the backrlshwater zooe, the desig of the biddinggames ensuredthat protestbids and gene zero bids wouldbe cleary identified.In addition,the low and high swtg points to test for stang point bia wee Rs 30 and Es 50 respecly. 'ese were higher than in the brackis water zone becaue d exitng tadff was already Rs 20 per month comparedto Rs 12 in the braksh water zone and Es 10 in the sweet water zone In the mndomlntionof the sating poits, howueoldseiter had low suarg pointsor highstarting points for both the gamesoffered to them (exist and imprved sysm). C _nadcugth Suvy

12.18 The suvey was caded out in the 9 villages over a 10-day pedod (June 13-22, 1989). A tota of 401 interiews were completed in the 9 study vMages, as follows: Jawa : 44 Banda 48 : TypeA : 140 Dhalla 48 : (A1: 134,A2: 6) pin : 42 Payal : 48 : Type BI: 140 Goakhpur 50 Dhuddian : 30 Mohra 43 : Type B2: 121 Bodial : 48 12.19 The interviewswoer condued durn the peak summerseason duing which water needsare criticaL 12.20 A majordiflfrence fam the other two zoneswas the inclusionof femalerespondents whe available; 11.4peroent of the resodet were wome 1221 One visit was made to each of the 9 villages Sone basic socioeconomic characetics of samplehousebols and their attitudestowar water-relatedissues ane presented in Tables 0-2 and 0-3.

64 UNDP-WorldBoo* Wu.a and Satao PApw 13. Findings in the Arid. Zone: Analysis of Willingness-to-Pay Bids

S Op Evaatd 13.1 he folowing options were evaluated thrugh bidding games: * A scheme based on public taps (this option was offered to householdsin Type B2 vilges); * A household connection to a standard piped water system of the kdnd exisng in Type A villages (this option was offered to all householdsin the sample); and * A householdconnection to an improvedpiped water system (this opdon was offered to householdsin Types A and DI villages).

132 The public tap option was offered since, under PHED guidelnes, this is the service level that ought to be provided In the arid zone where the majority of villages a¢e below the critical populationsze of 5000, which makes a vilage eligible fsr house connections. The bidding games would enable an asssment of the willin to pay for public taps as wel as yield an estimate of the premium that householdsplace on domestic connecios.

13.3 The perceptionsof householdswith experience of piped water systems are presented in Tables G04 and 0-5. It can be seen that households ae reasonably satisfied with the service primariy because the systems are relatively recent and, despite their shortcomings, are a major improvement over the past situationL Nevertheless, wilingeto-pay bids were elicited for an imprve system. The imprved system was stipulatedto prvide clean and safe water on a continuous basis with an acceptable level of reliabilityand pres

Reponse of HeO bMIJ Villages wstout Pln for s of Publc Water Systems 13.4 Householdsin viages which are not under consideraion for the installationof public water systeMswere offered two supply options:A system based on public taps and an alternadve based on house connectionsto a standard PHED system (the details of these systems were descrbed ealier). 13.5 The distributionsof WIP bids for the two sevce options are presented in Table I- 1. A number of obserations can be made based on the data presented in the table: * The mean WTP bid for a system based on public taps was quite high (Rs 35 per month). * However, a sizeable minority (13 percent) of the householdswere not wiling to pay anythingfor a systembased on public taps. The stated tariff for such a system is Rs 5 per month, but ff a more reasonablerate of Rs 10 per month is considered for the arid zone, appimatey 16 percent of tie householdswould not be willing to join the system. * The mean WI? bid for the standard PHED system wnth house connectons was Rs 55 per month, a premiumof Rs 20 over the publictap system. It should be

Pro8ramRepon Series 6S kept in mind that the formeralso involvesadditional fixed costs due to coaneton and installation. Only 2.5 percentof the householdswere not willing to pay anything for the standardPHED system. If the existingtariff in the aid zone (Rt 20 per month) Is used as a cutoff,the data indicatethat 9 percentof the householdswould not connectIf such an opion were offered. W to Pay for Conection to a Standard PlIED Systemwith HouseConnections 13.6 he distributionsof WIP bids for a connctio to a standardpiped water system with houseconmections in the three typesof vs ate prasentedin Table1-2. 13.7 Ihe followingobseations can be madebased on the data prented in the table: * There is very little differece betweenthe meanbids obtainedfrom Type A and B1 vfllages. 'e meanbids are Rs 39 and Rs 4Z respectively.Ihe meanbid in Type B2 villages,however, is sgnificantlyhigher at Rs 55. * Thereappears to be a significantdifference between the connectionrates in Types A and Bi villages,95.7 perent and 78.6 percet, espectvely. How , the connectionratio in Type A villageis basedon actualobservation while that in Bl villagesis derivedfrom the WI? bids. Householdsbiding below Rs 20 per month are assumednot to connectif the systemwere installed. The comarable rate in Type B2 villagesis 90.9 percent * The aboveinformation suggests the followinginterpretations: (a) The pipedwater supplies in TypeA villagesare of very recentorigin (less than 1 year in Jawa and Dhallaand less than 2 yearmin Bandaat the time of the suvey). Type B1 villages have piped water sstems under construction.Thus the two types of vilages are quite simlar in one respect Unlikethe sweet and backish water zones,Type A householdshave not had sufficientnegative eeience with the systems to lead to lower bids in comparisonwith vilages withoutsuh expeiene. (b) If a strategicbias exist, both Type A and Type B1 householdsare lky to manifestthe bias in the same directon, i.e. by bidding (c) In thelight ofthe abovetwo arpmens, the osenessof the meanbids is undestandable. However, the underbiddhg in Type Bl villages 1s manifestedin a low coneton ratio (bids less than Rs 20 per monthbeing consideted as not likely to connect to the stanrda system) Such a manifestationis not possiblein Type A villageswhere the connectionchoice has alreadybeen made. (d) An upad sttategicbias mightbe expectedin TypeB2 villageswhere there are no plansfor the installationof pipedsystems but wherethe feltneed for such systems is equallyacute. Both the mean bid (lb 55) and the connectionrae (90.9pert) ar higherthan in Type BI villages Wlnnss to Pay for an Iaproved PipedWat System with Home Conneadons

13.8 The distributionsof WI? bids for a conewtionto an impoved piped water system with houseconnections in Types A and B1 villagesare presentedin Table1-3.

66 LWDP-Wrd Bank Water mad Saitaio Prgram 13.9 The folowing obsations can be madebased on the data prsen_td an the table: Omc again the man bids in TypesA and Bi ves ane faiy close, being Rs 51 and Rs 59, respecIvely.Also the connction raes fairlysimilar, beng 95.0 percentand 999 percent,respectivdy. Of the 6 householdsunonnced to the standardsystem in Type A villages,2 indicatedthat they would conec to the Improvedsystem, bidding Rs 35 and Rs 45, respcvely. Incidentally,they had ndicatedthe unreliabilityof the standardsystem as their primaryrason fo not cnnecting. Of the other 4, 2 were sigle pero householdsand 1 was occupying a rentedpremise. The foth hosehold indiated no need for pipedwater because of access to a private handpump ilsidO the house. * The mean bids are sgnficanty highercompard to the sandard system The premiumis 31 percentin Type A vlages and 40 percet in Type B1 villges Mt Amly of Wilgess-to-Pay Bid for Piped Wate Sysem 13.10 Rela& is Type R2 ViUaes. Cm in of Sytenm wh Public Taps and House Connectios. The rsults of a mulivarte analysisof WTP bids for a piped water systembased on publictaps I Type B2 villagesare presentedin Table H-1. 13.11 Lager householdsare wiling to pay more (it 3 per month for each additional member)as are householdswith highermonthly expenditure per capita. Householdsdisatisfied with dte quality of alteutive sources of water bid much higher tha houseods that wee satisfied. Hoeholds that hvored meteing of watersupples also bid higler. 13.12 Most otherbvaiables had the expectedsig but wre not signifcanmt.The varables rpresentig householdlabor supply are interesting.Households in whichlabor is scarcedid bid more than householdswith more labor availableto fetch water, but the differencewas not staically sigificant. Ti suggests that public taps he a limited attction for households with few labor plies, possibly becae public taps do not obviae the need for fetcing water from outside lhe house. Oe would ect the differencebetween the two types of houeholds to be much more marked In the case of piped connectionsinside the houe. Only 3 percentof the sample espondentsindicated that they are dsasfied with e qualityof water avaiable from her souces, and, as one would expect,these rsondent bid significantlymore fot publictaps. 13.13 The negativesipg of the coefficientfor extenal exposurernorc the inrpreadion that a systembased on publictaps is consderd an &friorgood by thase with expeienceof systems based on house connections. 13.14 The resultsof the multvariateanalysis of WrP bids for a standad systembasd on houe connectionsfor the samehouseholds ae presentedin TableH-2. Onceagin householdsize and monhy per capitaexpenditue am a significantpositive determia of WIP bids. In addition,for this opdon, water consumptionper capita is positivelyrelated to WI? bids althoughit is not statisdcallysignificant. 13.15 As expected,both labor supply vaables are significantand negativelyrelated to WT? bids. This negative lationshipand the positiveassociadon with water consumptionper capit clearly capturesthe differces betweensystems based on publictaps and houseconnections

Prm Rprt Series 67 13.16 Householdsdisatsfied withthe qualityof alternativesources of waterane stll willing to pay more for piped waterbut the coefficientis not sgnificant. Ihis indicatesthat otheradvantages are assocated with house comections. This is in contrastto the attitudetowards public taps where healthconsiderations had moreweight. 13.17 None of the variablesreflecting pemonal charactewistics or attitudesis significant, perhapsindicating the factthat pipedwater Is not considereda discretionayor luxurygood but a basic necessityin the arid zone. 13.18 In both options the village-levelvariables are significant. WTP bids decrease systematicallywith distancefrom the districtheadquartm and with proximityto a perennialwater source(vilage dummy= 1 for the one villagewith acces to the Soan River). 13.19 Ritsu in l)pe Bl ViUlages.The resultsof a multivariateanalysis of WIP bids for a stadard pipedwater systm in TypeBI villagesare presentedin TableH-3. ueseresult8 presented a pwle. The most obvious maifestation is the behavior of households that could be eopcted to value piped water. hus householdswith more animals bid significantly less than householdswith fewer animals. Similady,households satisfied with the qualityof alterative sourcesof water bid higher(Rs 17 per mouthmore) than householdsthat considerthe sourcesunsafe for health. 13.20 Further,households with more labor supplydid not bid lower than householdswith less labor supply. The signs of the coefficientsof the proportionof womenand childrt are the oppositeof what one wouldexpect and are insignificantas well. 1321 However,the bids are positivelycorrelated with householdsize, valueof houseand the owneishipof land or property(and significandyfor the first two). The valueof the coefficient for householdsize (Rs 1.5) is much smallerthan in Type B2 vfllages(Rs 4). 1322 One explanationfor these results is that nrspondentsbid stategically. The water suppliesin Type BI villageshad alreadybeen sanctioned and it couldbe possiblethat the respondents consideredthe bidding gomesto be an attempt to set the monthlytarift This would explaintbe systematicunderbidding by householdsthat could be expectd to bid (and that do so, based on evidencefrom B2 villagesin the arid zone and most otherexperiments in the other zones).higherfor piped water.1 13.23 The generalaffluence of the householdsseems to havedetemined a basefor the WTP bids with the more affluentstaiting froma higher level.However, beyond that, bids seem to have beenquite systematicallysuppressed by the respondents 1324 If this interpration is conect, it would suggestthat the mean bid receivedfor a standardpiped systemwith house connections (Rs 41 per month)is an undeestimate. However,even with this possibledownward bias, the WTPbids are quite high. Perhapsthe meanbid in Type B2 villages (Rs56 per month)can be consideredan upperbound yieldinga range of Rs 40 to Rs 55 per monthas the one withinwhich the 'true" averagewould lie.

8UnJikethe sureys in the sweet water and brackishwater zones, the surveysin the arid zone were caried out after the 1988general elections in Paistan Duringthe electioncampaign, many promises were made that water schemeswould be providedas a 'gift to the people. This could explainthe apparentstrategic responses of householdsin the arid zone.

68 UWDP-WorldBank Water and Santato Progm 13.25 The results of the multivariateanalWysis petining to an improved system with house conections are presented in Table HA4. No major modirtcationneeds to be made to the conclusions derived fhom the response to the standard system. However, as in the other environmentalzones studied, for an hypotheticalimproved system, affitudinalvariables become somewhat more significant. Thus, households aware of piped systms bid sgnificantly higher while households that consider the pwvision of water to be a governmentobligation bid significantlylower (Rs 15 per month). 13.26 Ressda in 2)pe A Vidlages. The results of multivadate analyses of WTP bids for standard and improved piped water systems with house connectionsin Type A villages are ptesented in Tables H-S and H-6. 13.27 The explanatory power of the models is poor in compadson with the models for the other types of villages. However, there is no distortionof responsesas witnessedin Type Bl villages, most of the coefficients having the corect sign without being significant. Tis is probably because the system is already in operdon in Type A villages so that the respondentsmight not have intepreted the bidding games as an attempt to set the tariff. 13.28 The labor supply variablesare both highly significantas expected, and in contrast to Type BI villages. The only other variable which is significant is monthly householdexpenditure per capita. The vfllage-levelvariable, distance fom district headquartes, is insignificant,perhaps because two of the villages are at the same distance although along different directions. The village dummy for Jawa, a village with a new water supply having operational problems, is insignificant. 1329 Ihe WTP estimationswere based on the responsesof connectedhouseholds (Type Al) only because of the very small number of unconnectedhouseholds in the sample (6 out of 140). When the latter are added to the regression, the variable representingconnection to the system emerges as highly significant. Not much change results in the coefficients or significace of the other variables but the overal significance of the regression improves considerably. As expected, the bids of the unconnectedhouseholds are sigificantly less than those of the connected households.

Pmror ReportSeries 69 14. Estimated Revenues and Cost Recovery Potential

CA etof bed War Syste 14.1 The estimationof capitalcost; of piped water systemsin the adidzone based on a otionalvalue of capitl cost per capitadid nOtprowve very useful. This was so beu the parameter is vey seasitiveto popuationsize; thus, whetherthe populationis 1000or 2000(the tpcal nge in the arid zone) make a tremendousdifference to the capitalcost per capita. 14.2 To overcomethis limitation,the actualproect costs for the six villagesincluded in the sumple(three each of Type A and Type BE) were obtained from the PHED. Actual O&M allocationsfor the three Type A villageswere also obtained. Usingthese figures,averages were computedfor a typicalviUage in the arid zone. The averagecapital cost was Rs 1,440,000for Type A villagesand Rs 1,309,000for the six villages. The averageannual O&Mcosts were Rs 68,937 for TypeA vMages.Thus annualO&M costs as a perentage of capitalcosts workout to 4.8 pert in Type A villagesand 53 peret inthe s villages& Thisis in conformitywith the 5 percent benchmar used by the PHED. 143 To obtain per capitacosts and the tariff requiredfor full recoveryof O&M costs, the populationsof the six villageswere avenged to obtainan estimatefor a typicalvillage. ITe 1981 ensmuyields an estimateof 1230.The 1989population was obtainedby assuminga 3 percentannual gwth rate. A simlar averagng procedureyielded a typcal householdsize of 63 in 1981(this is lwer than the estimate(7.5) obtainedfrom the sampledata). The abovetwo estmates togetheryield the numberof housesin a typicalvillage of the arndzone. For the threeType A villagesthis estimate is 222 while for the entire six villagesthe numberis 208. 14.4 The connectionrate in Type A villagesdetermined from the surveyinformation is 95.7 percent(only 6 out of 140 housessureyed were not connectedby choice). Thus, one could expc. 212 or 200 housesto be connectedin a typicalarid zone village,depending on whetherthe averagingis basedon TypeA villags or TypesA and BI villages. 14.5 Usingthe abovedata, the aveage capitalcost per capitain the aid zone wo out to be Rs 83& The monthlyO&M costs total Rs 5745. Thus, the monthlytariff requiredfor full recoveryof O&Mcosts vares betweenRs 7 (TypeA villages)to Rs 29 C(ypesA and Bl villages). All the abovedata and computationsare presentedin Table 0-6. 14.6 The existg monthlytariff for a standardPHED system with houseconnections in the adidzone is Rs 20. Provison of a Standard Piped Water Sstem 14.7 The connectionfrequencies and revenueestimates pertaining to the provisionof a standd piped water systemat difrent mouthlytariff rates are shown in Table 1-4. ITe plots of connectionsfreqcies and estimaftervues againstmonthly tariff are shown separatelyfor Types A, Bl, and B2 villagesin Figures14.1

ProgramReport Seri 71 2.500

100 - 2000

J75 1,500w

350- I,oo Connections

25 500

0 0 0 10 20 30 40 50 60 70 80 90 100 Monthlytariff (Rts) FIgr 14.1 C_nsloan fequeacks ad monthlyrevem (per 100 bousehods): Pmvn of a sltadad pipd water systemaIo a type A vlge

2,500

100 -.. 2,000

50 -1,000 Connections-

/ _ 0 0 10 20 30 40 50 60 70 80 90 100 Mtnypy tarff(Rs) elp 14.2 _ _uede alld DLf r0es (Per 100 bo eik): PmVIdo of a 8_dia P d 8 tom a type B1 Vi_v 14.8 Beyond a certain tueshold connection frequency is very sensitive to changes in the monthly tariff This threshold occumsat Rs 25 per month in Type A villages, Rs 15 per month In Type B1 villages, and Rs 35 per month in Type B2 villages. 14.9 As remarked earlier in compating Type A and Type BI villages, the connection frequency in Type A villages is a better guide to actual behavior. Therefore,we can expect that a rise in tariff from Rs 20 to Rs 25 per month would not cause any lowering of connection frequency. If we futher assume that there was strategic underbiddingin Type A vilages and overbiddingin Type B2 villages we can expect that atariff rate of between Rs 25 and Rs 30 per month wouldresult in connection rates ranging from 95 percent to 85 perent.

14.10 At these connection rates and tariffs, the estimated monthly revenue would be approximatelyRs 2500 per 100 households in the village. Using an average of 212 for the number of households in a typical vilage the total monthly revenue generated would be Rs 5,300, which is in the same neighborhoodas the monthly O&M requirementesimated from cost data (Rs 5745). 14.11 It seems clear that tariffs can be raised to Rs 25 per month from the existing Rs 20 per month without any negative impact on connection rates. Tariff rates up to Rs 30 per month rmain in the feasible range. Between Rs 25 to Rs 30 per month full recovery of O&M costs is possible. This would be even more certain with a very small increase in the number of households over the next few years (in calculating a relevant populationbase, the PHED projects population for 10 years ahead of the approval systems using projectedpopulations ten years from the date of apprv of a scheme as their relevant population base). Provision of an Improved Piped Water System 14.12 The connection frequenciesand revenue estimates pertaining to the provision of an improved piped water system at different monthly taiff rates are shown in Table 1-5. The plots of connection frequenciesand estimated revenues against monthly tariffs are shown separately for Type A and BD villages in Figures 14.4 and 14.5, respectivly. 14.13 The tariff threshold is Rs 35 per month. Beyond Rs 35 connection frequenciesfall steeply from around 85 percent to around 65 percent. At Rs 35 per month the monthly revenues generated in a typical vilage of 212 households would be approximatelyRs 6400. 14.14 We are not in a position to state the extent to which the O&M expenses would increase for the kind of improved system descriled earier, but it seems likely that imprved systems are prematue in the typical arid zone vilage. However, householdsin larger villages with sufficient length of experience with standard piped water systems might be willing to pay tariff rates that could make the policy of providing selective improvementsworth investiting. Provisio of a PipedWater System Based n PublicTaps 14.15 Householdsin Type B2 villages were asked to bid on two different levels of service provision: public taps and a standd system with house connections. The comparativeconnection frequencies and revenue estimates are presented in Table 1-6. The compartive plots of conection frequenciesand revenue estimates against monthly tariff are shown in Figures 14.6 and 14.7.

Program Report Seies 73 3,500

- 3,000

100 --

-~~~~~~~~~~~~~~~ ~~~~~-2,500

75 - ~~~~~~~~~~~~~~2,000

jso -1500 I ~~~~~~Connectons 25 ~~~~~~~~~~~~~~~~~~~~1,000

0 10 20 30 40 50 60 70 80 90 100 Monthlytarff (Rs) FIgur 143 Cumnnhc freqenese sd mont revenu (per 100 bousubdus): ProVidon Of a standard pipe water system in a type B2 vilag

. 3,000 100 N ~~~~~~~~~~~~2,500

75 - \ 2,000

s o 1,500i0

25

0~~~~~~~~~~~~~~~~~~0

0 10 20 30 40 50 60 70 80 90 100 Monthlytariff (Rs) FIg 14A Connln fequenes_ ad monat Mev e (per 100 houebhos): Pr Qoo ot en Iuwped ppe water systm In a type A vMle 3400

3,000

100 -s - > 2 2,500

N ~~~~~~~~2,0O0 J75

21,000

25 / - 500

0 0o 0 10 20 30 40 50 60 70 80 90 100 Monthlytwaiff(Rs) F.e 14S Coettm frsquaks mad moaty revu (per 100 bo..e,M): efla apedva ped waft sytem 1 8 type U1viEtp

100

75 -

0~~ jo25L>--

0 10 20 30 40 S0 60 70* 80 90 1t Mont tariff(Rs) FIgure 146 Cmcihm hiquuadm bf tp 12 vilte 14.16 If the t is achiewto atlet a 85 perontcodo n rate, the montly tariff for usig public taps cannotexceed Rs 15 while the same percentag of householdsare wiling to pay Rs 35 a monthfor houseconnections. 'he preferencefor a systembased on houe connectos seems to be quite clear, over 13 perent householdsae not wfiing to pay anyting for a systembased on publictaps while the comparablefigure for houseconnections is 2.5 percent 14.17 Althoughthe £owenuesgenerated firom a systemwith house connectons .n a TypeB2 vilage wouldbe sufficientto reover full O&Mcosts, this mightnot be the case for a systembased on publictaps sne It isreasonable to assumethat the O&M osts for the two systemswould not be significantydifferenr9. In addition,tho difficultiesin collectingpayments from public tap systemsare well known. 14.18 We have some evidenceavailable for the differen in capital costs for the two optoDn PHEDdata rveaed that a vilage in the same vicnty as the studyvillages was oiginally

3,500

2,000 /~~~~\:lou~Hose connctions

21.500-

1,000 Publictaps

0 p p 0 10 20 30 40 50 60 70 80 90 100 Monthlytaiff(Rs) FIgr 14.7 Mont rerves (per 100 hmushW ) In a type B2 villge

1'9Savin are likelyonly an disizibutionsystem repair. ee ae estimatedat 1/12percent per annum of tie capitalcost of the distrut system. A tical estimateof the latter is Rs 600,000. Thbis wouldyield a savig of Rs 500 per yea, whichis less than 1 peent of the typicalannual O&M cost.

76 EWDP-WorldBank Watarand saa Pmgrm scheduledto teceve a system based on publc taps?' Mm deed cot eamate prpard in 1986 was for a sum of Rs 1,600,000. Hower, on te stion of villa notables, suppod by eletd representativesfrom the ara, it was decied to piwle house cM co insead The reviwed esmate was for the sum of Rs 2280,200. The mvidon rflected a capita cost ation of 42.5 percent In per capitaterms the capitalcost rose fromRs S76 to Rs 821. (Ilie per capitacosts are basodon an estmated 1987populadon size of 2776).

bThe schemeprovided for four 2000 galloncapcity RCCgond stog tn each wit a battey of taps. 21Schemebased on one overead 4000 pflon capacitytank.

ProgramRepwt Swim 77 15. Conclusions and Policy Suggestions

h Natu of Dead fw Wte In Rural Arma lS.1 Th most Importantfindg of this study peran to the natmueof the demandfor watea in rua Pjab. Te actual water suly sitatio in lmy of the ample vilages was considerablymore complex thian commonly assumed. It seemsclear that the perspecdvewithn which policy-maers viewe rur watr supy i8 in the pp ss of being renere out of date by theype of development. 15.2 WithinPakistan, the policyreging u rral water supplyhas been motivatedby the objectivesof healthimprovemens via the provison of beter qualitywate and ime savingsvia the locatio of moreacceible sourcs. Implicitin this perpte is a catoizton of wateras ar end product4a commodityrequired for such directuses as drinking,cooking, bathing, etc Withinthe study are, this pspecve was foundto be valid only in the ard zone. 153 In thtecenttal canal-irrigatedpart of Punjab,there was a grwing demand,not for ater as an end product but for water-basedamenities like indoorplumbing, showers, and flush toilets. Hewever,this demandcamot be servicedby hantpumpsor even public standplpes. It requiresa higherlevd of service. Whatrdul householdsare doingto fulfilltheir aspirationshas both important ~nomic consequencesand far-reachingpolicy implications.

15.4 TheSwet Wj rAreaa anXllusradv Ca. TI sweetgundwater area of Centrl Punjabprovides the best illustrationof the abovementionedissues. The qualityof the groundwater is goodand it is availableat easilyaccessible depth. Householdshave already upgraded their service from the taditional source of supply,public wels, to privatehandpump insidet house. These handpmps provideconvenient access to good qualitywater fit for all usages. Ihe serviceis reliable, bing efficientlyinstalled and mantale by the local privatesector, and inexpensiveto mainta Everyhousehold in the samplehad a privatehandpump inside the house. Householdshave madethis expenditure(Rs 1000 capitalcosts, in curnt value, for the pumpand the shallowwell; Rs 6 per monthfor maintenance) to avoid the inconvenienceof havingto fetch water fromoutside the house. 155 The official policy in the sweet water zone is not to providepublic piped water .applies. This policy is based on the reasoningthat pivate initiativehas succeededin providing convenientaccess to good qualitywater. However,the surveyresults indicate that a secondupgrding of service is well under way. Manyhouseholds are willing and able to pay for a higherlevel of serviceand in the absenceof reliablepiped water suppliesthey have installedelectric motors on their wells at considerableexpen The electic pumpslift water to prvate overheadtanks from which water can be used for showers,issh toilets and other indoorplumbing services. Thus, the services whichcould be providedby publicpiped water supplysystems are being replicatedat the individual householdleveL Does the officialpolicy retai its validityin thesechanged circumstances? Prvate Upgading of ServiceIn Central Punjab 15.6 In villageswithout piped water a the sweet water zone, 30 pementof the sample householdshad installd an electricmotor. In the brackishwater zone the comparablefigure was 62 percent. The percentageis higher,perhaps, because, in additionto the other benefts, electic pumps

Progm ReportSeries 79 also provide somewhat better quality water than handpumpsin the bracldshzone by lifting water from greater deps This is a very high percantage and cleady indicates the trend in the Central Punjab villages. 15.7 In viliages with piped water, the percentage of bouseholds with electric motors falls to 11 peret in the sweet water zone and to 33 pnt in the brackish water zoe. T1is suggests that piped supply systems with house connectionsare cleady perceived as substitutes for electric motors. However, in such villages, 7 percentof the households in the sweet water zone and 29 percent in the brackish water zone continue to maintai both opio This bngs to the fore the critical issue of the reliability of the various serice options. 15.8 A household'sdecision on how to satisfy its water needs is heavily influenced by the reiability of the differnt service options. A private handpumpis almost completelyreliable because it can be repaired locally and is entirely under the control of the household. In the Jweet water zone the water from the handpump is adequate for all usages, but the handpump cannot provide the amenities of indoor plumbing,showers and flush toilets, If the householdwants and can afford such amenities, it has three choices in villages with public piped supplies: (1) to connect to the public system, (2) to install an electric motor, or (3) both. 15.9 In the brackish water zone, the first choice ought to be the dominantone. The piped supply not only providesbetter quality water but its private costs to the householdare much less. The fact that 29 percent of households maintain both options is only due to the limited hours of supply hom the public system and its poor reliability. his poor reliability Imposesa high cost on households and, at the same time, underminesthe economic viability of public piped water systems.

15.10 FinancialCost to an Average Houswold of Different Sevice Options. The approximatecosts to an average householdin the brackishwater zone of vaious water supply options are shown in Table J-4. The handpump providesa base of serice at a totd monthlycost of about Rs 18 per month A household with both a handpump and a connection to a piped water supply would pay Rs 12 per month to the PHED for the water taiff and Rs 5 per month for operationand mantenance of its handpump. The monthly capital costs for both the handpump and the connection to the distrbution supply would be Rs 19 per month for a totl monthlycost of Rs 36. A household with an electrdc motor on its handpump would spend much more - about Rs 58 per month Households witlh both an electric motor and a domestic connection are estimated to pay Rs 76 per month.

15.11 A reliable piped water service ought to cost a household appro;imately Rs 18 per month at existing tariff rates. Instead, households waning upg4& services are paying about Rs 76 per month for the same level of service. The piped water system is not fulfilling its potential. It is ftioning as a supplement to the handpump in the sweet water zone and as a substitute (better quality water) for the handpumpin the bracish water zone. Reliable upgaded services can only be secured at present through investmentin a private motor-operatedsystem.

nFor private connectionsthe costs in the swewater zone are essentially the same as in the brackish wat"r zone, except that the monthly tariff is Rs.10 instead of Rs112. For the water supply options which require a private well, the costs are basially the same in the two zones for a well of a given depto

80 MU P-World Bak Water and San o Pgrwun 15.12 WhoDemaund H0er and MoreRefiable Sevce? 'The surveyresults Idicate (see TableJ-5) that, as expected,it is the wealthier,more educatedhouseholds that are demandinghigher service. The avegW constructionvalue of housesof houeholdswith onlya pivate handpumpin the bnackishwater zoe (in villageswith piped water supplies)was Rs 62,000. For householdswith both a privatehandpump and an electricmotor it was Rs 115,000. For householdswith a handpump, electricmotor and a domestc connection,it was Rs 145,000. In villageswith a piped water system in the sweet water zone, in householdswith only a privatehandpump, the most educatedmember of the householdhad an averageof 6 years of education. For householdswith a hamdpumpand an electricmotor, the averagewas 12 years of educadon. 15.13 Viflageelites have alwaysexerted a powerfuldemonstration effect an the rest of the population. At one time the elite used to be coized by the possessionof a brick houseor an electricityconnection. Now the distinguiin characteristici indoorplumbing and flush toilet. If historicalexperience is any guide,the choiceof the elite is a clear pointerto the trend of the futu 15.14 Wdlingnessto Pay as a Percestageof HouseholdInome. Thesurvey results indicate that the amountof moneyhouseholds are willingto pay for improvedwater suppliesas a percentof householdincome (or expenditures)is bwer than oftenassumed In the sweetwater zone the mean WTP bid was only 1.1 percentof averagehousehold income (see TableJ-. The percentof income respondentswere wiling to pay for improvedwater was higherin boththe bracldshzone (2.4 peroent) and arid zone (3.5 percent),but both esimas ae still well below the 5 pementrule of thumboften used to estimatehow much householdswil pay for improvedwater supplies. The Emnom_ of Viage Water SupW*Op_os 15.15 Ih efforts by householdsto povide themselvesindividually with Impoved water sevices entailsubstantial expenditures in the aggegate. Table 3-6 presentsan estimateof the actud amount of money curnly being spenton privatewater provisionin a typical village with a population of 5000 peoplewithout a piped watersupply in the brackishwater zne Assuming62 percentof the hoseholds havea handpumpwith an electricmotor and 38 percenthave only a handpump,households in such a vilge havealready invested over Rs 1 million(in cunent value)for pivate handpumpsand electrc motors. The operationand maintenancecosts of thae privatelyprovided water systemsis esdmatedto be Rs 9800 per uonth The total monthlycosts of these systemsare about Rs 23,9V'b. 15.16 Basedon cost data fromthe PHEIDin Punjab,a new pipedwater system for a viage with a populationof 5000 wouldcost about Rs 1,800,000-including the cost of 100 percentof the householdsconnecting to the distibutionsystem. The monthlyoperatton and maintenancecosts of a piped waftr systemare estmated to be about Rs 3800 for a totalmonthly cost of Rs 19,800. Given the approximatenature of these esimates,the totalcosts of a pipeddistribution system are essentially the same as the amount households re alhdy spending for handpumpsand electric motos 15.17 ITe esdtmatesof the oss of th. piped system assume that 100 percent of te householdsin the viUageae connected,the estimatesof actut, expendituresassume that 38 percent of the householdsonly bave a bandpump.In this sensethe cost estimatesare not comparble because the pipedwater systemproie a higherlevel of servicefor a greaternumber of people. 15.18 W gnesto Pay for Piped Watr S&#pL. It is in theabove context that the wilinge to pay for pipedwater ought to be evaluated.Once asain we tae a villageof 5000 people (562 households)in the brackishzone as an example.In such a villagewithout piped wate, the mean willngas-to-pay bid for monthlytaiff for a standardpiped water system with houseconnecions was

ProgramReport Series 81 Rs 40 (tldI.was in additionto the one-timecosts of a iaty Rs 600 equiredfor conetig to th system) summaon of the househo wiligness-to-pay bids ylds a monthly total of Rs 22,500. This figureis essatially the sameas the amout houdds are alreadyspending on water (Rs 23,900). This oqnceases the confidencein the cedibility of the willingness-o- pay bid. 15.19 In practce, it is not poible tecover the entirewllingness-to-pay amount because of the unfeasibilityof enforcig dieifentialtrin the samevill. Mh swveyresults dicate that in a vMageof the type bdag discussed,78 pecent of the householdswould connectto a sandad piped wat systemwith hoe connectionsif the monty taff wer set at Rs 25. In addio, thy wold bear the coss of conectig to h distuon line (Rs 500, appximately) and pay a coction fee to the PIHED(is 80). The estimatedreveue based on te numbes is Rs 11,300 per month(see Tible 1-7). If a piped watersystem is designedfor 78 peen of the popuation,tho totat montily cost to dte PHED would amount to Rs 13,400(Ris 10,00 capItl, Rs 2900 0M lTe estimates suggest that a vey substantil proporion (85 pect) of th total coss of a public piped watersystem can be rcovered underthe conditionsdesmbed above. 15.20 The comparablecomputations for the sweetwater zone ae shownin TableI-& Only 44 percet of the total costs of a piped water sysm are potnly reovable compued to 85 perent in the brackishwater zone. This Is the rest of much lowerWTP bids in the sweet waterzone, not of diffeens in tho cost the sysm in the two zones. Me WIP bids ae lower in the sweet wa zone because unlable public water sysms ae much less at_active when good quaity water i easly available from pivate andpmps. However householdsare wli to pay a substntial prmimumfr improved labflit, which tggesdhata higherpeenage of the costs cau be recoveredin the sweetwater zone if rehbility is impoved. 1521 WPlhinge to Pay in Villags wih P4xd WUIr. Ime willn to pay for piped water in aes with exstng supply syst Issmuch lowercompred to villages without piped war systems. TI is, perhaps,bause the tta ar not awareof howthe servie woud operatewhie the faomerhave sufficient cause to be dissatisfied. IThemean WI? bid In a villagewith piped water in the backish zone was only Rs 16 per monthcompard to Rs 40 per monthin the villagewithout pped water. However,when the opton of an iprved, more reiable senice was offed to the householdsin the vilage with piped water,te mean bid insed to Rs 33 per month while the percentg of houeholds willng to connectwent up to 95 percentfrom 75 percet. Tis agai undercoe s the premiumwhich householdsplace on sIxtemreliaty.

The Arid Zone 15.22 The situaon in the add zone of Nohen Pnjab is quite dffrent from the sweet water and bracksh water zonesof Cetral Punjab. In the arid ze at th pre tme the demond is for wateritself and not for water-basedamenities. Ihis is becauseof the scarcityof waterand the low base of serviceavailable to villages(public wells and surfa water). Houeholds spend an average of 5 hour per day to fulill their water needs duig the peak summer months compared to 4 hours per week in the brackishzm' and ev less in the sweetwater mon

'OAcontutoy teasoncould be the anchoringefft of the existg tff (12 pr month)which is kown to the respondentsin the vlg with piped water.

82 UlDP-World Bank Water and Sania* Pmpw 1523 The demand for privatesources s, however,quite srong; 24 percentof the households have installed pivate sources (e.g., wells or boreholes) whUleanother 5 percent have made an unsuccessfulattempt. This is despite the fact that the costs of installng private sources in the ard zone are much higher (approximatelyRs 8000 with a monthly maintenancecost of Rs 40). However, these private sources, wells or boreholes, cannot, i general, provide acce to higher service levels. The installationof private sources is not systematicallyrelated to socioenomic charcteistics of householdssince the very possibilityis based on the fortuitouscicumstance of an appropriatelocation of the house such that the groundwateris accessibleY 1524 For the above reasons,connection rates to piped systemsare almost 100 percent,much higher than in the sweet water and bracldsh water zones. This is so despite the fact that the arid zone is economicaly less developedthan the other two zones, the average village size is much smaller and the monthly tariff is much higher (compsaativeinformation for the three zones is provided in Tables J-9 and J-10). 1525 The official poicy of the PHED, of not providinghouse connectionsto villages with less than 5000 inhabitants,is not being followed in the arid zone. This has resulted in much higher capital costs per capita (approimately Rs 840) because of the small vilage sizes. The survey results indicate that while O&M costs can be fully recovered,it is not possible to recover the ca.tal costs to any significant extent. Dbeusdsn

15.26 CollectiuveWater Supplies. As long as the highest service demanded was the private handpump,the policy of leaving the sweet water area to be served by the private sector was justifiable. However, with a significant proportion of the population replicating a piped water system at the individual level, the policy is in need of fresh evaluation. As mentioned earlier, while the capital costs of a piped water system and private upgrading through the installationof electric pumps are comparable, the O&M costs of the latter are almost two and a half times the costs of the former. Pivate upgading is not likely to be the socially opdmal option. 15.27 Tlhealternative to individualupgrading is not necessarilystte run piped water supplies but other collectiveanrangements. Privatelybuilt and managedwater systems should not only increase community participation but should also prve to be less expensive to build, thus enhankcig the pospects of ful cost recovery. In this context the obse4vationsof Briscoe (1987) need to reiterated:

In government-builtwater supply prjecs the role of the private sector is much more problematic. In the Punjaband Sind the privatesector has a major role - ddlling and civil and mechanicalworks are all contracted out to the private sector. In principle this should make for an efficient sector. In practice, however, the situation appears to be quite different. It is widely acknowledged-- by governmentofficials and others knowledgeable about the sector - that there is exteisive collusion between the contractors and the governmentagencies, a relationshipbased on kickbacks to the governmentofficials, and resulting in construcion costs which are much higher than need be the case. In such a system there is an incentive for both officials and contactors to build over-designed, expensive systems. As an illustration of the

M'Thereis very little residential mobility in these vilages. Better-off households probably do not locate their houses in areas with easier access to ground water.

Program Report Seriss 83 inefficienciesin this sysem, the Oangi Pilot Seweage PMect in Karaci, it which the community,not the goment, finces the works,the costs of seweringhouses has beentordued by over 80 percentY Proct officialsattrbute the cost reducions to two factors - attention to reducing coss wherever possible,and eliminationof the contractor-officicorpd 15.28 Mor , it hasbeen faily wel establishedthat the tsition to higherservice levels (pipedconnections and electricpumps) leads to an increasein the use of water. Ihis was confirmed in the samplevillages wbere drainagehas emergedas a seriousproblem. Althoughit is possibleto upgradethe watersupply at an dividuallevel (even if socily inefficient)it is not possibleto resolve the rewltingdrainage problem on an individualbasis. Thus,a collectivesolution becomes unavoidable. 15.29 Reliabiili. Became of the poor quality of the groundwater,public piped water suppliesare being built i the brackishwater zone. In principle,because the amountbeing speniton privateupgading of servicesis so substantial,the dscssion regading private,collective water systems shouldbe equallyapplicable to the brackishzone. However,as mentionedearlier, piped supplies, irrespectiveof whetherthey are publicor pnvate,have to be muchmore reliableto be able to compete effectivelywith the individualupgrading of service. 1530 The srey suggeststhat the reliabilityof piped systemsis a cnrcial elementin achievingcost recovery. It is quiteclear that peopleare willingto pay significandymore for a reliable system. This is demonstratedby household investmentin multiplewater systems: handpumps,piped connectionsand electricmotos It was mentionedearier that in villageswith piped water in the brackishwater zone, almost30 percentof the householdsmaintained both electricmotors and piped connectaons,spending Rs 76 per monthas againstthe Rs 18 for the piped connectionalone. It is quite clear that for publicutlities to competeeffctively againstpmrate providers of handpumpsand electricmotois, their reliabilitymust be improved. 1531 In a hot and dry ag 4 ultLralarea such as Puonabit does not seem possiblefor reliabilityto be effectivelyincreased W..out the meteringof householdconnections. TIis is because the demandfor wate at zero margn cost (Le, unmeteredconnections) is immense. One findingof the surveyillustrtes this point welL In villageswith piped water supplieshouseholds with animals werewilling to pay m forconnectons. Investigtionsindicated that the reasonwas the convenience in washinganimal. Buffaloes,which need to be kept cool in the summer,could be hoseddown at home,tather than be takento the canalor the villagepond, thus savingon timeand supersion costs. Waterfor dgnkiag by animalscould also be providedthrough the domesticconnection. 1532 This findng highlightsan oversightwhich has sitgificantimplications for the design of ruml water supplysystems in Pakisu. The PHEIDdesigp criteria at presenttake into account humanneeds only and ue a consumptionbenchmark of 10 to 15 gallonsper capitaper day. However, if householdsuse the water to caterto animalneeds also, including water-intensive ones like washing, the designestimate could be easilyexceeded. It was estimatedin the surveythat while the minimum quantityof water requiredfor essenta humanconsumption vawied between 20 to 30 liters per capita per day, the minimumquantty equiredfor animalneeds varied between 40 and 60 liters per animal per day (estimatebased on the amountof waterwhich, in situationsof emerency, would haveto be fetchedfrom outside the housein the sweetand brackishzone and purchasedfom vendorsin the arid

5TIs figuremay be too high Ihe orginal costs of seweringa house wereabout Rs 2,500. The OnangiProject's cost was about Rs 800.

84 UNDP-WorldBwn Water ad Sana Prormn zone). Ihis means tnat a householdwith 8 membersand 2 cattle (the avenagein the arid zone) would need approximately200 liters for human consumpdonand 100 litets for animal needs. Tis 50 pecent icrease, which should hold for notmal, nonemergencyconsumpdon patrs also, is much more than the 5 percent to 10 perent margin added on by the PHED to total esimated human consumptionfor all other usages (public buildings, schools, mosques, animal etc.) Perhaps, this is one explanation for the ubiquitous problem of low pressure which charateriz ru water supply systems in Paistan 15.33 In such an enviroment, if connectionsare not metered, then water must inevitably be rationed by reducing the reliabilityof the system. If reliability is reduced, people must secure odter sources, and their domesticconnection becomes a back-up supply. Table A-12 povides eAidencefrom the sweet water zone. The majority of households stated that they would keep their handpumps operatonal either pemnanentlyor at least until such time as the piped water system could supply water with reguarity. In a vilage in whih the piped supply system had fallen into disuse, 12.5 percent of the households had dismantled their handpumpsat the time of obtaining the connection. Ihey were forced to reinstall them because of the poor performanceof the system. It is not surprising, therefore, that 94 percent of the householdsin this villaga indicateda resolve to keep their handpumpsoperational permanently even if the piped supply was restord. 1534 In such situationsthe willinness to pay for piped water is much reduced. The PHED then cannot collect te resources it needs to run the system efficienly. A vicious circle of system detioration and lowered willingnessto pay ensues. 1535 The concept of metering householdwater connectionsis quite acceptable in the ruri areas of Punjab. Tne percentage of households ftat favored metering vaied from 54 percent in the sweet water zone to 81 percent in the brackish water zone. The statisticl rsults also reveaed that such households were consistendy willing to pay more in tariff for piped water systems. 1536 The economicviability of public piped water systems requires enhanced tariffs. Households are willing to pay significanty higher rates (see Table J-11) but only if they are accompaniedby the imprved performanceof existingsystems. A packlge of tariff inreases tied to selective improvementswould be quite acceptable to the householdssureyed. This seems the only feasible way to move towards economic viability and to break the vicious circle mentionedearlier. 1537 Pubfic Stodpip Systems. Householdsprefer piped systms with domesticconnections and are willing to pay much higher rates than thought possible earlier (see Tables J-11, J-12, and J- 13). Even so, if such service is to be extended to small villages of 1000 to 500 people, as is being done in the arid zone, it would bhve to be subsidized to a considerableextent. 1538 It seems unlikely that the higher per capita capital expenditure can be recovered. In such situations it would be worthwhile to experiment with less expensv standpipe systems.

1539 Standpipesystems should be adequate in the smal villages of the arid zone since the overwhelmingdemand is for conveniet access to water and not for water-basedamenities Contray to the prevailingview about the acceptabilityof staadpipesystems, households in villageswithout piped water were willing to pay reasonablyhigh ra for such systems. The mean bid was Rs 35 per month and 68 percent of households would subsnbe to theses iceat a tuiff as high as Rs 25 per month (this tarff exceeds the present monthly tariff for piped systems with domesticconnections (Rs 20) in the arid zone).

PrWon ReportSerios 85 15.40 Ihe main drawbackof standpipe systemsas they have been organizedthus far relates to the problems of revenue collection owing to difficulties in monitoring actual usage. Some experimentationon a pilot basis with alternative structires is certainy warranted. For instance, a system based on manned kiosks shouldovercome the problem and require no more manpowerthan the present structure. Since piped systemsoperate for four hours a day at most, the existing full-time staff is under-utilized. The option of gving the kiosks on private contract could also be explored. Even if households purdase only 100 liters of water per day (about the average essential daily consumption for family needs in the winter) at Rs 0.10 per can of 16 liters (one tenth the existing price of vendeu water), the revenues realied per householdwould be of the order of Rs 20 per month. These would be sufficientto recover O&M costs and some part of the capital costs as well.

1SA4 Organztonal Issues. Any serious attempt to encourage communitiesprivately to consttuct and manage water supply systems would involve difficult orgaaional issues. Experience has proved that collective resposes to such needs do not arse by themselves but need patient groundwork.the availabilityof external technical expertise and accessible credit facilities. 15.42 The experienceof the successful Otangi Pilot Sewerage Prect in Karachi bears out the above point, even though the projectrequired only an intermediatelevel of collectiveeffort between the completelyindividual and the completelycollective. In a sewerageproect, the lane could function as the organizationalunit, and a successfuldemonstration could induce other lanes to participate. This is not possible in a water supply project where a much larger proportionof the populationhas to reach an understandingbefore the project can be initiated. 15.43 The orgaizational problem is made more difficult by the fact that, contrary to the popular presumptionin developmentcircles, the villages were not willing to assume responsibilit- the provisionof water. Over 65 percent of householdsin all three environmentalzones believed tuat a water supply system would be best managedby a governmentagency in preferenceto local politcal bodies, village water committees,or pnvate entrepreneurs. Tis is perhaps a reflection of the sharp clan and political divisions that charactize villages (especally those with sizeable populations) in Pfjab. These divsions affected the peweived effectivenessor fairness of elected political bodies or M'lagecommittees to mnage collective systems. It was considered a lesser evil to leave the system to a neutralagency, external to local vibage politics,even though it was inefficient Perhaps a publicly provided but privately managed water utility would be just as acceptableand more efficient, although this alternative has not been experimented with hi Pakdstan. The concept was not familiar to respondentswho remain wary of the lack of accountabilityof the privatesector and the general failure of governmentattempts at its regulation in many other spheres of activity. 15.44 At the same time, it is clear that the solution does not lie in raising more revenues and giving them to the PHED as it is structured at present. ITe PHED is primarily an engineering agency and is not equipped to involve the communityin decision-makingor even to carry out the tasks of revenue accounting and collecton in a satisfactorymanner. In this, it is handicapped by the fact that in principle the PHED is supposed to hand over the managementof the systems to local councils after an initial period of two yeats. However, the latter have neither access to the technical experdse required for maintenanceand operations nor the political unity to impose effective revenue collection in the factional villages. As a result, the systems are inevitably handed back to the PHED after a period of mismanagement. The consequenceof the lack of a clear policy in this regard has been that neither body has been equipped to the degree necessary to manage rural water supply systems in an effective manner. A resolution of this situation should be the first step in the reform of the ruml water supply sector.

86 UNDP-WorldBozk Waterand SaoniationProgram References

8dscoe, J. astn: Secor Waokand Willkng to Pay for WaterStdy." 'World Bak Waterand Urban DevelopmentDeprment, Apil 26, 1987. Mem nm. Brbooe, J. and D. De FerrantL Waterfor Rra Cwawte: He!pingPeople Help hemselves. WashingIon,DC: Mm Wodd Bank, 1988. Cummings,R G., D. S. Boolbite, and W. D. Schulze,(edits). ValuingEn,wimetata Goods. An Asss_m of the Contigen VabladonMedsoi Totowa, New Jersey: Rowman and Alanheld, 1986. Govnment of Pakist PopublatonCensus, Distict CensusReports for Sheiupra, Faisdabad,and RawalpindiDistr 1981. Whington, Dalo,Mar Muiwahuz,Gad McMahon,and Ky_on Ch Widlingnessto Payfor Water bi Newala DisFC T _.nia Sawtegiesfbr Cost Recoy. HeddReport No. 246 Washington,DO USAIDWater and Sanitationfor HealthPruiect, June 1989,pp. 205. Whitington,D., et aL 'Estimatingthe Wlingness to Pay for WaterSerices in DevelopingCounties: A Case Study of the Use of ValuationSurveys in SouthemHaiti.' Econaik Devwopmentand Cuwl Cange 38, no. 2 (January1990): 293-312. Wodd Bank. -WMins to Pay for Water in Rural Areas. Research proposa, Water and Urban DevelopmentDepatment, March 1987.

Prgr= Rpot Seies 87 Appendixes Appendix A

ThbleA.1 Vilage Profiles Ghazi Bhaddroo Kharian Mirza Jandiala Minara Minva Wala Vikan Sher Khan Area* (acres) 2087 1-44 4852 3S60 4868 Population* 4514 2173 89g0 5630 7581 Householdsize* 6.7 6.2 6.7 6.9 6.6 (fromsample, 1988) (9) (9) (10) (10) (9) Literacyratio* 28.6 17.6 22.7 14.0 24.0 Males-school age 8.9 5.4 6.0 4.7 9.5 or older (%) Females-school age 1.7 i.9 0.8 0.4 2.6 or older (%) Distancefrom district Headquarters(km) 2 2 18 10 14

On main Off On mn Off On road seny road secondoy secondary Road locaion road road road Prwsenceof facilities Electricity Yes Yes Yes Yes Yes Telephone Yes No Yes No Yes Medical No No Hospital Clinic Hospital Male education Middle Middle High High College school school school school Fmale education Middle Middle Mididdl e ddle College school school school school Bank No No Yes No Yes Union councilmembers from village 5 2 4 5 9 Districtcouncil membersfromvillage 0 0 1 2 1

* Figures from 1981population census. AHl others fom field visit

Prorm ReportSeries 91 Tabi A.2 VMagePofes Ohazi Bhaddroo Kharan Mrza adala Minara Minara Wala Virm Sher Khan Householdsize 9 9 10 10 9 Adult womenin household(%) 28 24 24 25 24 Childrenin household(%) 34 35 41 38 38 Yearsof educadonof most educated Memberof household 9 7 9 7 8 Wom inlhousehold 4 4 2 2 4 Essentialwater consumion 24 27 23 17 20 (litescapita/day) Householdsinvolved in 21 29 19 42 26 faing (%) Householdswith land or 54 61 73 76 55 other property(%) Constructionvalue of house 80 65 98 51 104 (000 Rs) Householdswidh external 43 23 31 29 37 exposure(%) Householdsthat believewat 64 48 22 44 50 shouldbe suppliedfree (%) Householdsthat believewate 64 35 29 34 39 can be suppliedfree (%) Housholdsthatbelievewater 54 71 71 78 72 supplysystems should be managd by PliED (%) Householdsthat favor 75 68 55 60 42 meting of water (%) Note: All staistics are derived fom the sampleobservaos.

92 UNDP-WarldBank Water ad Sanition P,i= Table A.3 Prfles by Villge and HousehoId Type VillageType HouseholdType A BI B2 Al A2 Bll B12 B2 Householdsize 9 10 9 11 8 10 10 9 Adultwomen in household(%) 24 24 26 24 24 25 22 26 Childrenin household(%) 38 41 34 42 34 40 44 34 Yearsof educationof most educated Memberof household 8 9 8 9 6 9 8 8 Womanin household 4 4 3 5 3 5 4 3 Essentialw'e-r consumption 19 23 26 20 17 22 25 26 (liters/capita/day) Houswholdsinvolved in 31 19 25 29 34 20 17 25 faning (%) Householdswith land or 61 72 58 68 52 75 67 58 other prperty (%) Constuction value of house 87 97 73 110 57 113 60 73 (000 Rs) Householdswith external 35 31 31 37 32 37 17 31 expose () Householdsdat believe water 48 22 56 44 52 23 22 56 shouldbe suppliedfree (%) Householdsthat beLHevewater 38 29 49 33 43 30 28 49 can be suppliedfroe (%) Hu seholds that believewater 74 71 63 73 74 73 67 63 supplysystems should be maed by PHED (%) Householdsthat favor 47 55 71 47 48 63 39 71 meting of water (%) Not: All staistcs are denved from the sampleobservations.

ProhramRpt Series 93 Table AA Patter of Housebhd Choice over AvailableServie Optlons and ApproimdnatePrivat costs AvailableService Options Manual Motodized Domestic Domestic Domestic Handpunp Hanpump Connecion Connecion Connection Only Only* Only + Manual +Motorized VillageType Handpump Handpump Type A vlhpgehouseholds: Number 59 6 0 69 10 Percentage 41.0 42 0 47.9 6.9 Type B village households: Number 82 35 NA** NA NA Percentage 70.1 29.9 NA NA NA Costs**** (Rs)

Capital Handpmp 1,00 100 1,000 1,000 Motor 2,000 2,000 Overheadtank 500 500 Comnecioncost 500 500 SOO Total 1,000 3,500 500 1,500 4,000 O & M (per month) HEndpump 6 6 6 6 motor 16 16 Elcticity NA*** NA*** Tariff 10 10 10 Total 6 22+ 10 16 32+

* A motorizedhandpump can be used manuallyin case of power failures. ** NA indicatessevice level not avaible. Domesticconnections are not availablein Type B villages. ** Rerpondentswere uwbketo estimae the electricity cnares attibutable to opeting the motur. **** Capitalcosts wieapproximate values in currentpices; O&Mcosts are obtainedfrom sample respoe.

94 aNDP-Wrd 20k Water ed Swuttc Imgrm Table A.S Reasons Mentoned by Households for Obtaininga DomesticConnecton to Piped WateeSystem % cf Households ReasonNo.1 ReasonNo2 Convenice 60.9 42.1 Betterquality 20.9 29 Reliability C.2 10.5 Othe reasons 10 18A

Table A&f Reasoa `--stioned by Households for Not i-elnl~ng a DoetcConnection - TypeA Villag % of Househols ReasonNo.1 ReasonNo.2 ReasonNo.3 Expense 57.8 44 0 No need 21.9 44 50 Not reliable 17.2 8 50 To avoid iLlegaltransactions 3.1 4 0 Poorquality 0 4 0

P~~ogiwaRpaR t Swai 95 TableA.7 Relatlouslp between Economic Standing and Choice of Service Level - Type A Vilage Choice of ServiceLevel AverageValue* of House(Rs) Manualhandpump only 48,473 Manualhandpump and 96,174 domesticconnection Motorizedhandpump only 137,500 Motorizedhandpump and 208,500 domesticconnection

* The monetaryamount that would be needed to reconstructthe type of house the respondentis livingin is used as an indicatorof economic standing.The averagevalue of this indicatorover all householdsin the Type A village is Rs 87.229.

Table A.8 Level of Satisfaction with Piped Water Supply System- Type AVile Al A2 Households Households Total Levelof Satisfaction No. % No. % No. % Satisfied 48 60.8 44 67.7 92 63.9 Dissatisfied 31 39.2 19 29.2 50 34.7 No response 0 0 2 3.1 2 1.4 Total 79 100 65 100 144 100

96 UNDP-WorldBank Waterand Sanian Pn sm TableA.9 MajorShorteomngs of PipedSupply System Mentioned by Households - Tpe A Vil,a_e % of Households Shorcoming# 1 Shortcoming# 2 All Connected AR Connected Households Households Households Households Reliability* 38.5 38.7 19.2 22.6 Maintenance/design** 28.8 25.8 13.5 16.1 Interactionwith staff** 25.0 32.3 17.3 25.8 No response 7.7 3.2 50.0 35.5

* Includesinsufficient pressure, frequentfailures, and power breakdowns. ** Includessubstendard matrial, poor distributionlayout, and lack of cleanliness. ***' Includesirresponsible, uncooperative, and corupt staff.

Table A.1O Most Important Improvement in Piped Supply System Desired by Households- Type A Village Characteristic % of Households* Greater pressure 31.6 Cleanliness 10.1 Increasedsupply 7.6 Laer storagetank 6.3 Use of standardmaterial 6.3 Reliability 5.0 More competentstaff 3.8 Lowerprice 2.5 Improvedbilling procedure 1.3 Satisfied 11.4 No response 13.9

* Permentagesdo not add up to ICObecause of rounding.

Progran Report SrM, 97 able A11 Household'sInforWalndilelks Regarding C Pp W and Alternative Supply SystUMS

Tyeof HwsidCh I- Al A2 B11* B12 B2**

Priccharged(%respondingyes) 98.7 - 100 - - Billingfrequency (no. of dmes per yea) Mea 12 - 9 - - Standarddeviatin 1.9 - 4 - - Exisg montly iff R) Mean 10 10 10 9.8 15.8 Stndard deviation 0 0.1 1.8 1.1 16.5 Households pding (%) 97.4 76.9 100 61.1 6.8 Totalno. of hoursof water supplyper day Mean 9 9 9.6 7.6 3.7 Standarddeviation 6.3 5.1 4.5 53 3 Householdswit4ho4bunuayorless (%)30.6 23.3 10 313 912 Numberof hous of additional supplyrequired to meet needs Mew 9.1 - - - - SUd deviatior 6.3 - - - - Frequencyof supplyhf{llUa*** mean 1.65 - 0.35 - - Stndard ddeviaton 1.50 - 0.02 - - Pipedwater (% respong yes) Satisfactionwithtaste 100 - 975 - - Sasaction with clean s 97.4 - 9S - - Satisfactiowith bygiene 98.7 - 95 - - Altonatve water(% respond yes) Satdsfationwith tae 89.7 100 87.S 88.9 96.6 Satisfaton withclanliness 97.4 100 95 100 100 Satisfactonwith hygiene 97.4 96.9 80 94.4 94.9 * Infodmto regardin pipedwater petisto periodwhen sceme wasopeationaL Wamnonatonregading piped wate pertainsto estMWaes. ** Reguar supply=0, < oncea week.1, oncea week2. > oncea weekw3. Other= 4; indexcntucd by smmin and taing men

98 UNDp-Wor Bmk Water and Saai P _qwu ThblA.12 Numberof Yea Househods Would Keep mandpumpsOperaiond No Piped Water Supply Syem Fxisted In Thir Vla Vi ype Respon(%) A Bl B2 Until piped supplyis regular 52.1 S.6 25.4 Foreve 5.6 94.4 69.5 Forupto6yeans 0 0 5.1 Dismantleimmediately 1.4 0 0 Noresponse 41 0 0 bide AM3 Housebold Preferene Regding Respnst for Operatic and M _ent of Piped Water Supply Sysms Comected Unconnected All Managme Option Households(%) Households(%) Households(%) PHED (goveam agenc) 73.1 68.3 70.5 Local councils(elected) politicalbody) 11.8 9.9 10.7 Vllage commI 7.6 9.9 9.8 Private enteprn 4.2 4.2 4.2 No prerence, 3.3 7.7 5.7

PAVra Reo Series 99 Appendix B

Adusted W _lgm-sto.PayBads 1. In the statisticalanalysis of the factom affecdng the williHgnessto pay for improved serviamsthe dependent variable is the wdllHgn-to-pay bidc Bocause of some incompletenessand inconsistencyin the desig of the bidding games cerain adjustents have had to be made to a few of the bids elicited fom the respondents. These adjustmentsare explained below. 2. When a respondent said 'no" to the lowest bid value in a bidding game without having said 'yes' at any stage it remained unclear whether his bid should be treated as a protest bid or a genuine zero bid or whether it ought to be assigped a value between zero and the lowest bid value. nuis issue was of some significancebemause in some games the lowest bid value was Rs 15 which is higher than the existing tariff of Rs 10 per month for a domestic conection. 3. TI problem was resolved by relying on the respnnse received to another question The respondentshad been asked to indicate the characteristicsof what they considered to be an ideal piped water system and to Indicate w-at monthly tariff they would be wiling to pay if such a system were made available. The following rule was adopted for adjusting the bid teceived in the bidding game (only for those who did not say 'yes" at any stage in the bidding game but said "no" at Rs 15 in the bidding game).

Bid Offered for Ideal Sysiem Adjusted Bid Value for System Offred in Bidding Game

No Response Protest Bid

Zero Genuine Zero Bid Rs 1- 5 Response laced in interval Rs 0-5 Rs 6-10 Response placed in interval Rs 5-10

Greater than or Response paced in interval Rs 10-15 equal to Rs 11 Simiar adjustmentswere made for those bidding games where the lowest value in the bidding game was Rs 5 per month. 4. In one bidding game (StandardSystem Village B2) there was an inconsistencybetween the high and low starting point versions. Ihe low starting point version did not contain a bid value of Rs 15 per month so that the bid inteval was Rs 10-20, whereas the high sarting point version contaied both intervals Rs 10-15 and Rs 15-20. The two versions were made consistent by allocating a bid received in Rs 10-20 interval in the low startng point version to either the Rs 10-15 interval oi the Rs 15-20 interval based on the value offered for the ideal system. If the latter was greater than Rs 15, the bid was placed in the Rs 15-20 interval. Othewise it was placed in the Rs 10-15 interval.

100 UNDP-WorldBank Waer and Sankadon Progr Table B.1 Decsion to Install an lectric Motor. Results of a Logit Model Dependent V kble:Probability that a householdwould install an electricmotor Pramet Parameter Mean IndependentVaiiables Estmat T-Ratio Estimate T-Ratio Values Intercept -4.71 -2.74** -4.87 -5.40** Householdsize -4.28E-l -0.62 9.23 Waterconsumption 0.34E-2 -0.54 21.62 Proportionof adult women 0.43 0.20 0.26 Proportionof children 0.35 0.24 0.41 Constuton value of house 0.67E-5 3.12** 0.65E-5 3.26** 84,432 Ownershipof land or property (1 if yes) 1.43 2.46** 1.25 2.43** 0.63 Age 0.14E-1 -0.81 50.29 Education 0.34 4.25** 0.33 4.31** 8.26 Occupation(I if faming) -0.21 -0.42 0.27 Meter (1 if yes) -0.78 -1.75* -0.78 -1.82* 0.55 Free supply(1 if yes) 0.48 1.07 0.34 1.83 0.43 Externalexposure (1 if yes) 0.41 0.95 0.33 Householddummy A -1.12 -1.72* -1.23 -2.04** 0.56 Al -1.39 -1.86* -1.22 -1.74* 0.32 sli -0.86 -1.39 -0.93 -1.65* 0.16 B12 -0.87E-1 -0.11 0.07 Numberof observations 244 244 Log-likelihood -80.46 -82.22 Restrictedlog-likelihood -119.56 -119.56 Chi-squa (16) 78.20 (8) 74.68 Significancelevel 0.201E111 0.32E-13 Proportionof orrect predictions 0.86 0.86

Prwn Rep Series 101 Table D2 Deeision to Conect to a Piped Watw Supply System: Result oftaLogit Model Depen&datVariabk.: Probablity that a householdwould onnect to a piped water supplysystm Pa_ameter Parameter Mean IndependSetVariables Estimate T-Ratio Estimate T-Ratio Values Intercept -2.12 -1.59 -2.25 -3.10** Householdsize 0.75E-1 1.38 0.68E-1 1.35 9.29 Waterconsunption 0.23E-1 1.90* 0.24E-1 2.02"* 19.00 Proportionof adult women 1.23 0.65 0.25 Proportionof chldren 2.11 1.67* 1.86 1.88* 0.39 Constuction value of hse 0.46B-S 1.48 0.46E-5 1.56 84,394 Ownershp of land or property (I if Yes) 0.35 0.78 0.61 Private water source (1 ff Motor) -0.98 -1.37 -0.99 -1.39 0.12 Age -0.70E-2 -0.48 49.80 Education 0.13 2.27** 0.13 2.44** 7.91 Occupation(I if fiaring) -0.84 -1.81* -0.72 -1.65* 0.31 Meter (1 if yes) -0.25 -0.58 0.49 Free Supply(1 if yes) -0.26 -0.62 0.48 ExtenadExposwe(lIf yes) -0.40 -0.85 0.34 Distace fiom distibuton ine -0.26E-1 -2.16** -0.2S6I -2.26** 13.01 Numberof observations 137 137 Log-likelihood -75.52 -76.77 Resetred log-ikelihood -93.90 -93.90 Chi-sque (14) 36.77 (8) 34Xi6 Significancelevel 0.80E-3 0361E4 Proportionof corect predictions 0.74 0.74

102 uNDP-Warld Bank Waw ae SanaidaionP)igrasn TbleB1.3 Stadt WaAnaysis ofWMlngnm to Pay for a Standard iped Water Supy Systm: Reslts of Ordaylus Squares 1reu Mode DependentVoiabke: Mdpoint of intervalin biddhngame withn whichrespondent's WTP bid falls Pa;M-ster Parmeter Mean IndependentVariables Estmate T-Ratio Estimate T-Ratio Values Itercept 2131 2.40** 20.06 ,.40** Householdsize 0.40 1.29 0.41 1.58 9.13 Waterconsumption 0.03 0.78 25.04 Propordonofadultwomen -11.81 -0.98 0.26 Profptoof chidien 1.01 0.13 0.42 Constuctionvalue of house -3.162-6 -0.44 84,524 Ownersip of lnd or proety (1 if yes) 4.62 1.75* 5.15 2.0841* 0.64 Privatewater soure (1 if motor) 7.11 2.52** 6.76 2.56** 0.29

A5ge . 3. 2.33I* -0.27 -3.28'* 50.84 Education 059 1.82* 0.49 1.63 8.67 Occupaon (I f farmning) -4.02 -1.44 -4.27 -1.65 0.23 Meer (1 if yes) 5.18 2.11** 5.78 2.56** 0.63 Free supply(I if yes) -2.8 -0.89 0.38 Extexna exposure(I if yes) 0.55 0.23 033 Householddummy Bli -6.02 -1.98* -5.49 -2.03** 0.36 B12 -1.88 -0.56 0.16 Staing point dummy (1 if low) -6.19 -2.22** -6.94 -2.60** 0.34 Numberof observadons 106 106 Mkanofdependentvaribbe 18.16 18.16 Sta. dev.of dependentvariale 12.52 12.52 F-vaIue (16,89) 3.31 (9,96) 5.83 Significanceof F-st 0.00 0.00 R-squared 037 0.35 Adjustd R-squed 0.26 0.29

P1r_m Rpot Seri 103 abbleBA Statisical Anaysis of Willgness to Pay for an Improved Piped Water Supply System by Households tn Villageswith an xisting Piped Water Supply System: Result of Ordinary Least Squares RegressionModd DependentVwab: Midpointof intervalin biddinggame withinwhich respondent's WTPbid falls Patameter Parametr Mean lndependentVariables Estitnaie T-Ratio Estimate T-Ralio Values Intercept 4.56 0.62 2.10 0.44 Householdsize -0.19 -0.85 -0.21 -1.07 9.82 Waterconsumption -0.08 -1.26 -0.07 -1.23 19.37 Proportionof adult women 14.69 1.54 15.33 1.68* 0.26 Proportionof children 6.54 1.00 7.86 1.31 0.41 Constructionvalue of house 1.75E-5 135 1.93E-5 1.65 90,669 Ownerhip of land or property (1 if yes) -0.68 -0.29 0.62 Privatewater source (1 if nrtor) 5.4Z 1.59 5.19 1.68* 0.14 Age -0.04 -0.54 49.57 Education 0.02 0.05 8.22 Occupation(I if fanming) 3.63 1.55 3.18 1.49 0.33 Meter (I if yes) 5.61 2.61** 5.87 2.90** 0.53 t'iee supply(1 if yes) -0.88 -0.40 0.48 Externalexposue (I if yes) 3.06 1.23 3.15 1.34 0.31 Satisfaction(1 if yes) 0.16 0.07 0.62 Householddwmny Al 1.34 0.60 0.59 Strting point dummy (1 if law) 0.18 0.09 0.51 Numberof observations 118 118 Meanof dependentvariable 13A1 13.41 StaL dev. of dependentvaiable 11.27 11.27 P-value (16,101) 1.57 (9,108) 2.84 Significanceof r-tes 0.09 0.00 R-squared 0.20 0.19 Adjused R-squaed 0.07 0.12

104 UNDP-WorldBank Waterand Sanitatin Prgrain Appendix C

Table C.A Distibution of W lP Bids for a Standard Piped Water System In Villages witheut an Operational Piped Water Supply HouseholdGroup Mean Bid* Bil B12 B2 (Rs) No. % No. % No. % 0 1 2.5 0 0 2 3.4 2.5 0 0 1 5.6 0 0 7.5 10 25.0 2 11.1 1^ 20.3 12.5 5 12.5 4 22.2 1I 28.8 17.5 14 35.0 8 44.4 5 8.5 22.5 6 15.0 2 11.1 9 15.3 27.5 0 0 0 0 1 1.7 32.5 1 2.5 0 0 1 1.7 37.5 0 0 0 0 0 0 45.0 0 0 0 0 0 0 50 2 5.0 1 5.6 11 18.6 No response 1 2.5 0 0 1 1.7 Total 40 100.0 18 100.0 59 100.0 Validresponse 39 97.5 18 100.0 58 98.3 No. of bids 28 70.0 15 83.3 44 74.6 >Rs 10 Mean Bid*** 16.67 16.81 20.73 (Rs) Mean of Bids 20.54 19.00 25.28 >Rs 10 (Rs) Note:The followingapplies to TablesC-I to C-3. * Mean bids are the midpointsof the intervalsin which the respondent'sbids fell (except0 and 50). ** Computedover the numberof total responses. *4w' Computedover the numberof valid responses.

Program Report Series 105 Tlble Ca Disibutdon of WTP Bidsfor a Standard Piped Water System with a Flexible Fiancing Arrngem t In Villageswithout an Operational Piped Water Supply HouseholdGroup MeanBid B12 B2 (Rs) No. % No. % 0 0 0 4 6.8 2.5 0 0 7 11.9 7.5 4 22.2 2 3.4 12.5 0 0 4 6.8 17.5 8 44.4 12 20.3 22.5 1 5.6 7 11.9 27.5 2 11.1 2 3.4 32.5 0 0 2 3.4 37.5 1 5.6 4 6.8 45.0 0 0 0 0 50 1 5.6 13 22.0 No response 1 5.6 2 3.4 Total 18 100.0 59 100.0 Validresponse 17 94.4 57 96.6 No. of bids 13 72.2 44 74.6 >Rs 10 Meanbid (Rs) 19.71 24.04 Meanof bids 23.46 30.40 >Rs 10(Rs)

106 UNDP-WorldBa* Wat ad and.Sat Program Tlhe C.3 Distribution of WTP Bids for an Improved Piped Water Supply In ViUages with an Operational Piped Water Sysem HouseholdCroup MeanBid Al A2 A2* (Rz) No. No. No. % 7 8.9 9 13.8 7 10.8 1 1.3 3 4.6 3 4.6 ,.5 23 29.1 12 18.5 9 13.8 ,1.5 4 5.1 8 12.3 4 6.1 17.5 27 34.2 13 20.0 21 32.3 22.5 5 6.3 6 9.2 4 6.1 27.5 0 0 0 0 0 0 32.5 0 0 1 1.5 2 3.1 37.5 2 2.5 0 0 0 0 45.0 0 0 0 0 2 3.1 50 2 2.5 1 1.5 1 1.5 No response 8 10.1 12 18.5 12 18.5 Total 79 100.0 65 100.0 65 100.0 Validresponse 71 89.9 53 81.5 53 81.5 No. of bids 40 50.6 29 44.6 34 52.3 > Rs 10 Mean bid (Rs) 13.87** 12.12 14.86 Meanof bids 20.25 18.79 20.96 >Rs 10 MRs)

* WTP bids for a standardsystem with a flexiblefinancing arrangement ** Mean bid would be Rs 15.19 if it is assumedthat nobody at present connectedat a tariff of Rs 10 per month would disconnectif the servicelevel is improvedat the same tariff.

PraM ReportSeries 107 TableCA Connecti Frequencles and Esmated Revenues: Provison of a Standard Piped WaterSystem in a Type B1 Viae Monthly Households EstimatedRevenue Tariff (Rs) Conmcted(%)* (Rs100 households) 0 100 0 2.5 96.6 242 7.5 94.8 711 12.5 74.1 926 17.5 58.6 1026 22.5 20.7 466 27.5 6.9 190 325 6.9 224 37.5 5.2 195 45 5.2 234 50 5.2 260 - The overal connectionfrequency for the Type BI vilUageis derived frm the responses of Bil and B12 households (Table C.1) based on their respective proportionsin the sample, .69 and .31.

Table C5 Connectio Frequenciesand Estimed Revenues: Provison of a Standard Piped Water System in a Type B? Vilage WlthoutFlexible Financing With FlexibleFinancing Estimated Estimated Moently Households Revenue Households Revenue Taiff Connected (Rs/100 Connected (Es/lOG (Ms) (%) households) (%) households) 0 100 0 100 0 2.5 94.9 237 89.8 225 7.5 94.9 712 77.9 584 12.5 74.6 933 74.5 931 17.5 45.8 802 67.7 1,185 22.5 37.3 839 47.4 1,067 27.5 22.0 605 35.5 976 32.5 20.3 660 32.1 1,043 37.5 18.6 698 28.7 1,076 45 18.6 837 21.9 986 SO 18.6 930 21.9 1,096

Joe UNDP-WoddBank Waterand San n Proram Tabk C6 ConnectionFrequencies and EstimatedRevenues: Providon of an ImprovedPiped Water SystemIn a TypeA Vlglage Monthly Households EstmatedRevenue Tariff(Rs) Connected(%)* (Rs/lOOhouseholds) 0 100.0 0 2.5 75.0 188 7.5 72.2 542 12.5 47.9 599 17.5 39.6 693 22.5 11.8 266 27.5 4.1 113 32.5 4.1 133 37.5 3.5 131 45.0 2.1 95 50 2.1 105

3 The overallconnection frequency for the TypeA villageis derivedfron the responsesof Al and A2 households(Table C.3) based on their respective proportionsin the sample,.55 and .45. TableC.7 Connectio Frequenciesand EsdimatedRevenues: Responses to Different Optons by UnconnectedHouseholds in a Villagewith an Operatil Piped Water Supply StandardSystem ImprovedSystem with FlexibleFinancing Estimated Estimated Monthly Households Revenues Households Revenues Tariff connected (Rs/100 Connected (RswlOO (Rs) (%) households) (*) households) 0 100 0 100 0 2.5 67.7 169 70.7 177 7.5 63.1 473 66.1 496 12.5 44.6 558 52.3 654 17.5 32.3 565 46.2 809 22.5 12.3 277 13.9 313 27.5 3.1 85 7.8 215 32.5 3.1 101 7.8 254 37.5 1.6 60 4.7 176 45 1.6 72 4.7 212 50 1.6 80 1.6 80

Prgrm Rewt Series 109 Appendix D

Tabb D.1 vuhp ProM.. Gatwala Santpura Bhaiwala Aklgarfh Manawala Sudhar Are (acrw) 1,899 1,999 2,342 2,384 - 2,121 Populaton* 6,181 7,339 11,049 7,885 20,586 8,333 Householdsize 6.7 6.8 6.5 6.5 7.1 6.5 (from sample, 1988) (10) (8) (9) (9) (9) (9) Literay ratio 38.6 32.3 32.8 30.1 58.5 19.7 Males-School age 10.4 8.4 11.6 9.5 24 3.9 and above(%) Females-School age 3.1 1.4 2 2.6 14.1 0.6 and above(%) Presenceof facilities Electdcity Yes Yes Yes Yes Yes Yes

* Figs from 1981populaton census.All ots from field visit.

110 UNDP-Worl Rank WaWerand Sniao Pp Tble D2 VillageProfiles Gatwala Santpura Bhaiwala Akalgarh Manawala Sudhar Householdsize 10.0 8.1 8.6 9.2 8.9 8.7 Adultwomen in housold (%)28 28 28 28 28 29 Childrenin household(%) 35 37 37 34 40 40 Yearsof educationof most educated memberof household 10 8 9 10 8 7 womanin household 6 4 5 6 4 2 Essentialwater consumption(Liters/ capita/day) 30 35 32 29 31 31 Householdsinvolved in farming(%) 22 27 19 20 9 18 Householdsowning ninals (%) 42 50 52 49 28 52 Waterconsumption of animas (liters/ animalJday) 62 63 63 50 76 35 Householdswith land or oier property(%) 63 63 67 60 35 38 Constructionvalue of house(000 Rs) 200 144 143 149 125 90 Householdswith external exposure(%) 46 48 32 52 25 28 Householdsthat believe water shouldbe supplied free (%) 59 52 53 54 47 52 Householdsthat believe water can be supplied fre (%) 27 29 33 24 19 31 Householdsthat believe water supplysystems should be managedby PHED (%) 61 75 62 62 72 67 Householdsthat favor meteringof water (%) 85 87 93 77 75 78 Note: All statisticsare derived from the sampleobservations.

Pror ReportSeri 111 TableD3 ProflHesby Vllage and HouseholdType VillageType HouseholdType A BI B2 Al A2 Householdsize 9 9 9 9 8 Adult womenin household(%) 28 28 28 29 28 Childrenin household(%) 40 35 36 39 42 Yearsof educationof most educated: Memberof household 7 9 9 8 6 Womanin household 3 5 5 4 2 Essentialwater consumption (liters/capita/day) 31 31 33 31 31 Householdsinvolved in fanming(%) 13 19 25 12 18 Householdsowning animals (%) 40 51 46 37 47 Waterconsumpton of animals (liters/animal/day) 48 57 63 53 35 Householdswith land or other property(%) 37 64 63 40 27 Constructionvalue of house (O0 Rs) 108 146 169 121 69 Householdswith external exposure (%) 27 43 47 28 21 Householdsthat believe water should be suppliedfree (%) 50 53 55 46 59 Householdsthat believe water can be suppliedfree (%) 25 28 28 21 37 Householdsthat believewater supply systemsshould be managedby PHED (%) 69 62 69 70 69 Householdsthat favormetering of water (%) 76 85 86 80 65 Note: All statisticsare derived from the sampleobservations.

112 UNDP-WorldBak Waterand SanitationPogram Table D4 Pattern of Household Choke over AvailableService Opdons and Approximate Private Costs ServiceOptions Domestic Domestic Manual MotDrized Domestic Connection Connection Handpump Handpump Connecticn + Manual + Motorized Villagelype Only Only* Only Handpump Handpump Type A Villagehouseholds Number 42 8 13 80 58 Pecentage 20.8 4.0 6.4 39.6 28.7 Type B Villagehouseholds Number 112 181 NA** NA NA Percentage 38.2 61.8 NA NA NA Costs****(Rs.)

Cti .andpump 1,000 1,000 1,000 1,000 Vtotor 2,000 2,000 Overheadtank 500 500 Connectioncost 500 500 500 Total 1.000 3,500 500 1,500 4,000 3 & M (permonth) landpump 3.5 3.5 3.5 3.5 Motor 9.5 9.5 Electricity NA*** NA*** Tariff 12 12 12 Total 3.5 13.0+ 12 15.5 25.0+

* A motorizedhandpump can be used manuallyin case of power failures. ** NA indicatesservice level not available.Domestic connections are not availablein TypeB villages. *** Respondentswere unable to estimate the electricity charges attributableto operating the motor. **** Capital costs are approximatevalues in curent prices; O&Mcosts are obtainedfrom sample responses.

Progam Rot Sriat 113 ThbIeD.5 ReasonsMentioned by Households for Obtaining a Domestc Connecion-Type A Vilae % of Households ReasonNo.1 ReasonNo2 Healthconsiderations 74.8 14.9 Cleanwater 11.9 39.6 Convenience 9.3 32.1 Otherreasons 4 13.3

TIbleD.6 ReasonsMentioned by Households for Not Obtainlg a DomesticConnecUlok-Tpe A Village % of Households ReasonNo.1 ReasonNo.2 Highcost 74.5 7.8 Hightariff 3.9 39.2 Low pressure 9.8 2 Low reliability 0 3.9 Noneed 11.8 0 No response 0 47.1

114 UNDP-WorldRank Watff and SantatiPronP r TableD.7 Relationshipbetween Economic Staudig and Choce of Service Level-Type A Village Village* Manawala Sudhar AverageValue AverageValue Choice of ServiceLevel No. % of House (Rs) No. % of House(Rs)

No privatefacility - - - 1 1 15,000 Manualhandpump only 26 24.5 67,423 16 16.7 53,437 Mowoizedpump only 2 1.9 60,000 6 6.3 133,333 Domesticconnection only i2 11.3 72,083 1 1 15,00J Manualhandpump and 44 41.5 142,380 36 37.5 75,735 domesic connection Motoized pumpand 22 20.8 193,181 36 37.5 115,833 domestc connection Total 106 100 124,692 96 100 89,680

* Tlheaverage value of a house is significantlydifferent in the two villages. Therefore,their statisdcsare presentedseparately.

Table D8 Leel OfSatisfacion wuh Piped Water Supply Sysem-Type A Vlage Al A2 Households Households Total Levelof Satisfaction No. % No. % No. % Satsfied 51 33.8 28 54.9 79 39.1 Dissatisfied 100 66.2 20 39.2 120 59.4 No response 0 0 3 5.9 3 1.5 Total 151 100 51 100 202 100

Program Repot Srie 115 Table D.9 MaJor Shert_omig of Piped Water Supply System Mentined by Houwebolds TyvpeA Villag % of HouseholdsIndicating Chamcteiistic as Shorcoming # 1 Shrtcoming # 2 AR Connected AD Conmected Charactersc Households Households Households Households Reliability* 70 69 32.5 33 Insufficientsupply** 15.8 17 21.7 22 Maintenance/design*** 10 10 15.8 15 Interactionwa". staff**** 4.2 4 20.8 21 Noresponse 0 0 9.2 9

* IncludesInsufficient pressure, frequent failures, and power breakdowns. * bIncludesInsufficient hours and low storagecapacity. "'" nlcludesSubstandard material and lack of cleanlness. -*** IncludesIrresponsible, uncooperaive, and comruptstaff.

116 uNDP-Workl&rak Waterad Saniion Progrn Table D.10 Extent of Informaton Rerding Parameters of Piped Water Systems Viliagelype

Manawala Sudha Parameter Al A2 Al A2 BI B2 Pricecharged (% respondingyes) 96 - NA* - - - Billing fiequ. ncy (no. of tmes per year) Mean 2.0 - NA - - - Standarddeviation 0.2 - NA - - - Householdsresponding (%) 96 - NA - - - Existingmonthly tariff (Rs) Mean 12.0 12.0 NA 11.0 15.6 16.9 Standarddeviation 23 1.0 NA 1.0 7.0 11.9 Householdsresponding (%) 100 68 NA 30 18 14 Connectionfee (Rs) Mean 180 257 85 86 79 144 Suandarddeviation 224 203 2.8 6.3 42 129 Householdsresponding (%) 87 57 100 78 4 9 Connectioncosts (Rs) Mean 721 947 443 445 453 503 Standarddeviadon 357 782 204 267 634 495 Householdsre ding () 100 82 100 87 42 17 Totalno. of hours of water supplyper day Mean 5.6 5.5 3.0 3.1 4.8 5.3 Standarddeviadon 1.0 1.1 0.8 1.6 2.0 1.9 Householdsresponding (%) 100 86 97 74 28 43 No. of hours of additionalsupply requiredto meet needs Mean 3.4 - 2.4 - - - Standarddeviaton 4.6 - 2.9 - - - Householdsrsonding (%) 99 99 - - - Frequencyof supplyfailure** Mean 0.5 - 1.8 - - - Standarddeviation 0.9 - 0.9 - - - Households esponding(%) - 100 - - -

* The supplyin Sudharwas kss than 6 monthsold and the fist billinghad no been made at the tme of the survey. Responses to some questions were not obtained because of the misinterpretationof a skip instrucdonin the questionnaire. ** For explanationof index see TableA-11.

Proam RepartSeies 117 Table D.i Cost of Piped Water Systems Tubeweli Canal Water Conts Based Based CapitalCost (RsA/apita) 300 S0 TotalCapit Cost* (000 Rs) 3,000 5,000 O&M Costs at 3% of CapitalCosts (Rs/Month) 7,500 12,500 Chargeshh/month**for full recovetyof O&M costs at a connectionfrequency of: (Rs) 100% 6.67 11.12 75% 8.90 14.83 50% 13.35 22.24 O&MCosts at 5% of CapitalCosts (Rs/Month) 12,500 20,833 Charges/hh/Anonthfor full recoveryof O&M costs at a connectionfrequency of: (Rs) 100% 11.12 18.53 75% 14.83 24.71 50% 22.24 37.07

*For averagevillage size of 10,000inhabitts. ** Averagehousehold size in the brackish-waterzone is 8.9. Table D.12 PerceptionsReprding Qualt of Water

Ty_pe of Houshold Perceptions Al A2 B1 B2 Piped water (% respondingyes) Satisfiedwith taste 98.7 - - - Satisfiedwith cleanliness 97.4 Satisfiedwith hygiene 93A - - - Alternativewater (% respoi yes) Satisfiedwith taste 19.2 41.2 23.5 72 Satisfiedwith cleanliness 92.7 96.1 89.5 94.6 Satisfiedwith hygiene 17.9 37.3 14 36.6 Householdsthat have visited village with opeational piped water system(%) - 52 63 Householdsthat feel piped water would be superiorto availablewater (%) - 90 92

118 UNDP-WmidBan Watera, I Saao Progaan Tble D.13 HouseholdP erene Regading ResponibUily for Operation and Maintenance of P WaterSystem Typeof Household Al A2 BI B2 Tbtal Mag.eOpt.on (%) (%)( (%) (%) PHED (govenment agency) 69.5 68.6 62.0 68.8 66.3 Lca councils(elected politicalbody) 9.3 9.8 14.5 7.5 11.1 Vilagemnuttee 9.9 9.8 11.0 6.5 9.7 Privateentreneur 2.0 0 2.0 3.2 2.0 No pefce 9.3 11.8 10.5 14.0 10.9

ProgramReport Serla 119 Appendix E

Table Ll Decsion to Install an Electric Motor: Results of a Lgit Model DependentVariable: Probabiity that a householdwould instai an electic motor Paramete Parameter Mean IndependentVaiables Estimate T-Ratio Estimate T-Ratio Values Intercept -1.54 -1.60 -1.72 -3.67** Householdsize 0.38E-1 1.21 0.19E-1 0.74 8.87 Waterconsumption 0.20E2 0.49 31.23 Animals -0.97E-1 -1.48 -0.11 -1.74* 0.88 Proportionof adult women -1.49 -1.32 -1.05 -1.19 0.28 Proportionof children -0.42 -0.58 0.37 Expenditureper capita 0.53E-3 0.22 216.81 Expenditureper capita squared 0.62E6 0.20 72,939 Constructionvaue of house 0.36e-5 2.46** 0.41e-S 2.85** 135,330 Ownenhip uf land or property(1 if yes) 0.79 3.30** 0.76 3.30** 0.53 Quaity of altenative water (1 if satisfied) -0.23 -0.86 -0.20 -0.76 0.23 Time 0.10E-2 1.50 0.83E-3 1.25 93.78 Age 0.37E-3 0.04 51.38 Education 0.13 4.26** 0.14 4.63** 9.12 Occupation(1 if farming) -0.15 -0.48 0.18 Free supply(I if yes) -0.12 -0.57 0.52 Extemalexpose (1 if yes) 0.49E-1 -0.21 0.38 Householddummy Al -0.60 -1.88* 0.47 1.88* 0.30 A2 -1.36 -2.81** -1.29 -2.93** 0.11 B1 -0.19 -0.62 0.40 Numberof observaions 481 481 Log-lielihood -264.51 -266.02 Restrictedlog-likelihood -333.38 -333.38 Chi-square (19) 137.73 (10) 134.72 Significancelevel 0.32E-13 0.32E-13 Proportionof correctpredictions 0.73 0.74

120 UNDP-World ank Water and Santadon Program Table E.2 Decsion to Connect to a Piped Water Supply System: Results of a Logit Model DependentVariable: Probability that a householdwould connectto a piped watersupply system Parameter Parameter Mean IndependentVariables Estimte T-Ratio Estimate T-Ratio Values Intercept 3.29 1.38 0.96 0.75 Householdsize 0.87E-1 1.36 0.66E-1 1.09 8.82 Waterconsumption -0.69E-2 -0.86 31.08 Animals 0.11 0.61 0.68 Proportionof adult women -2.57 -0.88 0.28 Proportionof children -4.03 -2.08** -2.29 -2.34** 0.40 Expenditureper capita -0.12E-1 -1.45 -0.13E-1 -1.60 203.67 Expenditureper capita squared 0.23E4 -1.56 0.25E-4 1.59 57,177 Constructionvalue of house 0.76E-5 1.64 0.82E-5 1.87* 108,640 Ownershipof land or property(I if yes) -0.13 -0.22 0.37 Qualityof altrative water (1 if satisfied) -0.78 -1.75* -0.68 -1.54 0.26 P.rivatewater source (1 if motor) 0.77 1.37 0.67 1.24 0.34 Age -0.17E-1 -1.09 51.12 Education 0.12 2.14** 0.12 2.32** 7.83 Occupation(1 if farming) -1.95 -2.66** -1.74 -2.81** 0.13 Meter (1 if yes) 1.02 2.19** 0.95 2.12** 0.77 Free supply(I if yes) -0.28 -0.68 0.50 Extenal exposure(1 if yes) -0.44E&1 -0.09 0.28 Distancefrom distribution line 0.66E-1 -2.18** -0.64E-1 -2.23** 5.86 Villagedummy Sudhar 0.49 1.13 0.43 1.05 0.48 Numberof obs.rvations 196 196 Log-likelihood -82.38 -84.04 Restrictedlog-likelihood -112.36 -112.36 Chi-square (19) 59.96 (12) 56.64 Significancelevel 0.39E-5 0.25E-8 Proportionof correctpredictions 0.82 0.81

Program Repo"t Sei 121 Table E3 Satistcal Analysisof Willi to Pay for a Standard Piped Water System: Results of Ordinary Least Squares Reression Model DependentVariable: Midpoint of intervalin biddinggame within whichrespondent's WTP bid falls Parameter Paraneter Mean IndependentVariables Estimate T-Ratio Estimate T-Ratio Values Intrcept 34.99 4.00** 32.51 5.50* Householdsize 0.48 1.63 0.57 2.16** 8.87 Waterconsmption 0.08 1.96* 0.08 2.10** 31.23 Animals 0.31 0.50 0.88 Proportionof adult women -6.48 -0.60 0.28 Proportionof children -2.18 -0.31 0.37 Expendture per capita 0.02 1.37 0.02 1.73* 216.81 Expentue per capita squared -1.20 -1.25 -1.31E-5 -1.41 72,939 Constructionvalue of house 8.00E-6 0.84 135,326 Owneship of landor property( if yes) -2.04 -0.84 0.53 Qualityof alteatve water (1 if satisfied) -3.27 -1.30 -3.47 -1.41 0.23 Privatewater source (1 if motor) 0.95 0.41 0.51 Tie 0.01 1.99** 0.01 2.10** 93.78 Age -0.21 -2.71** -0.20 -2.88** 51.38 Education 0.12 0.43 9.12 Occupation(1 if faming) 0.86 0.30 0.18 Meter (1 if yes) 5.58 2.11** 5.36 2.06** 0.82 Free supply(1 if yes) -0.42 -0.21 0.52 Extrnal exposue (1 if yes) 4.04 1.79* 4.09 1.88* 0.38 Awarenessof water systems(1 if yes) 0.01 0.01 0.74 Householddununy Al -17.18 -5.62** -17.52 -6.64** 0.30 A2 -31.57 -7.58** -32.60 -9.00** 0.11 B2 -3.03 -1.03 -2.97 -1.04 0.19 Staring point dummy (1 if high) 0.10 0.05 0.50 Numberof observations 481 181 Mean of dependentvariable 30.47 30.47 Stan. dev.of dependen variable 25.63 25.63 F-value (23,457) 8.62 (12,468) 16.61 Significanceof F-test 0.00 0.00 R-squared 0.30 0.30 AdjustedR-squared 0.27 0.28

122 UNDP-WorldBRan Water and Sntatimn Progrom Table L4 Statistical Analysis of WliUngpessto Pay for an Improved Piped Water System: Results of Ordinr Least Squares Regrsson Model DependentVariable: Midpoint of intervalin biddinggame withinwhich respondent's WTP bid fals Paameter Parameter Mean IndependentVariables Estdmate T-Ratio Estimate T-Ratio Values Intercept 37.46 3.39** 36.32 4.71** Householdsize 0.89 2.36** 0.92 2.64** 8.87 Waterconsumption 0.11 2.31** 0.11 2.29** 31.23 Animas 0.45 0.57 0.88 Proportionof adult women -1.13 -0.08 0.28 Proportionof children -2.24 -0.26 0.37 Expenditureper capita 0.03 1.97** 004 2.12** 216.81 Expenditureper capitasquared -2.08E-5 -1.72* -2.19B-5 -1.83* 72,939 Constructionvalue of house 1.77E-5 1.47 1.97E-5 1.72* 135,326 Ownershipof land or property (1 if yes) 1.94 0.63 0.53 Quality of altemativewater (1 if satisfied) -8.55 -2.69** -8.52 -2.73** 0.23 Private water source (1 if motor) 1.75 0.60 0.51 Tmie 0.02 1.96* 0.01 1.99** 93.78 Age -0.31 -3.09** -0.31 -3.44** 51.38 Education 0.69 1.98** 0.77 2.31** 9.12 Occu- -ion (I if farming) -2.44 -0.67 0.18 Metei kt if yes) 7.14 2.14** 7.03 2.14** 0.82 Free supply(1 if yes) -0.41 -0.16 0.52 Externalexposure (1 if yes) 5.28 1.86* 5.51 1.97** 0.38 Awarenessof water systems (1 if yes) -2.70 -0.78 0.74 Householddummy Al -14.30 -3.70** -16.25 -4.84** 0.30 A2 -20.82 -3.96** -23.48 4.95** 0.11 B2 -4.52 -1.22 -5.19 -1.43 0.19 Staruringpoint dumy (1 if high) 1.29 0.50 0.50 Numberof observations 481 481 Mean of dependentvariable 46.77 46.77 Stan. dev. of dependentvariable 31.72 31.72 F-Value (23,457) 7.51 (14,466) 12.35 Significanceof F-test 0.00 0.00 R-squared 0.27 0.27 Adjused R-squared 0.24 0.25

Prgram Report er:es 123 Tlble E.5 Stadstial Analysisof Willnness to Pay for a Standard Piped Water System-Type A Vilage: Resultsof Ordinary LeastSquares n Modd DependentVariable: Midpoint of intervalin biddinggame within which respondent's WTP bid falls . Parameter Paameter Mean Indpendet Variables Estimate T-Ratlio Estimate T-Ratio Values Intercept -6.79 -0.79 -3.18 -0.77 Householdsize 0.11 0.48 . 8.82 Waterconsumption -0.05 -1.53 -0.04 -1.43 31.08 Animals 1.93 3.50** 1.96 3*73** 0.68 Proporion of adult women 0.39 0.04 0.29 Propordonof children 3.19 0.47 0.40 Expenditureper capita 0.07 3.21** 0.06 3.39** 203.67 Expenditureper cata squared -6.55E,5 -2.42** -5.90E-5 -2.47** 57,177 Consuvion value of house -5.811E-6 -0.87 108,638 Ownershipof land or property (I if yes) -0.41 -0.21 0.37 Quality of alernadve water (I if satisfied) -0.60 -0.32 0.26 Private water source (1 if iflot) 0.03 0.02 0.34 Time 2.13E-3 0.09 12.16 Age -0.06 -1.07 -0.07 -1.39 51.12 Education -0.05 -0.26 7.83 Occupation(1 if fanring) -2.52 -0.91 -3.67 -1.59 0.13 Meter(1 f yes) 2.70 1.45 2.57 1.46 0.77 Freesupply (I if yes) 0.72 0.45 0.50 Fxtenld exposwue(1 if yes) 1.66 0.92 0.28 Sabtsfaction(1 if yes) -1.05 -0.64 0.39 Villagedummy Sudhar 2.85 1.64 3.03 2.02** 0.48 Householddummy Al 16.19 7.13** 16.00 9.38** 0.74 Stating poi dummy (1 i high) 0.48 0.30 0.50 Numberof observations 196 196 Meanof dependentvarilde 16.49 16.49 Stan. dev. of dependentvariable 13.18 13.18 F-vaIue (22,173) 6.32 (9,186) 15.96 Significanceof F-test 0.00 0.00 R-squared 0.45 0.44 AdjusedR-squared 0.38 0.41

124 UNDP-Wrd Bank Water and Sanua Program lTbl EA6 Stastical Analysisof WilHUe to Pay for an Improved Piped Water Systen-Tlype A Viae: Reslts of OrdhinaryLeast Squame Regresio Modd DependentVariable: Midpoint of intervalin biddinggame withinwhich respondent's WTP bid fals Parameter Parameter Mean IndependentVaiables Estimate T-Ratio Estimate T-Ratio Values ntrcept 9.53 0.59 3.22 0.35 Householdsize 0.89 2.08** 0.99 2.44** 8.82 Waterconsumption -0.06 -0.92 31.08 Aninals 0.31 0.30 0.68 Proporion of adult women -21.94 -1.12 -22.03 -1.86* 0.29 Proportionof children -2.69 -0.21 0.40 Expenditureper capita 0.14 3.47** 0.14 3.91** 203.67 Expenditureper capita squared -1.28E-4 -2.51** -1.42E-4 -2.97** 57,177 Constructionvalue of house -1.44E-S -1.14 -1.22E-5 -1.02 108,638 Ownershipof land or property (I if yes) 7.05 1.91* 6.40 2.04** 0.37 Qt-.ty of alteiative water (1 if satisfied) -9.30 -2.63** -9.56 -2.80** 0.26 Private water source (1 if motor) 1.50 0.42 0.34 Thue 0.05 1.06 0.05 1.24 12.16 Age -0.19 -1.75* -0.19 -1.90* 51.12 Educaion 0.11 0.30 7.83 Occupation(1 if farnning) -4.95 -0.95 0.13 Meter (1 if yes) 5.44 1.55 5.65 1.69* 0.77 Free supply(1 if yes) -1.03 -0.34 0.50 Extmemalexposure (1 if yes) 4.03 1.19 3.88 1.21 0.28 Satisfaction(1 if yes) -1.43 -0.47 0.39 VRllagedummy Sudhar -0.41 -0.13 0.48 Householddummy Al 10.63 2.49** 12.30 3.1S** 0.74 Startingpoint dummy (1 if high) 4.01 1.36 4.02 1.42 0.50. Numberof observations 196 196 Meanof dependentvariable 33.06 33.06 Stan. dev. of dpedent variable 22.02 22.02 F-value (22,173) 3.35 (13,182) 5.66 Significanceof F-test 0.00 0.00 R-squared 0.30 0.29 Adjustd R-squared 0.21 0.24

Progn RepontSries 125 Table E7 Statistical Analysp of Wilinpess to Pay for a Standard Piped Water System-Type B Village: Resilts of Ordary Lest Squares R Model DependentVariable: Midpoint of intrval in biddinggame withinwhich respondent's WTP bid falls Parameter Parameter Mean IndependentVaiables Esdtmaf T-Ratio Estinate T-Ratio Values Intercept 33.36 2.42** 33.06 4.00** Householdsize 0.90 1.86* 0.67 1.76* 8.90 Waterconsumption 0.1S 2.46** 0.15 2.65** 31.34 Animals -0.76 -0.80 1.02 Proportionof adult women -11.38 -0.73 0.28 Proportionof children -6.59 -0.64 0.36 Expenditureper capita 0.01 0.67 225.86 Expenditureper capitasquared -9.61E-6 -0.74 83,779 Constructionvalue of house 2.10E-5 1.26 2.07E-5 1.41 153,681 Ownershipof land or property(1 if yes) -2.04 -0.54 0.64 Qualityof alternativewater (1 if satisfied) -5.59 -1.32 -6.42 -1.63 0.21 Privatewater source (1 if motor) -0.80 -0.22 0.62 Tume 0.01 1.58 0.01 1.64 149.92 Age -0.35 -2.73** -0.31 -2.69** 51.56 Education 0.41 0.86 10.01 Occupation(1 if farming) 2.43 0.57 0.22 meter(1 if yes) 7.99 1.75* 7.40 1.67* 0.85 Free supply(1 if yes) -0.39 -0.12 0.54 Extenal exposue (1 if yes) 4.93 1.40 5.11 1.55 0.45 Awarenessof water systems (I ff yes) -1.40 -0.41 0.57 Householddummy BI 3.00 0.82 0.68 Startingpoint dummy (1 if high) -1.00 -0.31 0.50 Numberof observations 285 285 Mean of dependentvariable 40.09 40.09 Stan. dev.of dependentvariable 27.62 27.62 F-value (21,263) 1.75 (8,276) 4.20 Significanceof F-test 0.02 0.00 R-squared 0.12 0.11 AdjustedR-squared 0.05 0.08

126 UNDP-Wordank Water and Santaton Program Appendix F

Ibble F.1 Distibution of WTP Bids for a Standard Piped Water System

e Village ype Meanbid* A BI B2 (Rs) No. No. % No. % 0 16 7.9 3 1.5 3 3.2 5.0 13 .6.4 0 0 1 1.1 7.5 22 10.9 4 2 5 5.4 12.5 50 24.8 25 12.5 11 11.8 17.5 53 26.2 9 4.5 3 3.2 25.0 29 14.4 47 23.5 22 23.7 35.0 14 6.9 20 10 11 11.8 45.0 2 1 39 19.5 18 19.4 62.5 1 0.5 38 19 13 14 87.5 1 0.5 2 1 2 2.2 100+ 1 0.5 13 6.5 4 4.3

Total 202 100 200 100 93 100

No. of bids 151 74.8 193 96.5 84 90.3 > Rs 12

Mean bid 16.42 41.19 36.92 (R) Mean of bids 20.65 42.52 40.40 > Rs 12 (Rs)

Meanof bids 3.91 4.29 4.50 < Rs 12 (Rs) Note: The followingapplies to TablesF-I to F-4. * Mean bids are the mid-pointsof the intervalsin which the respondents bids fell (except0, 5, and 100).All genuinezero bids are includedin 0; all bids greaterthan zero and less than or equal to Rs S are includedin S; andall bids equal to or greaterthan Rs 100are included in 100+.

ProgramReport Scries 127 Tible F2 Distibution of WTPBids for an Inproved PipedWater System

Meanbid A Y Bl B2 (Rs) No. % No. % No. 9 0 4 2 0 0 1 1.1 5.0 0 0 0 0 0 ) 7.5 6 3 2 1 2 2.2 12.5 26 12.9 6 3 3 3.2 17.5 29 14.4 11 5.5 6 6.5 25.0 47 23.3 22 11 8 8.6 35.0 32 15.8 29 14.5 15 16.1 45.0 32 15.8 37 18.5 23 24.7 62.5 19 9.4 43 21.5 22 23.7 87.5 2 1 24 12 5 5.4 100+ 5 2.5 26 13 8 8.6

Total 202 100 200 100 93 100

No. of bids 192 95 198 99 90 96.8 >Rs 12

Mean bid 32.60 58.04 50.81 (Rs)

Meanof bids 34.06 58.55 52.33 > Rs 12 (Rs)

Mean of bids 4.50 7.50 5.00 < Rs 12 (Rs)

128 LWDAWarld Bank Water and Sant Program TableF3 Dishibution of WTPBids for a S ardPiped Water System with AlternativeFnaning Arrang nts: A2 Households Financingarangement Mean bid ExlrraSnem a gement Nn= eingeent2** (Rs) No. % No. % No. % 0 16 31A 7 13.7 4 7.8 5.0 13 25.5 8 15.7 2 3.9 7.5 22 43.2 12 23.6 13 25A 12.5 0 0 15 29.4 9 17.6 17.5 0 0 5 9.8 12 23.5 25.0 0 0 3 5.9 7 13.7 35.0 0 0 1 2 3 5.9 45.0 0 0 0 0 0 0 62.5 0 0 0 0 1 2 87.5 0 0 0 0 0 0 100+ 0 0 0 0 0 0

Total 51 100 51 100 51 100

No. of bids 0 0 24 47.06 32 62.75 > Rs 12

Meanbid 3.91 9.76 15.24 (Ps)

Meanof bids - 16.04 20.78 > Rs 12 (Rs)

Meanof bids 3.91 4.19 5.89 < Rs 12 (Rs)

* Under Amgement 1 thewater authoritiesbear 50% of the connectioncost. F* UnderAnangement 2 the wat authoritiesbear100% of theconnweion cost

programReport Sries 129 Table F4 DIstrIbutlonof WTP Bids for Standard and Improved Piped Water Systems Vig with Operatng Piped Water Systems Standardsystem Improvedsystem lTpe A1 Tpe A2 lype Al lype A2 Meanbid Households Households Households Households (RS) No. % No. 9 -wNo. % No. % 0 0 0 16 31.4 0 0 4 7.8 5.0 0 0 13 255 0 0 0 0 7.5 0 0 22 43.2 1 0.7 5 9.8 12.5 50 33.1 0 0 16 10.6 10 19.6 17.5 53 35.1 0 0 21 13.9 8 15.7 25.0 29 19.2 0 0 36 23.8 11 21.6 35.0 14 9.3 0 0 28 18.5 4 7.8 45.0 2 1.3 0 0 25 16.6 7 13.7 62.5 1 0.7 0 0 17 11.3 2 3.9 87.5 1 0.7 0 0 2 1.3 0 0 100+ 1 0.7 0 0 5 3.3 0 0

Total 151 100 51 100 151 100 51 100

No. of bids 151 100 0 0 150 99.3 42 82.4 >Rs 12

Meanbid 20.65 3.91 35.94 22.70 R8i)

Meanof bids 20.65 - 36.13 26.67 > Rs 12 (Ri)

Mean of bids - 3.91 7.50 4.17 < Rs 12 (Rs)

130 UNDP-WordBan Watr and Sad. NW= Tbble FS Connectio Frequenies and fEtimated Revenues: Proviion of a Standard Piped Water System

YillameMyE A Bi b2 Monthly Households Est.Rev. Households Est. Rev. Households Est. Rev. tariff Corn. (Rs/lo Conn. (Rs/l1O CoM. (RsIlOO (Rs) (%) households) (%) households) (%) households) 0 100 0 100 0 100 0 5.0* 92.1 461 98.5 493. 96.8 484 7.5 85.7 643 98.5 739 95.7 718 12.5 74.8 935 96.5 1,206 90.3 1,129 17.5 50.0 875 84 1,470 78.5 1,374 25.0 23.8 595 79.5 1,988 75.3 1,883 35.0 9.4 329 56 1,960 51.6 1,806 45.0 2.5 113 46 2,070 39.8 1,791 62.5 2 125 26.5 1,656 20.4 1,275 87.5 1.5 131 7.5 656 6.4 560 100** 05 50 6.5 650 4.2 420 Note: The followingawly to TablesF-S to F-7. * All bids betweenRs 0 and Rs5 areonsolidated in e Rs5 category. ** All bids greaterthan Rs 100 are conSlidated in the Rs 100 category.

be F6 Connecto Frquendes and Estmated Revenues: provio of an Improved Pipe Wae system Yiae type A B1 B2 Monthly Households Est Rev. Houshols ESt.Rev. Households Es. Rev. tariff CoMn. (Rs/100 Conn. (Rs/l00 Con (RsI00 (RS) (%) households) (%) households) (%) households)

0 100 0 100 0 100 0 5.0 98 490 100 500 98.9 495 7.5 98 735 100 750 98.9 742 12.5 95 1,188 99 1,238 96.7 1,209 17.5 82.1 1,437 96 1,680 93.5 1,636 25.0 67.7 1,693 90.5 2,263 87 2,175 35.0 44.4 1,554 79.5 2,783 78.4 2,744 45.0 28.6 1,287 65 2,925 62.3 2,804 62.5 12.8 800 46.5 2,906 37.6 2,350 87.5 3.4 298 25 2,188 13.9 1,216 lOD" 2.4 240 13 1,300 8.5 850

Proram ReportSeries 131 TableF.7 Connecton Frequenies and Em d Revenues: Provson of Optons to A2 Housebolds

Standardssn Std sstem Sndardsystem bIprovedsystem (^_Amet1) (ArrAnMen2)4- Monthly HouseholdsESLRov. Households EsLtRev. Houselds BEstlRev.Househlds EtRev. tarff Coo (Rs/100 Cam (R/OO Coim. OR100 COMn (Rs/100 (Rs) (%) househds) (%) housods) (%) household) (%) households) 0 100 0 100 0 100 0 100 0 5.0 68.6 343 86.3 432 92.2 461 92.2 461 7.5 43.1 323 70.6 530 88.3 662 922 692 12.5 0 0 47 588 62.9 786 82.4 1,030 17.5 0 0 17.6 308 45.3 793 62.8 1,099 25.0 0 0 7.8 195 21.8 545 47.1 1,178 35.0 0 0 1.9 67 8.1 284 25.5 893 4S.3 0 0 0 0 2.2 99 17.7 797 62.5 0 0 0 0 2.2 138 4.0 250 .87.5 0 0 0 0 0 0 0 0 100 0 0 0 0 0 0 0 0 + UnderAngeinent I the waterautdoies bear 50% of theconnection cost ++ Underan=nWeent 2 the waterasuthoities bear 100%of the connectioncosL

132 UNDP-WmdBa* Waterand S_wio P nram Appendx G

TableGA1 lageProles Gorakh- Dlud- Jawa Banda Dhalla Papin Payal pur dian Moha Bodhial Area*(acres) 1263 3664 2801 3298 508 2228 277 629 1441 Population* 835 2129 1129 941 854 1563 610 1222 480 Housdeoldsize 5.8 6.0 5.2 7.1 6.0 6.2 5.5 4.4 6.3 (fromsample, 1989) (7.9) (7.3) (7.2) (6.5) (8.0) (7.6) (7.1) (6.8) (6.5) Litacy mtio 30.9 36.1 25.8 15.6 33.4 28.2 59.6 36.9 38.0 Males- schoolage and above(%) 5.5 11.3 7.2 1.0 7.9 6.0 16.8 9.0 10.4 Females- schoolage and above(%) 0.9 2.0 1.1 0.2 1.6 0.6 8.3 4.4 0 DisUac from district (kcm) 24 45 24 56 35 14 37 48 25 Roadlocaior. off Rawat- main off off main Rawat- main off Rawat- Chak Adiala Rawat- Rawat- Adiala Chak Rawat- Adiala Banda Beli Road Chak Chak Road Beli Chak Road ULnk Link Beli Bell Link Beli Road Road Road Road Road Road Presenceof facilites Electricity Yes Yes Yes Yes Yes Yes Yes Yes Yes CJinic No No No No No No Yes No No Educaion(boys) Yes Yes Yes Yes Yes Yes No Yes No Education(gids) Yes Yes Yes Yes Yes Yes Yes Yes Yes

* Figuresfrom 1981 populaton census. Al othersfrom field visit

ProgrwnRepot eries 133 Thble G.2 Village Profiles*

Gorakh- Dhud- Jaw Banda Dhaia Papin Paya pur dim Mohra Bodhia

Household size 8 7 7 6 8 8 7 7 7 Age of head of househOId(years) 53 50 49 56 55 51 54 52 49 Adldt womn in household.(%) 30 30 33 37 28 28 32 30 28 Chldken in household(%) 39 37 35 32 43 42 32 36 38 Years of education of moSt educated: Memiberof houseold 8 8 9 8 8 8 10 8 7 Womnanin household 2 3 3 2 3 4 4 2 2 Essential water consumion llitescapits/dar) Sunmer 20 23 25 21 23 20 23 29 23 Winter 13 16 18 14 17 14 15 21 17 Householdsinvolved in famiung(% 30 15 23 52 40 38 10 44 56 Householdsowning aninas (%) 77 63 73 86 86 58 83 86 83 Wate cosumption of animals (literslanimal/day) Summer 42 40 26 44 47 20 34 70 41 winter 29 24 20 26 31 IS 23 51 27 Household.with land orotherpmperty(%) 73 52 56 83 88 90 60 93 83 Consotnctionvalue Of house(OOORs) 91 75 85 69 115 116 102 95 77 Per capita monthly expenditre(Rs) 238 232 201 188 204 182 268 212 238 Househoklswith extenal expcswe (%) 27 23 25 36 31 16 43 26 25 Householdsthat believe wat should be suppliedfiee (%) 93 85 83 100 90 80 100 95 90 Householdsthat believe watercan be suppliedfie(%) 66 58 54 71 79 56 77 77 83 Houhobds tha believe water supply systems dhouldbe managedby PHED(*) 91 100 90 55 67 86 60 77 71 Househols that favor numetingofwaer(%) 68 44 46 57 67 68 70 67 63

* Al msm are derved from the sample-observaions TableG3 Profie by Villge lType Villagetype A BI B2 Houseold size 8 7 7 Ageofheadofausehofd 51 54 51

Adultwomen in household(%) 31 31 30 Cldldn in household(%) 37 39 36

Yearsof educationof mosteducated Memberof household 8 8 8 Womanin household 3 3 3

Essential water consumption (liters/apita/day) Summer 23 21 25 WVter 16 15 18

Householdsinvolved i fing (%) 22 43 41 Householdsowninganimals(%) 71 76 84

Waterconumptin of snmals (literslanimal/day) Summer 36 37 50 Witer 24 24 35

Householdswith land or other property(%) 60 87 81 . Contuctionvalueofhouse(OOORs) 84 102 90 Per capitamonthly expendie (Rs) 224 191 236 Householdswith ext aexposure (%) 25 27 30

Householdsdia believewae shuld be suppliedfree (%) 87 89 94

Householdsthat belive waer can be applied free (%) 59 69 79

Househods tat beleve watersupply syste shod be mangd by PHED)(V) 94 70 70

Householdstdatfavormeteringofwater(%) 52 64 66

* MAstatstcs are derivedfrom the sanmleobservations.

PwgramReport Series 135 Table GA Level of Satisfaction with Piped Water Supply System -Tlpe A Village Villages Jawa Banda Dhalla Total Level of satisfaction No. % No. % No. % No. % Satisfied 10 24.4 33 71.7 47 100 90 67.2 Not satisfied 31 75.6 13 28.3 0 0 44 32.8 Total 41 100 44 100 47 100 134 100

* Water supplies were installed in Jawa, Banda, and Dhalla in 1989, 1987, and 1988, respectively.

TIble G.5 Major Shortcomis of Piped Supply System Mentioned by ConnectedHouseholds-Type A VMflages % of HouseholdsIndicating Chaateristic as Shortcoming# 1 Shortcoming# 2 Characteristic Jawa Banda Jawa Banda Rcliability* 61.3 38.5 29 0 insufficientsupply** 16.1 7.7 29 46.2 M4aintenance/design*** 3.2 0 0 0 Interactionwith staff**** 19.4 53.8 35.5 15.4 No response 6.5 38.5

* Includesinsufficient pressure,frequent failures, and powet'breacdowns. ** Includesinsufficient hours and low storagecapacity. ** Includessubstandard material and lack of cleanliness. ***' Includeshiesponsible, uncooperadve, and coffuptstaff.

136 UNDP-WorldBank Watr and Sanaton Pogmw TableG.6 Cost of Piped Water SystemsIn the Arid Zone Villagetype A Villagetype BI Averages rype A Gorakh- TypeA &BI Jawa Banda Dhalla Papin Payal pur villages villages Population(1981) 835 1,974 1,129 941 935 1,563 1,312 1,238

Population(1989) 1,667 1,562

Capitalcost* (000 Rs) 1,704.7 1,152.8 1,461.8 783.3 1,713.3 1,035.8 1,439.8 1,308.6 O&Mcost* per year (000 Rs) 72.51 56.2 78.1 - - - 68.94 -

Yearcompleted* 1989 1988 1989 1989-90 1989-90 1989-90 - -

Householdsize (1981) 7.1 6.0 5.2 7.1 6.0 6.2 6.1 6.3

(ftom sanple, 1989) (7.8) (7.5) (7.3) (6.5) (8.0) (7.6) (7.5) (7.5)

No. of houses (1989) 222 208

No. of householdsconnected 212 199 [Using0.957 connection raio of A villagesl

O&Mcost per householdper month at 100%connection rate (Rs) 25.9 27.6 at 95.7%connection rate (Rs) 27.1 28.9

* FrmPHED projectdocuments.

TIble G.7 Reasons Mentioned by Households for Obtahninga Domestic Conunecton-Type A Vilae % of householdsindicating reason as Reasons ReasonNo.1 ReasonNo.2 Need 82.1 12.1 Convenience 15.7 56.4 Cleanliness 0.7 14.5 Other 1.5 17.0

ProgramReport Series 137 Table G.8 Extent of Information Regarding Parameters of Piped Water Systems Villagetype A Bi B2 Jawa Banda Dhalla Price chared (% respondingyes) 51 100 87 - - Billingfequency (no. of times per yar) Mean 2 2 2 - - Standarddeviation 0 0 0 - _ Householdsresponding(%) 51 100 85 - - Pxisfingmonthy tatiff (R.s) > Mean 20 20 20 19.30 22 Standarddeviation 0 0 1.07 8.08 9 Householdsrespzding (%) 71 100 89 35 33 Connectionfee (Rs) Mean 238 229 204 150 325 Standarddeviation 91 82 46 107 340 Householdsrsoding(%) 100 100 100 18 6 Connectioncosts Ots) Mean 725 835 715 431 1396 Standarddeviation 462 438 581 316 1652 Househols sponding(%) 100 100 100 35 16 TotalNo. of hous of water supply perday Mean 0.84 2.37 2.10 4 4 Standarddeviation 0.68 1.16 0.73 3 4 Householdsresponding (%) 100 100 100 57 49 No. of hours of addition supply requiredto meet needs Mean 2 2.6 2 - - Standarddeviaion 1.2 1.5 0 - - Househols responding(%) 36 10 1 - - Frequencyof supplyfailwue* M1ean 2 0.35 0.17 - - Standarddeviaton 1.2 0.76 0.60 - - Householdsespmpding(%) 100 100 100 - -

* For explanain of index see TableA-Il.

138 UNDP-WorldBa* Water and Santatn Pmgr.n T3ableG. Household Preference Rerdn Responsbilityfor Operation and Maintenance of Piped Water Systms Villagetype Management option Al BI B2 Total (%) (%) (%) (%) PHED (govenment agency) 93.4 70.0 70.0 78.3

Local councils (Electedpolitical body) 2.2 8.6 10.7 7.0

Villagecommittee 3.7 11.4 7.3 7.5

Privateentrepreneur 0.7 0.7 12.0 0.2

Indifferent 0.0 9.3 0.0 7.0

Proram ReportSeres 139 AppendixH

Tabe H.1 Stadtstal Analyis of W hness to Pay for a Piped Water System Based on Publi Taps-Type B2 Vilges: Resultsof OrdinaryLeas Squares RegresdonModel DendeetVeda: Mipont of inval in biddinggame withinwhich respondent's W1P bid ls. Parameter Parameter Mean Indepdent Varibles Estimate T-Ratio Estimate T-Ratio Values Intercept 126.78 3.00** 116.54 3.38** Householdsize 3.27 3.12** 2.82 3.57** 6.77 Waterconsumption 0.07 0.27 25.25 Animals 1.67 1.36 1.45 1.40 2.33 Proportionof adultwomen -8.31 -0.40 0.30 Prpoztlonof children -12.09 -0.79 0.36 Expenditreper capita 0.05 2.07** 0.05 2.52** 236.05 Constuctionvalue of house -2.18E-5 -0.47 89,661.16 Ownershipof landor propety (I f yes) 0.91 0.1? 0.81 Qualityof alterativewater (1 if satisfied) -28.77 -1.94 -29.61 -2.17** 0.97 Privatewater source (I if yes) 0.57 0.09 0.37 Vendorusage (I if yes) 2.87 0.27 0.07 Age -0.20 -1.00 51.39 Educaion -0.12 -0.14 8.01 Occupation(I if farming) - 1.68 -0.26 0.41 Meter(1 if yes) 13.12 2.23** 13.48 2.58** 0.66 Freesupply (1 if yes) -10.32 -0.91 -12.33 -1.19 0.94 Eternal exposue(I if yes) -5.50 -1.03 -6.09 -1.26 0.53 Awarenessof waersystems(1 if yes) 3.07 0.56 0.65 Sex(I if male) 0.81 0.10 0.86 Ditanceof vilage fromdisL hedquats - 1.79 -2.39** -1.76 -2.91** 36.15 Vlllagedummy Bodhial 45.53 -2.99** -44.95 -3.66** 0.40

Numberof observations 121 121 Meanof dependentvariable 34.92 34.92 Standarddeviation of depedent vaiable 30.19 30.19 P-Value (21,99) 2.43 (9,111) 5.94 Significanceof F-test 0.00 0.00 R-squared 0.34 0.32 AdjusedR-squared 0.20 0.27

140 UNDP-WorldBank Water and SWatado PmV= TableH.2 SbtastIcal alyds of Wil_ngs to Pay for a Standar Piped Water System- Type B2 Villages: Resultsof Ordinry Least Squas Rression Model Depeewn Vaabl: Midpointof intervalin biding game withinwhich respondent's W1P bid falls. Parameter Prameter Mean Independentvariables esdmate T-Raio estmate T-Ratio values Intercept 124.53 2.37** 103.12 2.40** Householdsize 4.00 3.07** 4.37 4.12** 6.77 Waterconsumption 0.37 1.11 0.43 141 !3.25 Animals 1.75 1.15 2.33 Proportionof adultwomen -44.64 -1.71* 43.04 -1.77* 0.30 Propotionof children -33.60 -1.78* -34.05 -2.11*J 0.36 Expenditureper capita 0.08 2.83** 0.07 2.85** 236.05 Constructionvalue of house - 7.57E-6 -0.13 89,661.16 Ownershipoflandorproperty(1if yes) -3.15 433 0.81 Qualityof altemaivewater(1 if satisaied) -20.03 -1.09 -18.75 -1.10 0.97 Privatewater source (I if yes) 3.96 0.53 0.37 Vendorusage (1 if yes) 11.93 0.84 12.71 1.06 0.07 Age -0.16 -0.66 51.39 Education 0.57 0.54 8.01 Occupation(I if farming) - 3.21 -0.41 0.41 Met(erI(if yes) 9.01 1.23 7.89 1.19 0.66 Freesupply (1 if yes) -6.19 -0.44 0.94 Extenalexposure (I if yes) - 3.46 -0.52 0.53 Awatenessof waterysems (1 if yes) -4.02 -0.59 0.65 Sex (I if nmle) 3.18 0.32 0.86 DistanceofvUlagefrom dist headquate - 1.65 -1.77 -1.45 -1.88* 36.15 Villagedumny Bodhial -40.67 -2.15** -38.54 .2.47** 0.40

Numberof observations 121 121 Meanof dependetvariable 56.45 56.45 Standarddeviation of dependentvariable 37.21 37.21 P-value (21,99) 2.31 (10,110) 4.88 Significanceof F-test 0.00 0.00 R-quared 0.33 0.31 AdjustedR-squared 0.19 0.24

Program Report Series 141 Tables H3 Statistc Analyis of WIllingnessto Pay for a Standard Piped Water Sy8tem- Type Di Vlages- Resultsof Ordinr Least Squrs Repsson Model Depeadut Va*rk: Midpointof intervalin biddinggame withinwhich respondet's WTPbid falls.

Paramew Paraneter Mean Independentvariables estimate T-ratio estimate T-ratio values Intercept -23.43 -0.67 12.09 0.77 Householdsize 1.76 1.71* 1.53 1.97* 7.42 Waterconsumption 0.04 0.15 21.23 Aniials -1.74 -1.30 -2.13 1.80* 2.02 Proportionof adult women 23.89 0.78 0.31 Proportionof childrm 5.39 0.27 0.39 penditure per capita 8.94E-3 0.30 192.21 Constuction value of house 7.03B,5 1.88* 7.60E-5 2.42**101,742.65 Ownershipof land or prperty (1 if yes) 9.87 1.14 10.30 1.36 tJ.88 Quality of altemativewater (1 if sased) 18.33 1.57 16.76 1.62 0.94 Privatewater source (I if yes) 3.04 0.45 0.29 Vendorusage (I if yes) 13.39 1.07 14.17 1.27 0.05 Age -9.52E4 -0.01 54.06 Education 0.01 0.02 7.79 Occupation(1 if farming) -7.56 -1.25 -8.47 -1.60 0.43 Meter (1 if yes) 0.93 0.17 0.65 Free supply(1 if yes) -7.49 -0.82 -8.37 -1.05 0.89 Extenal exposure(I if yes) 1.41 0.24 0.46 Awarenessof water systems(I if yes) 2.92 0.46 0.74 Sex (1 if male) 5.77 0.60 0.90 Distce of village from dis headquarters 0.16 0.42 33.61 Vllage dummy Gorakhp 12.00 0.91 0.36

Numberof observations 136 136 Meanof dependentvaiable 41.31 41.31 Sndard deviationof dependentvariable 29.75 29.75 F-value (21,114) 1.34 (8,127) 3.41 Significanceof F-test 0.16 0.00 R-squared 0.20 0.18 AdjustedR-squared 0.05 0.12

142 UNDP-WorIdBa Waerand Saio2 Progum Table HA Statistical Analysis of Wilinness to Pay for an Improved Piped Water System- Type Bl Villages: Results of Ordny Least Squares RegressionModd DEpeldeniVaabke: Midpointof intervalin biddinggame withinwhich respondeint's WTP bid falls.

Parameter Parameter Man Indnt variables esfimate T-Ratio estimate T-Ratio values hercept , -4.94 -0.13 -7.35 -0.37 Householdsize 2.17 1.94* 2.19 2.34** 7.42 Waer consumption A4 0.75 21.23 Animals - 1.90 -1.30 -2.13 -1.63 2.02 Proportionof adult women 17.20 0.51 26.61 1.24 0.31 Prportion of chilren - 5.09 -0.23 0.39 Expenditurepercapita 0.05 1.45 0.06 2.13** 192.21 Constructionvalue of house 9.84E-5 2.42** 1.09B-4 2.99** 101,742.65 Ownemip of land or property(I if yes) 2.44 0.26 0.88 Qualityof altrative water ( if sadsfied) 21.59 1.70* 23.02 1.98* 0.94 Privatewater source (I if yes) - 1.27 -0.17 0.29 Vendorusage (I if yes) 5.33 0.39 0.05 Age -0.03 -0.13 54.06 Education 0.54 0.61 0.84 1.06 7.79 Occupation(I if farming) -4.71 -0.72 0.43 Meter(I if yes) 0.02 0.00 0.65 Free supply(I if yes) -13.66 -1.37 -15.27 -1.68* 0.89 Externalexposure(I if yes) 5.81 0.90 6.55 1.14 0.49 Awarenessof watersystems (I if yes) 8.76 1.27 10.33 1.70* 0.74 Sex (1 if male) 7.29 0.70 0.90 Distce of village fromdist headquarters -0.16 -0.38 33.61 Villagedummy Gorakhpur 6.97 0.49 11.28 1.79* 0.36

Numberof observations 136 136 Meanof dependentvariable 58.72 58.72 Standarddeviation of dependentvariable 34.46 34.46 F-value (21,114) 2.24 (10,125) 4.77 Significanceof P-test 0.00 0.00 R-squared 0.29 0.28 Adjustd R-squaed 0.16 0.22

Progran Report Se5 143 TIableH.S Statitcal Analysis of WHliges to Pay for a Standard Piped Water System -Type A VlUlages: Resultsof Ordhiary Least Squares RegressionModel DependenVaB:eg Midpointof intervalin biddinggame withinwhich respondent'sWTP bid falls.

Parameter Parameter Mean Independentvariables estimate T-Ratio estimate T-Ratio values Intercept 82.33 3.62** 74.03 4.85** Householdsize 0.59 0.78 * 7.52 Waterconsumption 0.03 0.20 22.83 Animals 0.40 0.41 2.07 Poportion of ad-It women 47.03 -1.97* 42.74 -2.08** 0.31 Proportionof children .44.22 -2.73** -38.10 -2.79** 0.37 Expenditureper capita 0.03 1.68* 0.03 1.83* 220.62 ConstuCiconvalue of house -4.26E-6 -0.13 84,458.65 Ownershipoflandorproperty(l if yes) 1.83 0.42 0.61 Privatewater source (1 if yes) -3.14 -0.41 0.08 Age -0.08 -0.60 50.45 Education 0.03 0.04 8.51 Occupation(I if fauming) -4.50 -0.85 0.23 Meter( if yes) 0.15 0.04 0.51 Free supply(I if yes) 2.90 0.50 0.87 Extemalexposure (I if yes) -3.14 -0.78 -4.16 -1.13 0.53 Sex (I if male) -9.62 -1.30 -10.96 -1.63 0.91 Distanceof villagefrom dist. headquartes -0.32 -1.41 31.26 VilUagedummy Jawa -4.95 -0.95 0.30

Numberof observations 133 133 Meanof dependentvariable 40.56 40.56 Standarddeviation of dependentvariable 21.20 21.20 F-value (18,114) 0.98 (5,127) 2.85 Significanceof F-test 0.48 0.02 R-squated 0.13 0.10 AdjustedR-squared 0.00 0.07

144 UNDP-WorldBank Water and Sanitati Pngrmm TIble 1L6 StatisticalAnalysis of Wifipess to Pay for an Imprved PipedWater System-Tlpe A Vlllages: Resultsof Ordiary LeastSquares Rerssn Model Dependnt Vibe: Midpointof intervalIn biddinggame withinwhich respondent's WTP bid falls.

Parameter Parameter Mean Independentvatiables estimate T-Ratio estimate T-Ratio values Intercept 83.49 3.12** 88.98 4.81** Householdsize 0.80 0.89 1.03 1.32 7.55 Waterconsumption 0.02 0.14 22.77 Animals 1.03 0.90 2.08 Proportionof adultwomen -28.24 -1.01 -29.15 -1.25 0.31 Prportionof children -49.31 -2.58** -50.63 -3.18** 0.37 Expenditureper capita 0.03 1.23 0.03 1.61 220.78 Constuctionvalue of house 2.OOE-5 0.53 84,643.94 Ownashipoflandorproperty(Iif yes) -1.03 -0.20 0.61 Privatewater source(I if yes) 1.16 0.13 0.08 Age -0.25 -1.49 -0.29 -1.88* 50.58 Education 0.78 0.99 8.51 Occupation(1 if farming) -4.51 -0.73 0.24 Metr ( if yes) 1.53 0.31 0.51 Frmesupply (1 if yes) -2.08 -0.31 0.86 Extnal exposure(I if yes) -6.67 -1.40 -5.93 -1.37 0.52 Sex(I if male) -6.87 -0.79 0.91 Distanceof vllagefrom disL leaquwters -0.12 -0.43 31.16 Vrllagedummy Jawa 6.19 1.01 6.13 1.31 0.30

Numberof observations 132 132 Meanof dependentvaiable 53.14 53.14 Standarddeviation of dependentvariable 25.21 25.21 F-value (18,114) 1.17 2.65 Significanceof F-test 0.30 0.01 R-squared 0.16 0.13 AdjustedR-squared 0.02 0.08

Proam Report Series 145 AppendixI

Tlbk LI Diribution d WTPBids for a PubUcTap System and a Stadard Systemwith House Connecto in Vies wihout Plas for stallaon of PubicWater Supplies

Mean*bid Publicta_s St____ard_sw (Rs) No. -,O- No. % 0 16 13.2 3 2.5 S . 3 2.5 0 0 15 20 16.5 8 6.6 25 17 14 10 8.3 35 31 25.6 25 20.7 45 12 9,9 10 8.3 55 5 4.1 28 23.1 65 4 3.3 10 8.3 85 8 6.6 15 12A 100+ 5 4.1 12 9.9

Tota 121 100 121 100

No. of bids Rs 20 82 67.8 110 90.9

Meanbid (Rs) 34.92 55.21 Note: Te followingapplies to ibles1-1 to 1-3. * Mean bidsare te mid-pointsof the intvals in whichthe rsponde's bids feil (except0, 5, and 100). All genuine zero bids are includedin 0, all bids greaterthan zero and less dtan orequal to Rs S are includedin 5;and al bids equal to rgreatertban Rs 100 are included in 100*.

146 UNDP-WorldBank Water and Sw*om Progrw Table L2 Distion of WTP Bids for a Stadad Piped Water System with House Connecto in lbp A, B51and 82 VillageI

Meanbid A Bi B2 (RS) No. % No. % No. % 0 2 1.4 1 0.7 3 2.5 5.0 3 2.1 5 3.6 0 0 15.0 1 0.7 24 17.1 8 6.6 2S.0 47 33.6 31 22.1 10 8.3 35.0 31 22.1 29 20.7 25 20.7 45.0 34 24.3 12 8.6 10 8.3 55.0 10 37.1 10 7.1 28 23.1 65.0 4 2.9 5 3.6 10 8.3 85.0 6 4.3 13 9.3 15 12,4 100+ 2 1.4 10 7.' 12 - 9.9

Total 140 100 140 100 121 100

No.of bids >Rs20 134 95.7 110 78.6 110 90.9

Meanbid (Rs) 39.07 42.16 55.21 Table U Ditibuton of WTP Bids for an Improved Piped Water System witb House Conections In TyvpesA and Di Vlage

Mean bid M_Agetp_e (Ra) No. % No. % 0 2 1.4 1 0.7 3.0 3 2.1 1 0.7 13.0 8 5.7 8 5.7 25.0 2 1.4 12 8.6 35.0 35 25 19 13.6 43.0 28 20.8 32 22.9 53.0 28 20 21 15 63.0 10 7.1 S 3.6 85.0 20 14.3 20 14.3 100+ 4 2.9 21 15

Total 140 100 140 100 No.of bids Rs20 133 95 130 92.9 Mean bid (Rs) 51.18 59.44

PW8ram Report Sewim 147 lIable I4 Connecton Frequences and Esimated Revenues: Provisionof Standard Plped Water System with HouseConnections In Types A, B1 and B2 Vllages Vlllagetype A Bi B2 Estimated Esimated Esfimated Households revenues Households revenues Households revenues Monthlytariff comected (Rs/100 connected (Rs/100 connected (Rs/100 (Rs) (%) households) (%) households) (%) households) 0 100 0 100.0 0 100.0 0 5.0* 98.6 493 99.3 497 97.5 488 15.0 96.5 1,448 95.7 1,436 97.5 1,463 25.0 95.8 2,395 78.6 1,965 90.9 2,273 35.0 62.2 2,177 56.5 1,978 82.6 2,891 45.0 40.1 1,805 35.8 1,611 61.9 2,786 55.0 15.8 869 27.2 1,496 53.6 2,948 65.0 8.7 566 20.1 1,307 30.5 1,983 85.0 5.8 493 16.5 1,403 22.2 1,887 100+** 1.5 I50 7.2 720 9.8 980 Note: Thefollowing applies to Tables 1-4to 1-6. * All bids between Rs 0 and Rs 5 are consolidatedin the Rs 5 catego . ** All bids greaterthan Rs 100 are consolidatedin the Rs 100 category.

Table LS Connection Frequenides and Estimded Revenues: Provisionof Improved Piped Water systems witHouse Connecios n TypesA and B1 Vilages ,iliae type Estimated Estimated Households revenue Households revenue Monthlytariff con d (Rs/100 comected (RsJ100 _ (Rs) (%) households) (%) households) 0 100 0 100 0 5.0 98.6 493 99.3 497 15.0 96.5 1.448 98.6 1,479 25.0 90.8 2,270 92.9 2,323 35.0 89.4 3,129 84.3 2,951 45.0 64.4 2,898 70.7 3,182 55.0 44.4 2,442 47.8 2,629 65.0 24.4 1,586 32.8 2,132 85.0 17.3 1,471 29.2 2,482 100+ 3.0 300 14.9 1,490

148 UNDP-WordBank Water d SantationPyrm TableL6 Coectio Frequenciesand EstimatedRevenues: Provion of PublicTaps and Standard HouseConnections In TypeB2 Vilages Publictaps Standardsystem Estimated Estimated Households revenue Households revenue Monthlytariff connected (Rs/100 connected (Rs/100 (Rs) (%) households) (%) households) 0 100 0 100 0 5.0 86.8 434 97.5 488 15.0 84.3 1265 97.5 1463 25.0 67.8 1695 90.9 2273 35.0 53.8 1883 82.6 2891 45.0 28.2 1269 61.9 2786 55.0 18.3 1007 53.6 2948 65.0 14.2 923 30.5 1983 85.0 10.9 927 22.2 .1887 100+ 4.3 430 9.8 980

P m ReportSeries 149 Appendix J

Table J.1 Compartve Size of Rual Localidesin the Three Evironental Zones by Populao Size (1961) % of rural localitiesby populationsize Total Zone rural >5,000 2000- 1000- 500- 200- Uninh. localities 4999 1999 999 499 <200 abited Pakistan 45,167 3.2 17 22 20.6 18.7 14.1 4.4

Punjab 25,266 3.1 18 24A 22.8 17A 10.2 3.9 (% of populaton living (16.6) (40.4) (25.7) (12.1) (4.5) (0.7) - in viage sze category)

Sweetwater zone

Sheikhupuradisict 1,090 5 21.1 23.4 19.9 15 8.6 6.9 (%of populaion living (26.1) (40.1) (20.9) (9.2) (3.3) (0.5) - in villagesize category)

Sheikhupurasubdistrict 284 11.3 44.4 24.6 12.7 5.6 0.4 1.1

Brackishwat zone

Faisalabaddistict 1,350 4.6 55.5 25.6 7.5 3.2 2.7 0.9 (% of populationliving (13.9) (66.5) (16.6) (2.4) (0.5) (0.1) - in village size category)

Faisalabadsubdistrict 259 15.4 65.6 15.4 2.3 0.4 0.8 0.0

Arid zone

Rawalpindidistrict 1,177 1.4 9.3 20.6 25.2 22.9 11.9 8.9 (% of populadonliving (12.5) (28.9) (29.4) (195) (8.3) (1.3) - in village size category)

Rawalpindisubdistrict 362 1A 8.3 185 28.7 27.9 11.9 3.3 Source: Handbookof Popuation CensusDa, Paklstm CensusOrganization, Statstics Division, Governmentof Pakistan,1985. Population census, 1981.

150 UNDP-WaoidBank Waterand Saai Phgrogi Table J.2 Sourees of Water and Light In the Three Environmental Zones (1981) (% of Housing Units with Acces) Sourceof drinkingwater Insidehouse Outsidehouse Spring/river/ Hand- Hand- sntm Elec- Zone Piped pump Well Piped pump Well Pond etc. tricity Punjab 3 37 5 3 15 16 4 17 15

Sweetwater zone Sheihupura District 2.1 72 0.8 0.5 19.1 5.2 0 0.4 23.6

Brackishwater zone FaisalabadDistrict 2.0 64.4 0.6 0.8 18.1 2.8 9.2 2.1 16.7

Ard zone RawalpindiDistrict 3.2 1.1 4.6 2.7 0.4 62.3 0.1 25A 16.7 Source: Housingcensus, 1980.Population census, 1981.

Table J.3 Occupaonal Proile of the Three Envonmental Zone Sweetwater Brackdsh-water Arid zone Zone zone zone (She-upur Dix) (FaisalabadDix) (Rwalpindi Dis.) Total woring populadon 516,838 920,700 245,440 (age 10 yearsand above)

Pecentae of woring Populationegaged in: Agriculture 54 54.9 52.9 Manufacturing 16.9 13.7 5.8 Constrction 4.4 4.2 5A Trade 6.1 6.6 6.7 Tran t 3.4 3 5.1 Services 13.4 12.2 17.6 Olhers 1.8 5.5 6.5 Source: Populationcensus, 1981.

Pram RepSees 15 Tbae J4 Fianl Cost* (its) to an Average Household for DifferentService Options (Bracish-Water Zone) Total Monthly Monthly Total Service Option capital capital O & M monthly (1) ba 1000 13 5 18

(2) Domstic connection 60(F 6 12 d 18 (3) Electricmotore 1500 20 20 40 (1+2) Handpumpand domestic 1600 19 17 36 Connection (1+3) Handpumpand electric 2500 33 25 58 Moor (1+2+3) Handpump,electric 3100 39 37 76 motor and domestic connection

Not includedin the aboveesuimates are averagecapital costs of indoorplumbing often associated with serviceoptions (1+2), (1+3), and (1+2+3):Overhead Tank, Rs 500; Indoor Piping,Rs 500; FlushToilet + SepticPit Rs 4,000-10,000. a Assumesan economiclife of 10 years; 10% real interest.Includes cost of shallowwell. b Twentyyears; 10% interest. c Connectionfee, Rs 100;Connection costs, Rs 500. d Monthlytariff paid by householdfor an unmeteredconnection. e Ten yeas; 10%interes_

* There is a lot of variationin the cost dependingupon the size of the septic tank and whether soak pit included.

152 UNDP-WaodBank Waterand S5ntation Pgran TableJ5 Households' Choice of Service Level by SocioeconomicCharaeterbtics (Vilages with Piped Water Supply) Sweetwater zone Brackish-waterzone Yearsof Yearsof education education (most (most Value educated Value educated of member of member house of house of Servicelevel (Rs) household) (Rs) household) Handpump 48,500 6 62,100 5

Handpumpand domestic 96,100 9 112,400 7 connection

Handpumpand electic 137,500 12 115,000 8 motor

Handpump,electric 208,500 12 145,200 10 motor and domestic connection

Progn ReportSeries 153 ThbleJ.6 Man WTPBlds for HouseConn_econs as a Percent of Household Income Sweet Brackish- Arid waler zone water zone zone 1. Mean WTPbid (Rs) 21 40 49

2. Date bidsobtained March August June 1988 1988 1989

3. Monthlyhousehold 1,995 1,679 1,409 income(Rs)a

4. Monthlyhousehold expendiu - 1,944 1,589 (f*omsurvey dat 1988-89)

Basedon this information:

WIPasapercentofincomeis: 1.1 2.4 3.5

WTPas percentof expenditre is: - 2.1 3.1 a Tmeincome in the survey villagesis likely to be somewhathigher than the district average becamueof proximityto the district headquarters.Also the adjustmentfor inflationwould ise income.Therefore, the percentageswould be lowerthan the alreadylow values.These are aveae rual householdincomes for the study districts from the 1984-85Household Incomeand Expendtu Survey.

154 UNDP-W Id aN* Water and Saitio Prgram TableJ.7 Costs of Village-LevelWater Supply Options (Rs) TypicalVillage with 5,000 Population (Sweetwater zone) Total Montly Monthly Total Serviceoption capital capital O & M monthly (1) Pipedwater sysem (100% of households connected) Costto PHED 1,500,000 13,500 3,800 17,300 Costto households 281,000 2,500 - 2,500 Total 19,800

(2) Actual current water 815,500 10,600 6,200 16,800 expenditues 30% - Handwumpand electric motor 70% - Handpumponly

(3) Summationof households' 11,800 willingness-to-paybids

(4) Estimatedrevenue based on 34,000 300 5,300 5,600 tariff of Rs 12.5 per month, 75% householdsconnected, and Rs 80 connectionfee.

(5) Cost of piped water system 1,125,000 10,000 2,800 12,800 to PHED for 75% households

Program Report Series 155 TableJ8 Cods of Vilge-Level Water Supply Options (Rs) Typicalvillage with 5000 populationa (Brackish-waterzone) Total Monthly Monthly Total capital capital O & M monthly Serviceopton (R) (Rs) (Rs) (Ra) (1) Pipedwater system (100%of households conected) Costto PHED 1,5WOOOP 13 ,5 00C 3,800w 17,300 Costto households 281,000( 2,5 0O - 2,500 Total 19,800

(2) Actual cunrentwater 1,084,00 14,100 g 9,800 23,900 expenditures 62% - Handpmnpand electricmotor 38% - Handpumponly

(3) Sumnation of households' 22,500 wilingness-to-paybids

(4) Estimatedrevenue basedon 35,000 300h' 11,000 11,300 tariff of Rs 25 per month, 78% householdsconnected, and Rs 80 connectionfee.

(5) Cost of piped water system 1,170,000 10S00F 2 ,900d 13,400 to PHE!Dfor 78%households a 562 houseswith 89 inhabitantsper household. b Basedon tubeweilat Rs 300 per capita capital costs. c Assumesan economiclife of 25 years, 10%real inteKst rate. d Assmes annual operon and maintenace costs equa to 3% of total capital costs (based on cost data from PHED). e Rs 500 connectioncost per household f Cost of eleclricmotor, Rs 1500;Cost of bandpump,Rs 1000. g Assumesan economiclife of 10 years, 10% real interestrate. h Computedover 25 yearsat 10%real interestrate.

156 UNDP-Wrd Ba* Wakr ad Sankado Fwgn Table J.9 Comparative Features of the Tbree EnrvironmentalZones? Sweet water Brackishwater Arid zone zone zone Aver8evilage sizeb (numberof inhabitants) 5,778 10,229F 1,085

Monthlyhousehold incomed (Rs) 1,995 1,679 1,409

Monthlyhousehold expenditue per capita(Rs) - 216 227

Constucion value of house (000 Rs) 86 134 103

Householdsize 9 9 7

Percentageof Adult womenin household 26 28 30 Childrenin household 41 37 38

Age of head of household(years) 50 51 51

No. of yearsof educationof most educated Memberof household 8 9 8 Womanin household 4 4 3

Householdsinvolved in faming (%) 27 18 31

Householdsowning animals(%) - 45 72

Householdswith land or other property(%) 63 52 75

Householdswith extemal exposure(%) 33 37 50 a Sample statisticsexcept whereindicated. b Populationof samplevillages from populationcensus, 1981. c There is one very large village in the sample (Pop. 20,586). The average excludingthis village is 8,157. d Source: HouseholdIncome and ExpenditureSurvey, 1984-85.The values are the average rurgalhousehold icomes for the three studydistricts. The comparativevalues for Punjaband Pakistanare Rs 1,533and Rs 1,545,respectively.

Proram Repot Se,a 157 Table J.10 Comparative Water-Related CharacterisfIs In the Thre Environmental Zones" Sweet water Brackish-water Arid zone zone zone Monthlvtariff for domesticconnection to piped water system(Rs)** 10 12 20

Percentageof connectedhouseholds in villageswith piped water systems 55 75 96

Essentialwater consumption (Liters/person/day) Villageswith piped water 19 31 23 Vllages withoutpiped water 24 32 23

Waterconsumption of animals (Liters/animal/day) Villageswith piped water - 48 36 Vlllageswithout piped water - 58 43

Householdsthat believe water should be suppliedfree (%) 44 52 88

Householdsthat believewater can be suppliedfree (%) 38 27 67

Householdsthat favormetering of house connections(%) 54 81 61

Householdsthat believe water supply systemsshould be managedby PHED(%) 71 66 78

* Sample statisticsexcpt whereindicated. ** Source:PHED

158 UNDP-World Bank Water and Santato Prora Table J11 Mean Wilhingness-to-PayBids for Monthly Tariff of Piped Water Sytems wit House Connectionsin the Three Zones Sweetwater Bmcish-water Aridzone zone zone (Currenttariff= (Currenttariff= (Ciunettariff= Rs IOper month) Rs 12per month) Rs 20 permonth) Villageswith piped supply MeanWTP bid for - 16 39 standardsystemn MeanWTP bid f8r 15 33 51 improvedsystem Villageswithout piped supply 1. (butwhich had a pipedsysten in the past)o =Mean WTP bid for 17 stanard system 2. (andin whichhouse- holdsknow a piped systemwill be installedsoon) - MeanWTP bid for 41 42 sandardsystem - MeanWTP bid for 58 59 im^provedsystem 3. (no pipedsystem in the past,no piped systemplanned) - MeanWTP bid for - - 35 standpipesystem -MeanWTP bid for 21 37 55 standardsystem - MeanWfP bid for 51 improvedsystem a Standardsystem refers to the kindof pipedwater system with house connections which has been installedby the PHEDin Punjab. b In the sweetwater zone the improvementconsists of the supplyof an extra4 howsof waterper day frmmthe standrd system.In the othertwo zonesit consistsof a continuouswater supply with improvedpressure and reliability. c No systemnsare planned to beinstalled in the sweet water zone. One village has an inoperative systm

Prram hRepotSeia TableJ.12 Actual and Hypothetical Frequency of Connection to Piped Water Systemswith House Connections in the Thee Zones Sweetwater zone Brackish-waterzone Aridzone (Currenttoriff= (CuTrnttaiff= (Currenttariff= Rs IOparmonth) Rs 12permonth) Rs 20 permonth) Villageswith piped supply

- Householdsactually 55 75 96 connectedat cument tariff(%) - Householdswho say they 60 95 94 wouldconnect at thecunent tariffif the pipedsystem wereimproveda Villageswithout piped supply

1. (butwhich once had a piped system) - Householdswho say they 84 - wouldconnect to a piped systemat current triff (%) 2. (andin whichhouseholds knowthat a pipedsystem willbe installedsoon) - Householdswho say they - 97 87 wouldconnect to a piped systemat thecurrent tariff(M) 3. (nopiped system in the the past,none planned) - Householdswho say they 85 90 94 wouldconnect to a piped systemat the cunrent tariff(%)

- Households that wolld - - 84 subscribeto a standpipe systemat Rs 15per month a Seefootnote 2, Tablel-I. b Whenthe systemwas operative, 69% of the householdswere connected.

160 UNDP-Wod Ban War and Sation Progran TableJ.13 Households'Willingness to Pay for Connectionto Piped WaterSystem by SocioeconomicCharacteristics (Vilages without Piped Water Supply) Sweetwater zone Brackishwater zone Aridzone Percent MeanWTP bid Percent MeanWT? bid Percent MeanWTP bid of sample (Rsper of sample (Rsper of sample (Rsper month) month) month) Educationof most educatedmember of household 0 - 8 years 44 15 38 36 48 44 9 - 12 years 41 21 41 40 48 53 > 12 years 15 33 21 47 4 55

Constructionvalue of house (Rs) 0-49,000 38 14 9 33 25 44 50,000 -99,000 40 20 22 36 31 41 100,000- 149,000 10 21 19 38 21 56 :>150,000 12 35 50 44 23 56

Occupation Non-faniing 75 21 79 40 58 53 Farming 25 17 21 41 42 43

Ownershipof animals yes NA. N.A. 49 41 80 48 No 51 39 20 53

Overall mean 21 40 49

I

PrograunReport Serw*e 161 OWRANWAIA ~~~~~~~~~~IBM233S

JRANWALA / /

r~~~~~~~~~~~

SHEKHUPURA DISTRICT

| ^t3e_ _~~ eA)/ / \ [ ~INDIA

S o AD,EALECTE C,ES - MAO RO S \ ,= - X v X ~~~LAH O REE

1~~~~~~ AK IS TA N SHEKHUPURADISTRICT a=5A.

0 SEtECrDMES - MAJORRODS;f - DISRTICTBOUNDAES -O-TER ROADSa INTNATIONALBOUNDARIES -RAILROADS

NOEWIJIB99 t ~~~~PA K I S T A N H iKUPUR FAISALABADDISTRICT

= FAJSALABADDISTrICT 7\ // SELECTMCITIEs RNMS

MAJORROADS JHANG

- - OTHERROADS

- RMLROADS SA AB -

DISTRICTBOUNDARIES0 B

4~~~~~~~~~~~~~~~~~~~~~~~

u~~~~~1 Stw I 2335

Havipw .~~~~~~~~~JMMU NORTHWEST AND FRONTIERKAHR

9~~~~~

{ { . \ s~~ISLAMABA D

JHLUM

P A KI S TA N ~ / M RAWALPINDIDISTRICT X

I=RAwAI.PD axsTmcr --- RiVER smoEcT- MAJORROADS rSIMCr BOUNDARIES OTH ROADS 0 2 4

- PROVINCEBOJNnAIes -RAI.ROADS __

NOMER 1991 UNDP-WrldSBna.iter ndSntioPrga

The World Bank -181.8 H Street, NW, WAashington.DC 20433 USA United Nations Development Programme One U'nitedNations Plaza NewYork. NY J10017 L'SA Regional Water and Sanitation Groups Regional Water and Sanitation Network Eastern and Southern Africa CerntralAmirica c/o The World Bank c/o UNICEF P.0. Box301577 P.O.Box 525-01901 Nairobi, Kenva Ciudadde Guatemala WestAfrica c/o The World Bank 01 B.P. 1850 Abidjan 01-,C6te dflvoire EastAsia and the Pacific C/oThe World Bank P.O. Box 324/JKT Jikarta, Indonesia South Asia CIO TheWorld Bank P.O.Box'416 NewDelhi 110003-.lndia

- k/I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~-