Geoforum 38 (2007) 855–868www.elsevier.com/locate/geoforum

Trickle Down? Private sector participation and the pro-poorwater supply debate in Jakarta, Indonesia

Karen Bakker

Department of Geography, University of British Columbia, Room 217, 1984 West Mall, Vancouver, BC, Canada V6T 1Z2

Received 10 August 2004; received in revised form 11 November 2005

Abstract

Private sector partnerships (PSPs) have been increasingly advocated as an instrument of ‘pro-poor’ water supply policies. This articleexamines the performance of the private sector with respect to network connections for poor households in Jakarta, Indonesia, drawingon three sources: data collected through a household survey of poor households in six Jakarta neighbourhoods in 2005; data provided bythe two private concessionaires and the Jakarta municipal government; and interviews with water supply managers, government oYcials,and NGO representatives in 2001 and 2005. The analysis concludes that the Jakarta PSP contract has not been pro-poor: new connec-tions were preferentially targeted at middle and upper-income households over the period 1998–2005, and the numbers of new connec-tions have been lower than the original targets. The paper argues that the failure to connect the poor is not solely attributable to theprivate operators, and identiWes disincentives to provide individual network connections to poor households on the part of the municipal-ity, the private concessionaires and poor households. The paper concludes by questioning the long-term ability of private sector operatorsto supply water to the poor.© 2006 Elsevier Ltd. All rights reserved.

Keywords: Water supply; Privatization; Jakarta; Indonesia; Poverty; Pro-poor; Development

1. Introduction: Private sector participation and the‘pro-poor’ debate

Proponents of private sector participation (PSP) inwater supply have argued that PSP is a means of improvingservice delivery to the poor (see, for example, Cross andMorel, 2005; Nickson and Franceys, 2003),1 critical in aworld in which an estimated one billion people – the‘unserved’ in development jargon – lack access to safe, suY-

E-mail address: bakker@geog.ubc.ca1 See, for example, the Global Water Partnership (www.gwpforum.org/)

and the World Water Council (www.worldwatercouncil.org/), two influ-ential networks of private water companies, governments, and lendingagencies. Also see the Business Partners for Development, which links theWorld Bank with private water companies and governments, and ‘aims toproduce solid evidence of the positive impact’ of PSPs (www.bpd-web.org).

0016-7185/$ - see front matter © 2006 Elsevier Ltd. All rights reserved.doi:10.1016/j.geoforum.2005.11.011

cient water supplies (WHO, 2000). SpeciWcally, througheYciency gains, improved management, and better accessto Wnance than public utilities, private companies improveperformance (including cost recovery rates) and increaseaccess through extending networks and providing new con-nections to previously ‘unserved’ customers. This will ben-eWt the poor, particularly in urban areas, who are oftenserved by a variety of informal arrangements such as watervendors, and typically pay much higher prices per unit vol-ume for poorer quality water than wealthier consumers(Johnstone and Wood, 2001; Shirley, 2002; World Bank,1994, 1997, 2004a).

Opponents of private sector participation argue thatPSPs are not reliable mechanisms to supply water servicesto the poor, because private companies are unable to sup-ply the poor on proWtable terms. As proof, critics point tothe withdrawal of the private sector from contracts andregions of the world, in light of risk-return ratios whichhave remained unacceptably high (Hall and Lobina, this

856 K. Bakker / Geoforum 38 (2007) 855–868

volume; Hukka and Katko, 2003; Smith, 2002),2 in partbecause of the low ‘ability to pay’ of poor consumers.Moreover, some critics argue that the potential contribu-tions and sustainability of private sector involvement willbe undermined by political conXict and civil society resis-tance to PSPs arising from the belief that water is ahuman right. Indeed, mobilization of social movements inopposition to PSPs has occurred in many countries, attimes resulting in the cancellation of contracts by govern-ments (La Paz and Cochabamba, Bolivia). In other cases,such as Manila, the opposition of social movements hasfactored into the decision of the private company to with-draw (Barlow and Clarke, 2002; McDonald and Ruiters,2005; Shiva, 2002; Trawick, 2003; Swyngedouw, 2005;WaterAid, 2003). Many of these critics argue that PSPsare not ethically appropriate, and call for the managementof water as a commons, often with reference to idealizedmodels of indigenous water governance. Even where crit-ics agree, in principle, to the management of water by theprivate sector, they argue that political conXict over thesocio-economic identity of water will further elevate risks,and decrease the likelihood of the private sector beingable to supply the poor on a proWtable basis.

Recently, multilateral Wnancial institutions seem to beaccepting some of these claims. In a recent study of privatesector provision of urban services provision to the poor, forexample, the Asian Development Bank states that: “the pri-vate sector is not willing or able to solve the problems ofunserved areas on its own” (ADB, 2003a, 56). Analysis ofthe discourse of the public statements of senior executivesof water supply services Wrms reveals a retreat from earliercommitments to pursuing PSPs globally, with senior Wgurespublicly acknowledging high risks and low proWtability insupplying the poor (Robbins, 2003). Britain’s inXuentialeconomic weekly The Economist has warned of a ‘retreat ofthe private sector’ from water supply in developing coun-tries (The Economist, 2004). These developments haveraised concerns that the contribution of the private sectorto reaching the UN’s Millenium Development Goals (thenew goalposts of the international development agenda) forwater will be relatively limited.3 In response, a debate abouthow best to implement a ‘pro-poor’ agenda through PSPshas arisen amongst consumers, governments, donors, and

2 For academic studies critical of the privatisation process, with a focuson developing countries, see the Municipal Services Project website (http://qsilver.queensu.ca/~mspadmin). For an international public sector unionperspective, see the very comprehensive PSIRU website (www.psiru.org).For a campaigning NGO perspective, see the Council of Canadians BluePlanet Project (www.canadians.org/blueplanet/index2.html) and the US-based Public Citizen’s campaign on water supply (http://www.citizen.org/cmep/Water/).

3 The eight United Nations Millenium Development Goals were agreedupon at the UN Millenium Summit (2000). The MDGs set a speciWc targetfor water supply: reducing by half the proportion of individuals withoutsustainable access to adequate quantities of aVordable and safe water by2015. See http://www.un.org/millenniumgoals/.

private water companies (see, for example, Franceys andWeitz, 2003).

This article examines some of the main claims made inthese debates through a case study of Jakarta Indonesia,one of the largest water supply PSP contracts in the Southto date.4 The Wrst section of the paper brieXy describeswater supply access in Jakarta, which has one of the lowestrates of water supply and sanitation services provisionacross Southeast Asia. The second section summarizes keyaspects of the Private Sector Participation (PSP) contractsigned in 1998, highlighting some of the key weaknesses ofthe contract. It also discusses relevant aspects of the rene-gotiation of the contract following the Asian economic cri-sis, the devaluation of the rupiah, and fall from power ofex-President Suharto. Building on this background, thethird section examines the question of whether the privatesector contract has been pro-poor. Here, the paper makestwo related arguments. First, the paper documents the factthat the Jakarta PSP contract has not been pro-poor: thenumbers of new connections have been lower than the orig-inal targets, and new connections were preferentially tar-geted at middle and upper-income households over theperiod 1998–2005. Second, the paper argues that this is notsolely attributable to the private operators, and identiWesdisincentives to provide individual network connections topoor households on the part of the municipality, the privateconcessionaires and poor households. In contrast to muchof the literature on ‘pro-poor’ water supply, which assumesthat these disincentives are largely economic, the paperargues that a range of non-economic factors must also betaken into account, and documents these factors based on ahousehold survey of 110 households in six neighbourhoodsin Jakarta conducted in mid-2005. The fourth section con-cludes with critiques of some of the core assumptions ofcurrent debates on pro-poor water supply services in urbanareas.

2. Watering Jakarta

In contrast to many other Asian cities, Jakarta has rela-tively little water supply and wastewater disposal infra-structure, and is characterized by lower rates of urbanservices provision than other national capitals. The WorldBank characterizes its water and sanitation sector as oneof the weakest in Asia (Brennan and Richardson, 1989;Leitmann, 1995; McGranahan et al., 2001; World Bank,2004). The oYcial (generous) estimate is that 56% of thecity’s residents are connected to the network (2004 Wgure)

4 This analysis is based on research carried out over the period of severalmonths in two research visits to Jakarta in 2001 and 2005. Data collectedthrough a household survey of poor households in six Jakarta neighbour-hoods in 2005; data provided by the two private concessionaires and theJakarta municipal government; and interviews with water supply manag-ers, government oYcials, and NGO representatives in 2001 and 2005. Per-mission to cite archival documents collected by Michelle Kooy duringresearch visits to Leiden and Amsterdam in 2003 is gratefully acknowl-edged.

K. Bakker / Geoforum 38 (2007) 855–868 857

(Jakarta Water Supply Regulatory Body, 2004).5 Thewater delivered through the network is not potable; medi-cal studies repeatedly Wnd faecal coliform contamination,and residents are advised to boil their water. Due to poorquality, low pressure and incomplete coverage of the net-work, most residents depend on a variety of water sources,including deep and shallow wells, water vendors, and bot-tled water (Berry, 1982; Gilbert and James, 1994; Loveiand Whittington, 1993; McGranahan et al., 2001). Domes-tic per capita water consumption is estimated to bebetween 70 and 80 L (one of the lowest of the 18 largeAsian cities surveyed) (ADB, 2003b). Less than 2% ofhouseholds are connected to a sewerage system (ADB,2003b); the vast majority of wastewater is disposeddirectly to rivers, canals, or to (often poorly functioning)septic tanks (Crane, 1994; McIntosh, 2003; Surjadi, 2002).Contamination by wastewater and industrial eZuent, aswell as salinisation due to seawater inWltration due toover-pumping have, in turn, polluted the shallow aquifer –the sole household source of supply for many poorer fam-ilies – in many areas of the city. Rivers and canals aresometimes too polluted to use even for washing clothing.The public health impacts of this situation are predictable,and have been well-documented: high rates of water-related diseases, including gastrointestinal illness due tocontaminated water and parasite-related illnesses due topoor drainage, particularly in poorer areas (Agtini et al.,2005; Leitmann, 1995; McGranahan et al., 2001; Siman-juntak et al., 2001; Surjadi, 2003).

International Wnancial institutions argues that one of themost important factors contributing to low levels of waterservices provision in Jakarta is the low levels of infrastruc-ture Wnance (see, for example, Akhtar, 2005; World Bank,2004), exacerbated by the Asian Wnancial crisis and cur-rency devaluation. Initiatives such as the Indonesia Infra-structure Summit (held in Jakarta in early 2005) haveexplicitly targeted foreign direct investment. The govern-ment has identiWed a signiWcant shortfall in Wnancingrequirements for rehabilitation and extension of urbaninfrastructure.

Jakarta’s infrastructure ‘crisis’ is, it should be noted, nota new phenomenon, nor is it solely a crisis of Wnance. Thecurrent lack of funding for services is compounded by thelegacy of deliberate under-investment during the 1960sand 70s, as policy-makers sought to discourage rural–urban migration, and articulated a policy stance that sew-erage was a ‘private concern’ (Argo, 1999; Cowherd, 2002).Moreover, in the post-Independence era, high priority was

5 Figure calculated using 2002 data from annual reports of the two pri-vate concessionaires operating in Jakarta. This was cross-referenced withADB (2003b) which reports a Wgure of 51.2%. Coverage ratios should beunderstood as rough estimates; their calculation is dependent upon a num-ber of variables which are only imprecisely measured, such as urban popu-lation and average size of household. Reported Wgures vary signiWcantly,and do not indicate the number of households which have a connectionbut which rely primarily on other sources (e.g. groundwater) due to qualityor service concerns (e.g. low pressure).

given to an urban redevelopment agenda focused on highlyvisible infrastructure (avenues, highways, parks, monu-ments, and sculptures). This ‘monumentalist’ infrastruc-ture served a dual purpose of urban development andsource of pride for the new nation, whereas urban servicesprovision and the urban environment were given relativelylow priority (Chifos and Hendropranoto, 2000; Firmanand Dharmapatni, 1994; Ford, 1993; Kusno, 1997; WorldBank, 2004b), despite sporadic national government-leddevelopment plans to accelerate service delivery (Silas,1992). In the case of Jakarta, rapid population growthexacerbated these factors; Jakarta City alone grew from1.8 million people in 1950 to 6.5 million in 1980, withequally rapid population growth in the surroundingmetropolitan areas (with the total population of thegreater metropolitan area now estimated at 18 million),implying rapid changes in land use on the expandingrural–urban fringe (Chifos, 2000; Firman, 1997, 1998,2000; Lo and Yeung, 1996).

Weak governance of water supply is a partial explana-tion for this situation. Jurisdictional fragmentation hasreduced the ability of any one level of government inJakarta to eVectively govern water resources within awatershed, or even within urban boundaries.6 Municipalgovernance structures are another factor. In Jakarta (aswith other cities in Indonesia and indeed around theworld), water utility budgets were not ring-fenced from thatof the municipality. Rather, a small tax base and presenceof few alternative revenue-generating activities for themunicipal government encouraged the use of water utilityrevenues for non-water related expenditure by municipalpoliticians and managers. For many water supply utilitiesin Indonesia, this had the eVect of reducing the amount ofrevenues available to cover operating costs and fund capitalexpenditure (notably infrastructure rehabilitation andimprovement), exacerbated by relatively low cost recoveryrates. Like water supply utilities across the South, Indone-sian water providers are often caught in a vicious cycle: lowcost recovery, low revenue, low investment, and low levels

6 In the Jakarta region, for example, the majority of the JMA is a politi-cally constituted as an independent territory with a status of a province –‘DKI Jakarta’ (Special Capital Region of Jakarta). The city governor isindependent from West Java province, and (together with the municipalgovernment) controls the city’s water supply company: PAM Jaya. Theprovince of West Java is responsible for the urban areas which fall outsideof DKI Jakarta, and for the watershed in which the main Jatiluhur reser-voir for Jakarta’s water supply is sited (well upstream from the city). Envi-ronmental and urban planning regulations are not systematically appliedwithin the watershed, and the open canals which act as conduits for Jatilu-hur water are polluted by residential and industrial eZuent, posing seriouswater quality challenges to the municipal water supply utility engineers.Meanwhile, within the city, tackling groundwater pollution is complicatedby the division of responsibility amongst the sewerage authority, the mu-nicipal water utility (which controls networked water supply), and the na-tional government’s Ministry of Mines, which bears responsibility forregulating deep (i.e. drilled) wells, from which a substantial proportion ofthe city’s residents draw water.

858 K. Bakker / Geoforum 38 (2007) 855–868

of service (Bakker, 2003a; Cross and Morel, 2005; Nunanand Satterthwaite, 2001).

In summary, a shortage of Wnance and weak governancehas been identiWed by IFIs as two key failings at the heartof water supply sector in Indonesia. The result, as in manyother countries, has been a twinned response: the supportof private sector participation in the water supply sector inthe hopes of attracting increased private investment; andbroad-based market-oriented water sector reform, includ-ing a controversial new Water Law establishing tradablewater rights and redeWning water as an economic good(Jakarta Post, 2003; World Bank, 1999, 2004b, 2005). Thesedevelopments are in line with the evolution of governanceframeworks internationally over the past two decades, inwhich state authority has been increasingly delegated tonon-state (usually private sector) actors (Pierre, 1995, 2000;Rogers and Hall, 2003), characteristic of neoliberal fra-mings of solutions to environmental problems (for recentcritiques of neoliberal approaches to water, see Bakker,2005; Goldman, 2005; Haughton, 2002; McDonald andRuiters, 2005; Swyngedouw, 2005).

As discussed in the following sections of the paper, thefocus on infrastructure, Wnance, and national governanceinstitutions overlooks some of the most important factorsaVecting water supply access: local urban governance,consumers’ entitlements, urbanization patterns, andurban socio-economy. Moreover, the characterization ofthe urban ‘unserved’ as eager consumers, with a high will-ingness-to-pay, obscures important barriers – ecological,political, and economic – to the extension of the watersupply network. The following sections present an analy-sis of the impacts of consequent ‘mis-reading’ water con-sumers, focusing on the private sector participationcontract under which the city’s water supply system is cur-rently managed.

3. Going private: the water supply concession contractfor the City of Jakarta

Jakarta’s government exhibited a renewed interest in theurban environment and services provision in the 1990s, astypiWed by the then-governor’s favourite slogan forJakarta: ‘Bersih, Manusiawi, Wibawa’ (Clean, Humane,Powerful) (Leaf, 1996). Concerns about the poor level ofservice in the water sector had persisted for decades, andwater shortages and water quality problems were perceivedto be increasingly acute (Berry, 1982; Lovei and Whitting-ton, 1993; Gilbert and James, 1994; Indonesia Times, 1996).One response (in Jakarta as in other Indonesian cities) waslimited private sector participation: out-sourcing of routinerepairs, billing and payment collection by Jakarta’s watersupply utility, PAM Jaya (Mandaung, 2001). Water supplywas one of many PSP initiatives ongoing in the country; theIndonesian government had passed legislation enabling pri-vate sector participation and privatisation for most publicsector utilities in the mid-1990s, and had embarked on pri-vate ventures in various sectors over the past decade, for

example privately funded toll highways throughout thegreater Jakarta area.7

Discussions regarding a long-term PSP concession con-tract with foreign Wrms began in the mid-1990s. Interna-tional water companies were keenly interested in enteringthe water services market in Indonesia, as a large, middle-income country with an expanding middle class with rela-tively low penetration of networked water supply services.After protracted negotiations, ‘cooperation agreements’ forthe management and expansion of Jakarta’s water supplysystem were awarded in late 1997 to two of the largestwater services companies in the world8: (British) ThamesWater International and (French) Ondeo (Suez-Lyonnaisedes Eaux). The process of awarding the contract forJakarta’s water supply was characterized by what the polit-ical science literature deWnes as ‘collusive corruption’(where government and private sector oYcials collude todeprive the government of revenues) (Bardhan, 1997;Shleifer and Vishny, 1993). No public tendering processoccurred; rather, international water companies put for-ward unsolicited proposals directly to the government.Under then-President Suharto, partnership with an Indone-sian Wrm was a prerequisite for international corporationshoping to take over the operations of a utility network.This is not unusual in the international water supply sector,in which private sector consortia typically have localminority shareholders. In the case of Indonesia, however,these private sector consortia were frequently linkeddirectly to the President; by the early 1990s the large Indo-nesian conglomerates had “already [become] active withinother public service areas, and these groups expected tobeneWt from the privatization of water services” (Baye,1997, 201). In the case of Jakarta’s water supply system, thetwo international Wrms were partnered with two local pri-vate Wrms, respectively members of two of the most impor-tant conglomerates in Indonesia – Salim Group (run by aBob Hassan, a crony of then-President Suharto) and SigitGroup (run by the Sigit Harjojudanto, Suharto’s eldestson).9

7 Private sector participation contracts in water supply have been signedin several other Indonesian cities: Bali; Batam; Medan; Lhok Seumawe;Sidoarjo; and Pekanbaru (Baye, 1997; ADB, 2003).

8 Sanitation services were not included in the contract, and remain theresponsibility of the various municipalities that make up the greater Jakar-ta area.

9 Corruption in Indonesia is internationally recognized as being particu-larly pervasive (Transparency International). ‘Market consumption’ and‘parochial consumption’ (where the latter hinges on kinship, caste, etc. andthe former on wealth) (Scott, 1969) were conXated in a system that came tobe known in Indonesia by the triad of ‘Corruption Collusion Nepotism’popularized as an acronym (Korupsi, Kolusi dan Nepotisme or KKN)which came to symbolize the Soeharto regime (Robertson-Snape, 1999).That the contracts were awarded despite national laws prohibiting foreigninvestment in drinking water delivery (Law No. 1/1967; Ministry of HomeAVairs Decision No. 3/1990) and local regulations (No. 11/1992 and No.11/1993) precluding private sector involvement in community drinkingwater supply was to be a source of conXict in the early years of the con-tract (Argo and Firman, 2001).

K. Bakker / Geoforum 38 (2007) 855–868 859

In January 1998, each consortium signed a 25-year con-tract with PAM Jaya, the municipal water supplier inJakarta, which retained ownership of the water supplyassets. The private consortia were to be responsible foroperation of the water supply system, including administra-tion of the customer database and billing. Thames’ contractallocated the company the exclusive right to operate andmanage the existing water supply system in the eastern halfof the city,10 supplying 2 million people connected to thesupply system out of a potential customer base of 5 million.Simultaneously, Lyonnaise des Eaux was given a contractto supply the western half of the city (Fig. 1), covering aslightly larger number of potential customers. Ambitioustargets were set: the private companies committed to reach-ing universal coverage by 2023 and to supply potable waterto consumer by 2007.

The contracts were expected to be lucrative for both thelocal and international partners. Under the terms of the con-tract, this proWt was not to be linked directly to the revenuesof the municipal water supply system. Instead, each consor-tium was to receive a fee on the basis of volume of water sup-

10 Indonesia Times (1998) “Privatised water supply begins soon” January16th, p. 3.

plied and billed, not on the basis of the water tariV (set by themunicipality), or the percentage of cost recovery. With nodirect equity stake, and with proWt de-linked from cost-recovery rates, the international water companies thussought to minimise the risk inherent in cost-recovery. Anadditional safeguard was built into the payment mechanism:an indexation formula, linked to the rupiah-US dollarexchange rate and the (Indonesian) inXation rate was builtinto the ‘water charge’ formula used to determine paymentsmade to the private operators – who are paid according tounit volume of water delivered to the distribution networkrather than billing revenue. Cost recovery and currency risks,in other words, were to be borne by the local government.

3.1. Re-regulation: tariVs, proWts, and the re-negotiation of the contract

The political and economic turmoil that unfolded inIndonesia in 1998 vitiated these strategies. Riots, the resigna-tion of Soeharto, and the abrupt and dramatic devaluationof the Indonesian rupiah11 threw the country into a period

11 From approximately 2,300 Rupiah/US$ in 1997 to 10,000 Rupiah/US$ in 1998.

Fig. 1. Jakarta’s private sector participation contract: Water supply service areas.

860 K. Bakker / Geoforum 38 (2007) 855–868

of chaos. After a tense interlude in which senior expatriatemanagers of the private concessionaires Xed the country,local managers cancelled the PSP contracts, and senior Brit-ish and French executives and diplomats pressured the fed-eral government to have the contracts reinstated, the privateconcessionaires resumed operations (having discreetly aban-doned their Indonesian partners, now tainted by their asso-ciation with ex-President Soeharto) (Harsono, 2005).

Confronted with public protest over rising prices of sta-ple food items and gasoline, the municipal governmentrefused to raise tariVs to compensate for the devaluation ofthe rupiah. This delay in tariV increases should not, in the-ory, have posed diYculties for the private water companies,as revenues are determined by a ‘water charge’ paid perunit volume of water delivered into the network. Thismeans that revenue of the private operators is not linked toamounts billed or collected from consumers. In otherwords, the revenue of the private concessionaires is, in the-ory, independent of cost-recovery as well as tariVs. Indexingthe water charge to the Rupiah–USD exchange rate pro-vided protection against currency devaluation; as therupiah fell in value, revenues from the water charge(expressed in rupiah) should have risen accordingly.

The limitations of this strategy were revealed whenreceipts in dollar terms plummeted from 1998 onwards.Given political unrest in Jakarta, the Governor was unwill-ing to implement agreed-upon tariV increases. The gapbetween the water charge required for compensating theprivate companies and the average water tariV increaseddramatically. Whereas, the water charge paid to the privateoperators was 11% below the average tariV in 1997, it roseto over 60% above the average tariV in early 2001. Subse-quent tariV increases did not raise the tariV above the watercharge until early 2004 (Jakarta Water Supply RegulatoryBody, 2004). The result was that the amount charged by theprivate concessionaires – via the water charge – to the gov-ernment increased dramatically, while revenue fell just asdramatically. PAM Jaya (and thus the local government)bore the sole risk for the revenue shortfall, and becameincreasingly indebted to the private companies. The cumu-lative deWcit by the end of 2001 was Rp 469 billion (approx-imately $46 million USD) and had reached Rp 990 billion(approximately $97 million USD) by September 2003 –excluding late payment interest and retroactive tariVincreases (Jakarta Water Supply Regulatory Body, 2005).

The time period for repayment of this debt by PAMJaya is likely to be protracted. With the fall in the value of

the rupiah, its operating revenues fell approximately four-fold in dollar terms. PAM Jaya’s revenue can be expectedto be on the order of 400 billion rupiah per year (approxi-mately one-twentieth of the outstanding ‘debt’); the negoti-ated tariV increases are likely to be less than 10% per year.Thus, although tariVs were raised and will continue toincrease, these increases will not generate suYcient revenueto quickly repay the ‘shortfall’. By mid-2001, the prospectof slow repayment of the stillincreasing ‘debt’ provoked arenegotiation of the contract, transforming it into a man-agement contract – with a guaranteed internal rate ofreturn of 22% – rather than the original concession agree-ment.12 Technical targets have been dramatically scaledback (Table 1), most notably, the commitment to providepotable water supply at the point of consumption wasdropped.

4. Connecting the poor?

Implicit in the original technical target of 100% servicecoverage, and explicit in public justiWcations of the PSPcontracts, was the belief that private sector participation inwater supply would lead to a higher rate of connection ofpoor households. Service coverage has increased since 1998,but the distribution of new connections has not been ‘pro-poor’, if this is deWned as a rate of connection equal orgreater to the percentage of poor in the urban population.

An important goal of the original concession agreementwas the extension of the network and increase in coverage,for which targets were speciWed in the original contracts.By 2002, however, service coverage levels for the two

12 As with many such contracts, proWts are ‘backloaded’. The InternalRate of Return is calculated over the lifetime of the contract, and is, todate, negative.

Fig. 2. Customers per tariV band (%, 2003). Source: PAM Jaya (2004).

Table 1Original versus renegotiated technical targets (1998–2004). Source: PAM Jaya (2004), TPJ and Palyja annual reports, ShoWani (2003)

Number of connections unit Service coverage ratio %

TPJ Palyja TPJ Palyja

Baseline (before privatization) 231,607 176,980 52.00 38.00Original targets (1997) 361,607 395,522 70.00 70.00Revised targets (2002) 335,413 301,048 62.00 45.00Realization (2002) 336,550 312,879 62.17 44.17

K. Bakker / Geoforum 38 (2007) 855–868 861

concession areas remained just above 50%, well below the70% target speciWed for 2002 in the initial contract (GlobalWater Report, 2002) (Table 1). New connections haveoccurred, but these have not targeted poor customers inproportion to their representation in the urban population.Fig. 2 illustrates the disproportionate weighting of con-sumer connections in middle-income tariV bands in 2003;whereas the majority of residents in Jakarta would fall intothe ‘lower middle’ and ‘low income’ categories, 87% of net-worked connections are provided to tariVs for middle-income households or above. This is, to some extent, thelegacy of public sector management, attributable to unwill-ingness by the municipally managed utility to extend thenetwork into poor areas due to fears about low cost recov-ery (Taylor, 1983a,b), and to a tariV pricing policy in whichwater rates for public hydrants (used by poor householdsand water vendors) was higher per unit volume than waterrates for individual households – implying a reduction inrevenue when a poor household was connected to the net-work (Crane, 1994).

This legacy of the public sector under-provisions of indi-vidual household connections to poor customers was not,however, redressed by the private concessionaires. Table 2provides data on the numbers of new consumers connectedin each tariV band by one of the two private concessionaries(Thames PAM Jaya) over the period 1998–2004. Only 25%of new connections were targeted in the two lowest tariVbands (public hydrants, intended to serve those withouthousehold connections; and ‘poor’ and ‘very poor’ house-holds). In other words, three quarters of new connections

were for middle-income and upper-income households,government enterprises and commercial enterprises.

Given that the private concessionaires are paid via a‘water charge’, which is linked to volumes of water deliv-ered into the water supply system but is independent of rev-enues and tariVs, this bias towards wealthier consumersmight seem surprising. There is no apparent direct disincen-tive to the private concessionaires to connect low-incomehouseholds. Why, then, were customers in the lowest tariVbands less likely to be connected? An important part of theexplanation lies in the pricing levels of the tariV bands(Table 3). The lowest tariV (May 2005 data) is 550 Rupiah/m3, well below the production cost (of approximately 3000m3).13 Increasing the number of connections in the lowesttariV band thus decreases the average revenue per cubicmetre supplied. Reducing the average revenue per cubicmetre by connecting poor households would lower themunicipality’s revenues, in turn reducing their ability to paythe water charge, and to repay the debt shortfall owed tothe private operators. A secondary disincentive is the higheraverage cost per connection in poor neighbourhoods,which raises installation costs: given the lack of land-useplanning in informal settlements, the highly dense anddisordered distribution of homes means that installing con-nections may be more time-consuming (if conventionalbelow-ground infrastructure is used).

13 Interview with Alizar Anwar, Advisor to the Jakarta Regulatory Body,May 2005.

Table 2New connections, East Jakarta 1998–2004, by tariV band. Source: Thames Pam Jaya, personal communication (May 2005)

TariV group #New connections % Increase

I Social institutions and public hydrants 1101 1II Public hospitals, poor and very poor households 21,898 24IIIa Middle-income households and small-scale businesses 51,847 58IIIb Upper middle income households and government oYces 11,150 12IVa Large hotels, highrise buildings, banks and factories 2323 3IVb Harbour/port 1849 2

Total 90,167 100

Table 3TariVs, Wxed charges, and tariV increases (2003–2005). Source: PAM Jaya (2004); DKI Jakarta (2005)

Average tariVs per tariVband Rp/m3 (2005)

Monthly Wxedcharges (Rp) (2005)

TariV group description (2003–2005) % Customers pertariV band (2003 data)

% Increase intariVs (2003–2005)

I 550 4695 Social institutions (e.g. religiousfacilities) and public hydrants

1.0 47

IIa 550 5060 Public hospitals and very poorhouseholds

11.7 47

IIb 2450 10,440 Low income households 46.5 44IIIa 3500 11,950 Middle income households and

small-scale businesses19.9 59

IIIb 5100 19,390 Upper middle income householdsand government oYces

14.7 32

IVa 9750 19,390 Large hotels, highrise buildings,banks and factories

5.1 48

IVb 11,500 27,665 Harbour/port 1.0 31

862 K. Bakker / Geoforum 38 (2007) 855–868

The municipality thus has two direct incentives not totarget poorer neighbourhoods for new connections. This isan important explanation for why the utility, when underpublic management, did not connect poor customers. Inturn, this produces an indirect disincentive for the privateoperators to connect poor customers: the revenue receivedby the municipality is the source of funds from which theprivate operators are paid, and decreasing revenues impliesa greater chance of debt, longer repayment period, andincreased possibility of municipal default.

The perverse disincentives built into Jakarta’s water sup-ply tariV structure are an example of how pricing strategiesintended to increase access have counterproductive goals(Whittington, 1992). The remedy, as most commonly pre-scribed by international Wnancial institutions, is to increasetariVs (a seemingly counterintuitive strategy), therebyremoving the disincentive for connecting poor consumersand providing more capital to Wnance new connections (e.g.Azdan, 2001; Yepes, 1999). This recommendation is sup-ported by studies, which assert that ‘willingness-to-pay’ and‘ability-to-pay’ of poor customers is higher than previouslythought. Frequently, the higher rates per unit volume paidby poor customers relying on water vendors are cited asevidence for this argument (e.g. Soto Montes de Oca et al.,2003; Winpenny, 1994). Indeed, the response to the prob-lem of low tariVs in Jakarta has been a series of negotiatedtariV increases, which have disproportionately raised tariVsfor poorer and middle income groups (Table 3).

The above argument is Xawed for at least two reasons.First, it overlooks the fact that water customers are ‘pricetakers’ rather than ‘price makers’. Water vendors typicallyoperate as spatial monopolists; in Jakarta, vendors do notcompete, but rather collude to establish monopoly supplyzones and a captive clientele (Susantono, 2001). Informa-tion about willingness-to-pay can not be extracted, I wouldargue, in this context – particularly where lack of surface orground water availability in urban neighbourhoods makeswater vendors the only source of water (apart from bottledwater, which are even more expensive per unit volume).

Second, this argument overlooks the fact that monopo-listic control of water vending creates a political barrier tonetwork expansion to the poor. Water vending is controlledby a complex network of middlemen running tankers,ambulatory water vendors and public standpipes connectedto PAM Jaya’s network (Lovei and Whittington, 1993).The monopoly rent extracted from the cities’ poor repre-sents an attractive source of proWts. Indeed, the potentialproWtability of extracting rent from the captive market ofwater consumers is recognized through the practice of sell-ing informal ‘licenses’ amongst water vendors (Susantono,2001). In Jakarta, as in other cities, this monopolisticbehaviour is sometimes linked with organized crime, and attimes characterized by intimidation (if not outright vio-lence) of customers, competing vendors, and police and cityoYcials who attempt to eradicate informal water vendingpractices (e.g. Swyngedouw, 1997). To put it crudely, maWa-like control of water vending in poor areas of the city is a

signiWcant barrier to network expansion, which an increasein tariVs will not address. This is signiWcant, as surveys havefound that approximately one third of Jakarta’s house-holds purchase water from street vendors (World Bank,1993; Crane and Daniere, 1996, 1997; and survey by authorin June/July 2005) (Table 4). This implies that solutions tothe problem of water supply in Jakarta must also addressissues of governance – particularly in the context of anapproach to state power and urban planning in Jakarta inwhich state activities are often geared towards the reaYr-mation of prestige and reinforcement of networks ofpatronage, rather than public welfare per se (Cowherd,2002; Kusno, 1998, 2000).

A third critique of the ‘raise prices to improve access tothe poor’ argument arises in situations where, as in Jakarta,other attractive sources of water are readily available. Theexistence of shallow and deep aquifers in the city meansthat groundwater is a viable alternative to networkedwater, for both wealthy residents (who rely on cleaner, butmore expensive deep wells) and poor residents (who rely onshallow wells, often contaminated by urban runoV andsaline due to seawater intrusion). Reliance on groundwateras one of multiple sources of water is common in poorareas of the city (Table 4); however, depletion of groundwa-ter has led to salinisation in some areas, rendering waterunWt for drinking and cooking (Braadbaart and Braadba-art, 1997). Reliability (in contrast to low pressure and inter-mittent Xow in the networked water supply system), lowcost, and quality are important factors determining con-sumer preferences for groundwater. Groundwater and wells

Table 4Household sources of water supply in six Jakarta kampungs (2005).Source: Household Survey, June/July 2005, n D 110

Total % exceeds 100 because some households use multiple water sources.

Water source # Houses %

(a)Ground water 39 37Ground water with bottled water/

vendedwater/public hyudrant41 39

Network water 10 9Network water with groundwater 2 2Public hydrant/vended water

with rainwater14 13

Total 106 100

Total households using atleast two sources

65 61

(b)DW 3 3Bottled water 12 11Ground water 70 64Vended water 34 31HU 7 6PAM 32 29Other: public toilet 4 4Public hydrant 8 7TA 13 12

Total 183 166

K. Bakker / Geoforum 38 (2007) 855–868 863

in the city are regulated by the national government’s Min-istry of Mines, which has no formal mandate to cooperatewith municipal water suppliers; this has meant that initialplans by the private concessionaires to prohibit users fromaccessing private wells have not been implemented. Raisingnetworked water prices may lead, in the absence of gover-nance reform and networked water quality improvementsto groundwater ‘crowding out’ networked water supply,and to a further reduction in revenues. Indeed, water man-agers for the two private companies have noted, with someconsternation, the growing number of ‘zero consumption’customers – some of whom may be switching to other watersources.

This, in turn, raises a more general point about waterpricing, pertaining to the limits of cross-subsidisationwithin a water-pricing regime in cities, such as Jakarta, witha large proportion of poor residents. With a ratio of domes-tic to industrial customers of 80/20 and with relatively fewusers in higher tariV bands, possibilities for cross-subsidiesin Jakarta are relatively limited. This suggests that in theabsence of subsidies external to the water supply pricingregime – as are used in Chile, and were used in OECDcountries such as the UK (Bakker, 2004; Gomez-Lobo,2001) – raising tariVs will not necessarily have the desiredgoal of increasing revenues and increasing rates of connec-tion of the poor. As discussed in the following section, addi-tional measures to reduce or remove disincentives for poorcustomers to choose network connections will also berequired.

5. Disconnecting or Misconnecting? Disincentives for poor customers to choose network connections

The previous section discussed the disincentives for thewater utility (whether publicly or privately managed) toconnect the poor. In this section, we discuss the disincen-tives which discourage poor consumers from connecting tothe water supply system. Our evidence is drawn from pri-mary survey data,14 and is supported by references to otherpublished survey data.

Understanding choices about water use must be situatedin context; in Jakarta, as in many mega-cities in the South,most residents use multiple sources of water in the home(Table 4). Residents of Jakarta obtain their water supplythrough a complex, heterogeneous set of sources, tech-niques, and modes of delivery. Few residents rely on onesource, using a combination of household piped networkwater connections, shallow and deep wells, public hydrants,

14 Survey of 110 households conducted in 2005 in eleven kelurahan (sub-districts) in North (Kamal Muara, Penjaringan (Marlina & GedongKompa), Penjaringan (Rawa Bebek), Pegangsaan Dua), West (Semanan),East (Kampung Melayu, Rawa Terate, Jati, Kampung Tengah) and Cen-tral Jakarta (Kebon Melati, Gunung Sahari Selatan). These kelurahanwere identiWed as predominantly poor and targeted for delivery of waterservices by the Kimpraswil Fuel Subsidy Reduction Compensation pro-gram, created to oVset the impacts of a reduction in fuel subsidies on poorhouseholds.

and water vendors for their water supply needs (Surjadi,2002, 2003). According to our survey of 110 households insix Jakarta neighbourhoods in 2005, 61% of householdssurveyed used multiple sources (the three most frequentcombinations being network and vended water, networkand groundwater, and groundwater and vended water).15

Use of diVerent water sources varies temporally and sea-sonally, due to quality and pressure concerns. Low pressurein the piped network means that households prefer to havea backup source – often a well. In some areas of the city,however, shallow groundwater cannot be used for drinkingdue to salinisation and pollution of the shallow aquifer(due to pumping, sea-level intrusion, and surface wastewa-ter disposal in the absence of a sewerage system).

This heterogeneity of use is further complicated byJakarta’s spatial pattern of urban development and urbanservices provision. Within the city, an ‘estate’ pattern ofblocks of commercial properties and colonial-era mansionsfronting on broad avenues is intermixed with dense ‘illegal’settlements of poorly serviced houses and self-built dwell-ings in the inner blocks (Cowherd, 2002; Ford, 1993; Leaf,1996; Porter, 1996), on empty lots, and along any streetswide enough to accommodate built structures while stillpermitting the passage of traYc, a pattern which has inten-siWed following the informalisation of much of the city’seconomy following currency devaluation in 1998. Manyneighbourhoods do not have any access to piped water, asthe water network is concentrated in wealthier areas of thecity (Martijn, 2005). The resulting spatial diVerentiation ofland use and income has created an ‘urban dualism’, withmiddle-class houses abutting informal housing in a highlyvariable urban micro-geography in which multiple watersources will be in use simultaneously.

Even in those areas with networked water supply, manyhomes will not have individual household connections.Susantono Wnds, in an extensive survey, that informal waterservices “thrive” in neighbourhoods where formal servicesare available, with households relying on water vendorseven when they have the option of house connections withthe municipal water utility (Susantono, 2001). Householdsrely on water vendors even when they have the possibilityof a direct house connection to the network. Physical prox-imity of the network (as indicated by the distribution of atertiary pipe network in the neighbourhood) is not, in otherwords, linearly correlated with residential network connec-tions.

Why would this be the case? The answer is that thechoice of which source to use is inXuenced by factors otherthan physical availability of a network. One important fac-tor is the total cost of water supply (as distinct from thecost per unit volume of water). In a pattern typical of thirdworld cities (Cairncross et al., 1990; Gulyani et al., 2005;Swyngedouw, 1997), piped water supply costs less per unit

15 These Wndings are similar to the results of surveys conducted by Surj-adi (1994, 2002, 2003) and McGranahan et al. (2001), the two most recentacademic studies available.

864 K. Bakker / Geoforum 38 (2007) 855–868

volume in Jakarta than other modes of water supply, par-ticularly vended water. In comparisons of the prices of ven-dor water versus networked water supply, the price per unitvolume was found to be from 10 times to 32 times moreexpensive in the case of vendor water.16 Poor householdstypically rely on vendor water, whereas wealthier house-holds have access to the networked water supply system; asa result, many poor households pay more per unit volumeof water than do wealthier residents of the city. Given theirlower incomes, many poor households pay a much higherproportion of their income for water than do wealthierhouseholds. In our survey of 110 households, 43% of house-holds spent more than 5% of their income on water bills(often cited as appropriate threshold by international aidorganizations).17 Wealthier households with a networkedconnection, in other words, receive water at a lower cost perunit volume, spending lower proportions of income formuch greater quantities of water. Unsurprisingly, levels ofwater consumption are positively correlated with wealth inJakarta (McGranahan et al., 2001).

On the basis of cost per unit volume alone, then, it seemscounter-intuitive that poor households would not connectto the water supply network where possible. However, thedisincentive for connection becomes more obvious when weconsider the total cost of connecting to the water supply sys-tem (as opposed to price per unit volume of water supply).Monthly bills include more than charges per unit volumesof water consumed. Fixed charges (such as the meter fee,and the annual charge) are also added on to the bill (Table3). For a poor household whose residents consume 50 L/per-son/day (the World Health Organization recommendedminimum), the Wxed charges will be anywhere from 5 to 10times as high as the volumetric consumption charge; theeVective cost per unit volume will thus be higher than thatof vended water for the poorest consumers. Moreover, anetworked water supply implies additional infrastructurecosts to be borne by the consumer, in the form of a watertank or holding device, made necessary because of the inter-mittent nature of water supply through the piped network(with cutoVs of several hours occurring daily in some areas).Transaction costs are also signiWcant; long waiting timesat water utility oYces to pay bills and clear up meter mis-readings raise transaction costs compared to the ease ofcomplaint handling and convenience of home visits by ven-dors to collect bill payments. Connection fees are also sig-niWcant (ranging from 200,000 to 350,000 Rupiah in thehouseholds surveyed), relative to average incomes of poorhouseholds (which averaged 1.4 million Rupiah in the

16 ADB (2003); McGranahan et al., (2001); and a survey conducted bythe author in the neighbourhood of Sunter Agung in January 2001. ADB,2003b gives a maximum Wgure of US $4.17/m3.17 A study of 1000 households in Jakarta which examined the diVerent

prices paid by diVerent wealth groups found that, overall, the poor pay onaverage twice as much per metre cubed as the wealthy (McGranahan et al.,2001), and that water expenditure represents, on average, 10% of incomein poor households.

households surveyed), and must usually be provided as alump sum; which may pose signiWcant barriers to householdswith small, irregular incomes. Connection fees also varydepending on distance from the network; poor householdsare more likely to live in areas of lower network density(Fig. 1), and thus to pay higher fees for connecting. For allof these reasons, overall costs to poor households of vendedwater may be lower than networked water supply, eventhough the latter has a lower price per unit volume. Giventhese cost barriers, payment Xexibility permitted by vendors(some of whom even allow customers to buy water on credit)is an important incentive for poor households, which mayhave limited budgeting ability, to choose vended water overnetworked water (Susantono, 2001; ShoWani, 2003).

Another important factor is land tenure. Deep wells areexpensive and have higher maintenance costs, which eVec-tively prohibits development by those without permanenttenure. A signiWcant proportion of the city’s populationlives in temporary (often self-built) accommodation withoutsecure tenure. In these instances, public hydrants andvended water become the sole or primary source of supply.Surjadi et al. found that over 20% of the city’s residents reg-ularly buy drinking water from vendors (Surjadi, 1994). Themost recent academic survey found that approximately 1/3of Jakarta’s households purchase water from street vendors(Crane and Daniere, 1996); these Wgures correspond withthe results of our household survey, which found that 31%of respondents regularly bought vended water (Table 4).

Another disincentive to connect to the network is its per-ceived low water quality. In our survey, networked waterwas perceived to be lower quality than other sources ofwater (particularly groundwater), particularly by more edu-cated respondents. Residents of Jakarta perceive ground-water to be of higher quality than either vended or networkwater. Indeed, the most comprehensive comparative surveyof water quality of diVerent sources in poor neighbour-hoods in Jakarta to date found that samples of drinkingwater from the network were more contaminated with fecalcoliform than groundwater (Surjadi, 1994). In some cases,vended water was perceived to be of higher quality thannetworked water supply. The fact that vendors check waterquality and may strain the water or let it settle before deliv-ering explains why perceptions of vended water qualitymay be higher, despite the fact that vendor water oftenoriginates in hydrants connected to the networked watersupply system.

Recognizing some of these barriers to connecting thepoor, both private concessionaires have undertaken limitedinitiatives to improve access for poorer households. To ren-der in-house connections more aVordable, Palyja intro-duced a policy allowing poorer households (on the lowesttariV bands) to pay the connection fee in 12 monthly instal-ments included in the monthly water bill.18 Partly as a result,

18 The pro-rated monthly connection fee of $0.71 included in the monthlywater bill. A household consuming 20 m3 a month will thus have a month-ly bill of about $1.50 ($0.0375 £ 20 + $0.71).

K. Bakker / Geoforum 38 (2007) 855–868 865

in West Jakarta, the number of poor people servedincreased from 72,816 in February 1998 to 177,164 inDecember 2000 (ADB, 2003a), but the monthly bills remainat a level above what many households can aVord. In theeastern concession area, the community of Marunda wastargeted by TPJ, which used a grant from its British parentcompany to subsidize the provision of in-house connections.Fewer than 500 households were connected; to facilitatepayment, connection fees were waived and households werewere instead required to pay a deposit (of approximatelyUS $2.50) (ADB, 2003a). Levels of water consumption havereportedly increased dramatically, while water bills havefallen substantially (ADB, 2003a). Prior to the concessioncontract, households in Marunda District generally receivedtheir water from private vendors who purchased water fromtankers. Households used to spend, on average, US$7.50 amonth for 3 m3 of water (Wve 20-l containers of water at$0.05 a container) now pay approximately $1.125 for 30 m3

of water (at $0.0375 per m3–most customers being on a lowtariV, reXecting the small size of their dwellings), consuming10 times as much water but paying approximately one-sev-enth of their previous monthly bills, partly the reason forhigh levels of cost recovery from the newly connectedhouseholds (BPD, 2003). Recognizing the limited penetra-tion of water supply network into poor neighbourhoods, thefederal government launched a water supply program insome of the poorest kampungs in 2004,19 but the provisionof household connections was severely limited by the disin-centives discussed above, compounded by an unwillingnessof private partners to extend the network in conjunctionwith the government, and by suspicion on the part of somepublic sector managers that publicly-provided infrastruc-ture would end up providing implicit subsidies to the privatesector (ShoWani, 2005).

Given, however, the high level of indebtedness of themunicipal water utility to the private concessionaires, littleinterest has been shown on the part of the private compa-nies in extending what are essentially charitable, loss-mak-ing initiatives. Accordingly, these ‘pro-poor’ initiatives haveremained limited in scope, and have not been duplicatedelsewhere in the city. Without an explicit ‘pro-poor’ policyon the part of the Government, and in the absence of spe-ciWc pro-poor targets in the contract, new connections inpoorer areas are likely to lag in proportion to the overallincrease in new connections for the reasons discussedabove. Recognizing this, donors have begun re-fundingcommunity water supply in Jakarta. The (American bilat-eral aid donor) USAID, through its Environmental Ser-vices Program is funding decentralized small-scalecommunity water supply systems in West Java, includingJakarta, with a budget of $40 million USD over Wve years;these community systems will not connect users to the net-work, but will rely on alternative technologies. The World

19 Under the auspices of the Kimpraswil Fuel Subsidy Reduction Com-pensation program, created to oVset the impacts of a reduction in fuel sub-sidies on poor households.

Bank has approved a $5 million US ‘output based aid’ con-cessional loan for expanding network coverage in Jakarta.20

With an explicitly ‘pro-poor’ focus, this latter project pro-vides cheap capital to the two concessionaires to connectthe poor.

This, in turn, raises questions about the long-term abilityof PSP contracts to supply water to the poor. Similar ques-tions were raised in the World Panel on Financing WaterInfrastructure report21 released at the Third World WaterForum in Kyoto in 2003. The report articulated the needfor a new Wnancial architecture to stimulate and supportXows of private capital for water and sanitation (Win-penny, 2003) including, controversial calls to use oYcialaid funding to support private sector involvement throughthe provision of low-cost Wnance and risk mitigation mech-anisms such as currency guarantees for private investors indeveloping countries.22 Implementing the pro-poor approachin this manner would entail a potentially dramatic transfor-mation in the premises and mechanisms of ODA Wnance, inwhich public funds are provided to subsidize poorer house-holds, enabling private sector operators to manage watersupply systems at a proWt. Ironically, one of the key prom-ises by advocates of PSP contracts has been the indepen-dent Wnancing that private companies could provide undersome types of PSP agreements. In contrast to these prom-ises, some private water companies have more recentlyargued that they must have access to public funds, on con-cessionary terms (from governments, bilateral aid agencies,or multilateral developments banks) if they are to meet uni-versal service coverage targets.

6. Conclusions

The private sector participation contracts signed in 1998with two international operators promised to improvewater quality, mobilize international Wnance for networkexpansion, and thereby improve and increase access towater supply for Jakarta residents – particularly the poor.As documented in this paper, a survey of performance ofthe private sector concessionaires indicates that key origi-

20 At the time of writing (October 2005), the USAID project was under-way and the World Bank project was in the tendering stage.21 Commissioned by the organizers of the Third World Water Forum,

and chaired by former IMF General Manager Michel Camdessus, thePanel brought together the Presidents of major multilateral developmentbanks (IADB, ADB, EBRD, WB), and representatives of the IFC, Citi-bank, US Ex-Im Bank, private water companies (Suez, Thames Water),government representatives (from Mexico, Ivory Coast, Pakistan, Egypt,and France) and two NGOs (Transparency International and WaterAid).22 These proposals, as well as the composition of the Panel and the lack

of public consultation on the report have been critiqued by a number oforganizations, which have raised numerous points: the focus on large-scaleinfrastructure and lack of emphasis on alternative technologies, levels ofservice, governance models, citizen input, and methods of improving pub-lic sector performance; the focus on encouraging private sector involve-ment to the exclusion of other business models; and the ethics andfeasibility of providing risk mitigation and cost reduction to the privatesector via the use of public funds (Bakker, 2003b).

866 K. Bakker / Geoforum 38 (2007) 855–868

nal performance targets have been dramatically scaledback. Moreover, there is evidence that new connectionshave targeted middle-class customers, and that tariVincreases have been higher for poorer customers, withoutconcurrent attempts to address issues of ability to pay,income thresholds, and cross-subsidy mechanisms. TariVpricing (with lower tariV bands below marginal costs),decided by the municipal government in negotiation withconcessionaires, is implicitly ‘anti-poor’, providing a disin-centive to both the municipality and the private concession-aires to connect the poor. The physical layout of thenetwork, which is spatially concentrated in wealthier areasof the city – a legacy of public sector management – is anadditional barrier to connecting the poor.

Moreover, poor users have multiple disincentives to con-nect to the network. Total costs of networked water supplymay be higher than alternative sources (such as groundwa-ter or vended water). Other disincentives include insecuretenure, the need for Xexibility of payment, convenience, sta-tus, and high ‘transaction costs’ associated with dealingwith the formal water utilities. ‘Transaction costs’ (infra-structure costs to build storage because networked watersupply is only intermittent; line-ups and time oV work topay bills (for those without bank accounts and regularincome); fear of time required to deal with meter mis-read-ing and bill over-charging) are other disincentives.

These Wndings echo results of other studies regardingurban water supply in the South (e.g. Almansi et al., 2003;WaterAid, 2003; Whittington, 1992). Much of this litera-ture is, however, characterized by a narrow economicreductionism, in which the failure of water utilities to reachthe poor is attributed largely or solely to an inability-to-pay, or to inappropriate pricing. In this literature, theconcept of poverty is used (to use Rahnema’s term) as a‘regulating Wction’ (Rahnema, 1992), obscuring the non-economic factors that act as important disincentives forpoor households, which choose not to connect to the watersupply system. In a manner analogous to the colonialstate’s active construction of the peasant (Mitchell, 2002),the concept of the poor functions as a ‘post-colonial device’(Bell, 2002) that has become a preoccupation and focus foraction by Northern aid and development organizations.This article has suggested that other questions shouldreceive greater attention in the pro-poor debate, namely:the nature of urban governance (which, in Jakarta’s case,has systematically prioritized monumental infrastructureand elite residential services at the expense of universalpublic services); the inequitable spatialisation of networkaccess (a legacy of public sector management to servelargely elite interests); and the multiple disincentives for thepoor to choose network connections, and for network man-agers – both public and private – to connect the poor.

Acknowledgements

Numerous individuals (water company employees, con-sumers, government oYcials, NGO representatives) gave

generously of their time during Weldwork research inJakarta in 2001, 2004, and 2005. Fieldwork research sup-port by Michelle Kooy, Nur Endah ShoWani and Ernst-JanMartijn is gratefully acknowledged. Helpful commentswere received from Michelle Kooy, Graciela Madanes-Schneier, Philippe Le Billon, and Andrew Leyshon as wellas three anonymous reviewers. Eric Leinberger’s expertassistance with graphics is, as always, much appreciated.

References

ADB, 2003a. Beyond Boundaries: extending services to the urban poor.Asian Development Bank, Manila.

ADB, 2003b. Water in Asian Cities: Summary of Findings of the Studyand a Regional Consultation Workshop. Asian Development Bank,Manila.

Agtini, M., Rooswanti, S., Lesmana, M., Punjabi, N., Simanjuntak, C.,Wangsasaputra, F., Nurdin, D., Pulungsih, P., RoWq, A., Santoso, H.,Pujarwoto, H., Sjahrurachman, A., Sudarmono, P., von Seidlein, L.,Deen, J., Ali, M., Lee, H., Kim, D., Han, O., Park, J., Suwandono, A.,Oyofo, B., Campbell, J., Beecham, H., Corwin, A., Clemens, J., 2005.The burden of diarrhoea, shigellosis, and cholera in North Jakarta,Indonesia: Wndings from 24 months surveillance infectious disease 5(89).

Akhtar, S., 2005. Mobilizing Finance for Infrastructure Development.Speech (Director General, Southeast Asia Department, Asian Devel-opment Bank), Indonesia Infrastructure Summit. 17 January, Jakarta.<http://www.adb.org/Documents/Speeches/2005/sp2005004.asp>(accessed 14:10:2005).

Almansi, F. et al., 2003. Everyday Water Struggles in Buenos Aires. WaterAidand Tearfund, London.

Argo, T., 1999. Thirsty downstream: the provision of clean water inJakarta, Indonesia. Unpublished doctoral thesis, School of Regionaland Community Planning, University of British Columbia.

Argo, T., Firman, T., 2001. To privatize or not to privatize? Reform ofurban water supply services in Jabotabek, Indonesia. Built Environ-ment 27 (2), 146–155.

Azdan, M., 2001. Water Policy Reform in Jakarta, Indonesia: A CGEanalysis. Unpublished doctoral thesis. Department of Agricultural,Environmental and Development Economics: Ohio State University.

Bakker, K., 2003a. From archipelago to network: urbanisation and privat-isation in cities in the South. The Geographical Journal 169 (4), 328–341.

Bakker, K., 2003b. The third world water forum in Kyoto. Flux 52/53,111–116.

Bakker, K., 2004. An uncooperative commodity: privatizing water inEngland and Wales.

Bakker, K., 2005. Neoliberalizing Nature? Market environmentalism inwater supply in England and Wales. Annals of the Association ofAmerican Geographers 95 (3), 542–565.

Bardhan, P., 1997. Corruption and development: a review of the issues.Journal of Economic Literature 35 (3), 1320–1346.

Barlow, M., Clarke, T., 2002. Blue Gold: The Fight to Stop the CorporateTheft of the World’s Water. The New Press, New York.

Baye, E., 1997. Experimenting with the privatization of water services inIndonesia – from Surabaya to Djakarta. In: Lorrain, D. (Ed.), UrbanWaste Management. Hydrocom, Levallois, pp. 195–211.

Bell, M., 2002. Inquiring minds and postcolonial devices: examining pov-erty at a distance. Annals of the Association of American Geographers9 (3), 503–527.

Berry, B., 1982. Clean water for all: equity-based urban water supply alter-natives for Indonesia’s cities. Urban Geography 3 (4), 281–299.

BPD, 2003. Jakarta, Indonesia: Water Supply Improvements to MarundaDistrict. London: Building Partnerships for Development in Water andSanitation. http://www.bpd-waterandsanitation.org/english/projects/jakarta.asp (accessed 3:01:2004).

K. Bakker / Geoforum 38 (2007) 855–868 867

Braadbaart, O., Braadbaart, F., 1997. Policing the urban pumping race:industrial groundwater overexploitation in Jakarta. World Develop-ment 25 (2), 199–210.

Brennan, E.M., Richardson, H.W., 1989. Asian megacity characteristics,problems and policies. International Regional Science Review 12 (2),117–129.

Cairncross, S., Hardoy, J., Satterthwaite, D. (Eds.), 1990. The Poor DieYoung. Earthscan, London.

Chifos, C., 2000. Meeting unmet expectations: initiatives in urban environ-mental service provision in Jakarta. In: Chifos, C., Yabes, R. (Eds.),Southeast Asian Urban Environments. Program for SEA Studies.Temple University Press, Arizona State University, pp. 29–72.

Chifos, C., Hendropranoto, S., 2000. Thirty Years of Urban InfrastructureDevelopment in Indonesia. In: Chifos, C., Yabes, R. (Eds.), SoutheastAsian Urban Environments. Program for SEA Studies. Temple Uni-versity Press, Arizona State University, pp. 153–181.

Cowherd, R., 2002. Cultural construction of Jakarta: Design, planning anddevelopment in Jabotabek, 1980–1997. Unpublished Ph.D. disserta-tion, MIT, Department of Architecture.

Crane, R., 1994. Water markets, market reform and the urban poor:results from Jakarta, Indonesia. World Development 22 (1), 71–83.

Crane, R., Daniere, A., 1997. The contribution of environmental amenitiesto low-income housing: a comparative study of Bangkok and Jakarta.Urban Studies 34 (9), 1495–1512.

Crane, R., Daniere, A., 1996. Measuring access to basic services in global cit-ies: descriptive and behavioural approaches. American Planning Associ-ation. Journal of the American Planning Association 62 (2), 203–221.

Cross, P., Morel, A., 2005. Pro-poor strategies for urban water supply andsanitation services delivery in Africa. Water Science and Technology51 (8), 51–57.

DKI Jakarta, 2005. Decree 138/2005. Daerah Khusus Ibu Kota JakartaRaya, Jakarta.

Firman, T., 1997. Land conversion and urban development in the northernregion of West Java, Indonesia. Urban Studies 34, 1027–1046.

Firman, T., 1998. The restructuring of Jakarta Metropolitan Area: a “glo-bal city” in Asia. Cities 15 (4), 229–243.

Firman, T., 2000. Rural to urban land conversion in Indonesia duringboom and bust periods. Land Use Policy 17, 13–20.

Firman, T., Dharmapatni, I., 1994. The challenges to sustainable develop-ment in Jakarta metropolitan region. Habitat International 18 (3), 79–94.

Ford, L., 1993. A model of Indonesian city structure. Geographical Review83 (4), 374–396.

Franceys, R., Weitz, A., 2003. Public-private community partnerships ininfrastructure for the poor. Journal of International Development 15(8), 1083–1098.

Gilbert, A., James, D., 1994. Water pollution in Jakarta Bay. In: James, D.(Ed.), The Application of Economic Techniques in EnvironmentalImpact Assessment. Kluwer, Dordrecht, pp. 111–142.

Goldman, M., 2005. Imperial Nature: The World Bank and Struggles forSocial Justice in the Age of Globalization. Yale University Press, NewHaven, CT.

Gomez-Lobo, A., 2001. Incentive based subsidies: designing output-basedsubsidies for water consumption Public Policy for the Private Sector22 World Bank, Washington.

Gulyani, S. et al., 2005. Universal (non)service? Water markets, householddemand and the poor in urban Kenya. Urban Studies 42 (8), 1247–1274.

Harsono, A., 2005. When water and political power intersect: a journalistprobes the story of water privatization in Jakarta, Indonesia. NiemanReports 59 (1), 45–47.

Haughton, G., 2002. Market making: internationalization and globalwater markets. Environment and Planning A 34 (5), 791–807.

Hukka, J.J., Katko, T., 2003. Refuting the paradigm of water services priv-atization. Natural Resource Forum 27, 142–155.

Indonesia Times, 1996. Water Shortage Maybe Come Out of Control inyear 2000. Indonesia Times. 19 August.

Jakarta Post, 2003. Water An Economic Commodity: Minister. TheJakarta Post. 20 September.

Jakarta Water Supply Regulatory Body, 2004. Regulatory Approach tothe Jakarta Water Supply Concession Contracts. Presentation: Lanti,A., (Member, Regulatory Body), Stockholm Water Symposium;August 16–20.

Jakarta Water Supply Regulatory Body, 2005. Urban Water in Jakarta.Presentation: Urban Water and Citizenship workshop, University ofIndonesia, 11 May.

Johnstone, N., Wood, L., 2001. Private Firms and Public Water: RealisingSocial and Environmental Objectives in Developing Countries.Edward Elgar, Cheltenham.

Kusno, A., 1997. Modern beacon and the traditional polity: Jakarta inthe time of Sukarno. Journal of Southeast Asian Architecture 1, 25–45.

Kusno, A., 1998. Custodians of transnationality: metropolitan Jakarta,middle-class prestige, and the Chinese. In: Dandekar, H. (Ed.), CitySpace Globalization: An International Perspective, pp. 161–170.

Kusno, A., 2000. Behind the Postcolonial: Architecture, Urban Space andPolitical Cultures in Indonesia. Routledge, London.

Leaf, M., 1996. Building the road for BMW: culture, vision and theextended metropolitan area of Jakarta. Environment and Planning A28 (9), 1617–1635.

Leitmann, J., 1995. Urban environmental proWle: a global synthesis ofseven urban environmental proWles. Cities 12 (1), 23–39.

Lo, F.C., Yeung, Y.M., 1996. Global restructuring and emerging urbancorridors in PaciWc Asia. In: Yeung, Y.M., Lo, F.C. (Eds.), EmergingWorld Cities in PaciWc Asia. United Nations University Press, Tokyo.

Lovei, L., Whittington, D., 1993. Rent-extracting behaviour by multipleagents in the provision of municipal water supply: a study of Jakarta,Indonesia. Water Resources Research 29 (7), 1965–1975.

Mandaung, M.E., 2001. Improving the eYciency of water supply manage-ment: a case study of the Jakarta public water supply enterprise, Indo-nesia. Canberra: Centre for Developing Cities, University of Canberra.cities.Canberra.edu.as/publications/Policypaper/Melva.htm (accessed12 January 2001 and 24 June 2002).

Martijn, E.J., 2005. Hydraulic Histories of Jakarta, Indonesia (1949–1997):Development of piped drinking water provision. Vancouver: Unpub-lished research report, Department of Geography, University of BritishColumbia.

McDonald, D., Ruiters, G., 2005. The Age of Commodity: PrivatizingWater in Southern Africa. Earthscan, London.

McGranahan, G. et al., 2001. The Citizens at Risk: From Urban Sanita-tion to Sustainable Cities. Earthscan, London.

McIntosh, A., 2003. Asian Water Supplies: Reaching the Urban Poor.Manila: Asian Development Bank and International Water Associa-tion. <http://www.adb.org/Documents/Books/Asian_Water_Supplies/(accessed 13:10:2005).

Mitchell, T., 2002. Rule of Experts: Egypt, Techno-politics, Modernity.University of California Press, Berkeley.

Nickson, A., Franceys, R., 2003. Tapping the Market: The Challenge ofInstitutional Reform in the Urban Water Sector. Palgrave MacMillan,London.

Nunan, F., Satterthwaite, D., 2001. The inXuence of governance on theprovision of urban environmental infrastructure and services for low-income groups. International Planning Studies 6 (4), 409–426.

PAM Jaya, 2004. Laporan Evaluasi Kinerja Tahunan. Annual ReportEvaluation 2003. Perusahaan Daerah Air Mimun, Jakarta.

Pierre, J., 1995. The marketization of the state: citizens, customers and theemergence of the public market. In: Peters, G., Savoie, D. (Eds.), Gov-ernance in a Changing Environment. The McGill-Queen’s UniversityPress, Montreal and Kingston.

Pierre, J. (Ed.), 2000. Debating Governance. Oxford University Press, NewYork.

Platts’ Global Water Report, 2002. Jakarta: test bed for public privatepartnership. Issue 146, 17 May, pp. 1–3.

Porter, R., 1996. The Economics of Water and Waste: A Case Study ofJakarta, Indonesia. Avebury, Aldershot.

Rahnema, M., 1992. Poverty. In: Sachs, W. (Ed.), The Development Dic-tionary. Zed Books, London, pp. 158–176.

868 K. Bakker / Geoforum 38 (2007) 855–868

Robbins, P., 2003. Transnational corporations and the discourse of waterprivatization. Journal of International Development 15 (8), 1073–1082.

Robertson-Snape, F., 1999. Corruption, collusion and nepotism in Indone-sia. Third World Quarterly 20 (3), 58–602.

Rogers, P., Hall, A., 2003. EVective water governance.Global Water Partner-ship Technical Committee TEC Background Papers, #7 GWP, Sweden.

Scott, J., 1969. The analysis of corruption in developing nations. Compara-tive Studies in Society and History 11 (3), 315–341.

Shirley, M., 2002. Thirsting for EYciency: The Economics and Politics ofUrban Water System Reform. Elsevier, London.

Shiva, V., 2002. Water Wars: Privatization, Pollution and ProWt. SouthEnd Press, Boston.

Shleifer, A., Vishny, R., 1993. Corruption. The Quarterly Journal of Eco-nomics (August), 599–617.

ShoWani, N.E., 2003. Reconstruction of Indonesia’s drinking water utilities:assessment and stakeholders perspective of private sector participation inthe capital province of Jakarta, Indonesia. TRITA-LWR M.Sc. ThesisLWR-EX-03-30, ISSN 1651-064X. Stockholm: Department of Land andWater Resources Management, Royal Institute of Technology.

ShoWani, N.E., 2005. Assessment of the Kimpraswil water supply project:fuel subsidy reduction compensation program for water supply. Van-couver: Unpublished report, Department of Geography, University ofBritish Columbia.

Silas, J., 1992. Government-community partnerships in Kampungimprovement programmes in Surabaya. Environment and Urbaniza-tion 4 (2), 35–36.

Simanjuntak, C.H., Larasati, W., Arjoso, S., Putri, M., Lesmana, M.,Oyofo, B.A., Sukri, N., Nurdin, D., Kusumaningrum, R.P., Punjabi,N.H., Subekti, D., Djelantik, S., Sukarma, Sriwati, Muzahar, Lubis, A.,Siregar, H., Mas’ud, B., Abdi, M., Sumardiati, A., Wibisana, S., Hen-darwanto, Setiawan, B., Santoso, W., Putra, E., Sarumpaet, S., Ma’ani,H., Lebron, C., Soeparmanto, S.A., Campbell, J.R., Corwin, A.L., 2001.Cholera in Indonesia in 1993–1999. American Journal of TropicalMedicine and Hygiene 65 (6), 788–797.

Smith, L., 2002. The murky waters of the second wave of neoliberalism:corporatization as a service delivery model in Cape Town. Geoforum35 (4), 375–393.

Soto Montes de Oca, G., et al., 2003. Assessing the willingness to pay formaintained and improved water supplies in Mexico City. CSERGEWorking Paper ECM-2003-11.

Surjadi, C., 1994. Households Environmental Problems in Jakarta. Stock-holm Environment Institute, Stockholm.

Surjadi, C., 2002. Public Private Partnerships and the Poor: DrinkingWater Concessions Case study – Jakarta, Indonesia. InternationalInstitute for Environment and Development, London.

Surjadi, C., 2003. Public Private Partnerships and the Poor: DrinkingWater Concessions, A Study for Better Understanding Public-privatePartnerships and Water Provision in Low-income Settlements. Water,Engineering and Development Centre, Loughborough.

Susantono, B., 2001. Informal Water Services in Metropolitan Cities ofDeveloping World: The Case of Jakarta, Indonesia. Berkeley: Doctoralthesis, Department of City and Regional Planning, University of Cali-fornia.

Swyngedouw, E., 1997. Power, nature, and the city. The conquest of waterand the political ecology of urbanization in Guayaquil, Ecuador: 1880–1990. Environment and Planning A 29 (2), 311–332.

Swyngedouw, E., 2005. Dispossessing H2O: the contested terrain of waterprivatization. Capitalism Nature Socialism 16 (1), 81–98.

Taylor, John Lewis., 1983. An evaluation of selected impacts of Jakarta’sKampung improvement program. Los Angeles: Unpublished Doctoralthesis; Urban Planning, UCLA.

Taylor, J., 1983. An Evaluation of Selected Impacts of Jakarta’s KampungImprovement Program. Unpublished phd thesis, Department of UrbanPlanning, UCLA.

The Economist, 2004. Third world water and the private sector – How notto help those in need. The Economist, 28 August, 11.

Trawick, P., 2003. Against the privatization of water: an indigenous modelfor improving existing laws and successfully governing the commons.World Development 31 (6), 977–996.

WaterAid, 2003. Does PSP BeneWt the Poor: The Synthesis Report. Water-Aid and Tearfund, London.

Whittington, D., 1992. Possible adverse eVects of increasing block watertariVs in developing countries. Economic Development and CulturalChange 41 (1), 75–87.

WHO, 2000. Global Water Supply and Sanitation Assessment. 2000Report. World Health Organization, UNICEF, and Water Supply &Sanitation Collaborative Council, Geneva.

Winpenny, J., 1994. Managing Water As An Economic Resource. Routl-edge, London.

Winpenny, J., 2003. Financing water for all: report of the world panel onWnancing water infrastructure. World Water Council, Third WorldWater Forum, Global Water Partnership.

World Bank, 1993. Indonesia: Urban Public Infrastructure Services.World Bank, East Asia Region, Washington.

World Bank, 1994. Infrastructure for Development World DevelopmentReport 1994 World Bank, Washington.

World Bank, 1997. The State in a Changing World World DevelopmentReport 1997 World Bank Group, Washington.

World Bank, 1999. Water Sector Adjustment Loan: Sectoral Environmen-tal Assessment. E-267. World Bank and Jakarta: Government of Indo-nesia, Inter-Agency Task Force on Water Sector Policy Reform,Washington.

World Bank, 2004a. Making Services Work for Poor People WorldDevelopment Report 2004 World Bank Group, Washington.

World Bank, 2004b. Indonesia: averting an infrastructure crisis: a frame-work for policy and action. Indonesia Infrastructure Department, EastAsia and PaciWc Region, World Bank.

World Bank, 2005. Implementation Completion Report (SCL-44690) on aloan in the amount of US $300 Million to the Republic of Indonesiafor the Water Resources Sector Adjustment. Washington and Jakarta:Rural Development and Natural Resources Sector Unit; IndonesiaCountry OYce; East Asia and PaciWc Region, World Bank. Report No32 6565. June.

Yepes, G., 1999. Do cross subsidies help the poor to beneWt from waterand wastewater services? Lessons from Guayaquil. UNDP – WBWater and Sanitation Program Working Paper.