The Anatomy of a Deal

A Close Look at the World Bank's Plans to
Privatize Ghana's Water System

Multinational Monitor, September 2002


Having successfully pushed developing countries to sell off most of their government-owned enterprises, the World Bank is now urging poor countries to privatize basic service provision.

Although water delivery is assumed as a public responsibility in much of the industrialized world-80 percent of the water works in the United States are publicly owned and operated-the Bank is aggressively pushing privatization of water systems in the Third World.

Citizen movements across the planet are rising to challenge the World Bank and corporate schemes to wrest control of now-public water systems. One the hottest flashpoints in the conflict between the people and the Water Barons is in Ghana.

There, the National Coalition Against the Privatization of Water (NCAP of Water) is aggressively opposing a Bank-advocated privatization scheme that would lease out the country's urban water systems for a song. The scheme was hatched in 1995, and may be implemented next year, unless NCAP of Water can thwart it.

An International Fact-Finding Mission recently visited Ghana to assess the privatization proposal. It was able to interview all of the players in the privatization debate, including key officials from the World Bank and within the Ghanaian government, as well as to access many of the most important consultant reports and documents surrounding the privatization proposal. The fact-finding mission's investigation offers a unique glimpse at the legal framework and contract issues surrounding the effort to lease out the country's water system - the anatomy of a deal.


All parties to the conflict in Ghana agree the present water-delivery system is abysmal. One-third of urban consumers are not connected to the piped water system. Many of those who are connected to the pipe system receive sporadic service-as infrequent as once every two, three or four weeks.

In response to the failures of the state-run system, Ghana decided to pursue water privatization in 1995. The decision came at a meeting at which a consultant presented a range of options for reforming the water service. Virtually all options involved some version of privatization; the notion of reforming the utility while keeping it in the public sector was effectively ruled out from the beginning. The meeting was not known to the public. Nongovernmental organizations (NGOs) were excluded, with the exception of two international NGOs.

Since 1995, the privatization proposal has gone through various iterations. Parliament has been consulted, but has not yet voted on the proposal.

The World Bank and various foreign donors have funded an array of consultant reports to suggest how the privatization process should be carried out.

Although final details remain to be hammered out, the framework of the proposal now appears set.

The government calls Ghana's current privatization proposal "private sector participation" (PSP). It is a 10-year lease arrangement rather than a full sell-off of the country's water system. Under the PSP, the government will retain control of two aspects of the water system that have no prospect of paying for themselves-the country's extremely rudimentary sewage system and the rural water supply. The urban water supply systems will be grouped together into two geographically defined business units, with roughly half of the country's urban consumers in each unit.

A group of prequalified bidders-all foreign, and including global giants such as Suez and Vivendi-will bid to operate the business units. The Ghanaian government will pick two, one for each business unit.

The lease operators' duties will be to provide water supply, rehabilitate and renew pipes, maintain assets, bill and collect, and make new connections. The operators are each expected to invest $70 million over the 10-year life of the lease.

Responsibility for expansion of the system remains with the government. Pipe extension is expensive, and there is no realistic scenario in which tariff revenues might cover the cost of expansion. There is thus a need for external financing, and the government anticipates relying on donor funds -in many cases, available only to public recipients-to pay for the costs of pipe expansion.

In preparation for privatization, the government moved to adopt a policy of achieving cost recovery, in which water tariffs are priced sufficient to meet operations and maintenance costs, without any public subsidy to keep prices in check. This, even though systems in the United States, among other industrialized countries, routinely rely on support from general tax revenues. While some government officials and the World Bank suggest that cost recovery is a valuable policy objective in its own right, the move to achieve cost recovery has proceeded in tandem with evolution of the PSP proposal.

Cost recovery is viewed as a prerequisite to attracting private operators. According a 1998 report from Louis Berger, S. A., a consulting group which analyzed the Ghanaian water system, "In the context of leasing contracts under PSP, a tariff level that covers at least operating and maintenance costs as well as the financial costs related to any capital expenditures made by the private sector is necessary to attract private operators since it represents the break-even point for these contracts." In particular, private operators do not want to be dependent on government subsidies for revenues to cover their fees and profits.

Compounding the rate hikes, the privatization scheme calls for the inclusion of an "automatic tariff adjustment"- with rates rising automatically to offset inflation and, most importantly, currency devaluations. That makes sense from the viewpoint of the foreign operators-they will want to maintain constant profits in dollar-denominated terms, not in cedis, the local currency. But Ghanaian consumers' cedi income does not go up just because the value of the cedi declines. Assuming future devaluations, Ghanaian consumers will find themselves paying a higher and higher proportion of their income to the water company.

The government will pay the lessees a fee for system operation. This fee will include costs incurred by the lessee, including a rate of return on investment.

The fee will be adjusted to reflect the degree to which the lessee meets or exceeds reductions in "non-revenue water" -water that is either lost to physical leaks, or delivered to consumers but not paid for.

The fee will be further adjusted to reflect the degree to which the lessee meets performance targets related to: residual chlorine (20 percent), non-revenue water (10 percent), availability (30 percent), metering rate (10 percent) and customer complaints (30 percent).

The lessee will post a performance bond, which will presumably be forfeitable upon mis-, mal- or non-feasance by the lessee.


Through reimbursement of indexed operating and maintenance expenses, the lessee will be reimbursed for the basic costs of running the water system. A 2001 Information Memorandum for donors to the project-prepared by consultants Stone & Webster and the most detailed public description of the contract framework-suggests there will be reimbursement for "prudently incurred" costs. This reimbursement provision provides an incentive to the lessees to pad costs, to engage in transfer pricing (buying products from the home country or other subsidiaries or related companies at inflated prices) and generally to be insufficiently disciplined in economizing in expenses accrued.

This danger does not mean the reimbursement provision is improper; but the danger must be acknowledged and addressed. It is unclear what, if any, safeguards will be built into the contract to prevent such actions. It is unclear what agency would have the capacity to review expenditures by the lessees. The duty presumably will rest first with the Ghana Water Company (GWC), and perhaps also with the Public Utilities Regulatory Commission (PURC), in case of request for tariff increase. At present, the GWC does have considerable knowledge and expertise on the costs of operations and management. However, much or most of that expertise will be transferred to the lessees. After the lease, the GWC will run a skeleton operation of 80, which among other responsibilities, will be obligated to oversee the expansion of the piped system. It is very difficult to see how the GWC will maintain capacity to scrutinize the expenditures of the lessees.

On the flip side, the required investment by the lessees is relatively riskless.

Although a condition of entering into the lease arrangement is that the lessees must each provide $70 million in investment capital over 10 years, this is not the sort of investment made by an entrepreneur. Indeed, it may be viewed as more of a loan to the business unit. Under the terms of the contract, each lessee will be paid a specified return on investment (with the exact level to be determined by bid), as well as compensated for the depreciated value of their investment.


The contract provides an incentive to the lessees to reduce non-revenue water. Non-revenue water is an umbrella term that refers both to water that is physically lost through leakage and to water that is not paid for, due to unauthorized connections, nonpayment of bills or other reasons.

Overall system loss of non-revenue water is unknown, but generally estimated on the order of 50-55 percent of overall production. The Berger Report estimates approximately 30 percent losses due to nonpayment, and 25 percent to physical leakage.

From an accounting point of view, nonrevenue water due to nonpayment and nonrevenue water due to physical leaks are equivalent. However, they are not equally important in terms of service delivery; nonpayment water at least reaches a consumer, leaked water is simply wasted, with no consumer benefits.

It is likely that the "low-hanging fruit" in reducing nonrevenue water will relate to nonpayment rather than leaked water. If this is correct, then-at least to the extent there are built-in incentives to reduce expenses-the lessee incentive will be to reduce nonpayment water.

There are some apparently easy means to reduce nonrevenue water. The Berger Report estimates that half of all nonpayment water is consumed by government agencies that are either not billed, or simply do not pay their bills. These are large water users, and a lessee with authority to cut off agencies for nonpayment is likely to be able to quickly reduce this nonpayment.

Forcing government agencies to pay for water may reduce wastage and improve the water system's balance sheet. Indeed, this could easily be accomplished by the GWC, if there was political will to do so. However, payment from non-paying government agencies simply constitutes an intragovernmental reallocation of resources, leaving the government no better or worse off than it was initially.


Each lessee's fee will be adjusted, beginning in year four, based on the extent to which it satisfies the performance criteria. Performance will be measured in comparison to the early years of the contract, since baseline data do not now exist.

This may give the lessees an incentive to underperform in the early years of the contract, to lower the baseline.

Another problem relates to the establishment of baseline data, and monitoring of performance indicators during the term of the contract. The lessees will have responsibility for determining baseline data and for reporting performance data during the term of the contract. The lessees will also have responsibilities to report information to the PURC, in matters covered by PURC regulations.

The global experience with self-reporting is quite poor, and there is reason to be skeptical of PURC's and GWC's capacity to scrutinize the self-reported data.

It is easy to imagine manipulation of almost all of the data relevant to the performance standards. In particular, data on water availability and customer complaints- making up more than half of the weighting for the performance multiplier-would be very susceptible to manipulation.

Improper reporting could completely undermine the beneficial intent of the performance criteria in the payment formula, enabling the lessees to be paid for targets not achieved.


The lease arrangement specifies the lessees will take out a performance bond, which would be forfeited in event of lessee failure. Inclusion of the performance bond in the lease is an important mechanism of ensuring operator accountability.

However, the import of this mechanism of accountability is dependent on the terms of the bond. Important questions include:

* Is the bond partially forfeitable, or only fully forfeitable?

* What are the standards for forfeiture?

* Who makes the determination of forfeiture? According to what rules?

This information is not in the public domain. It is not clear to what extent these details have been considered by policymakers in Ghana. Asked, neither government officials, including the Water Sector Restructuring Secretariat (the agency overseeing water privatization), nor World Bank staff indicated knowing the details, or even having given it much thought.

Whether there are other protections included in the contract, and on what terms, is also not in the public domain. Key questions requiring detailed public review and debate before any hand over of authority to the lessees could be responsibly undertaken include:

* What provisions exist related to potential lessee breach of contract?

* What provisions exist for damages and penalties against the lessees? What are the standards for imposing these penalties? Who makes such a determination?

* What are the conditions under which the government can terminate the contract? And what obligations does the government owe a lessee in the event of termination?

These last questions are particularly vital. Multinational operators generally seek assurances that contracts will remain in place, and not be subject to political reconsideration. However, such measures may provide one-sided assurances to private operators who may be able to demand large exit payments from governments, even in the event of poor performance.


In the PSP proposal, responsibility for expansion of the piped system rests with the government and the GWC.

Given the enormous need for expansion-approximately one-third of households in the urban areas covered by the PSP proposal do not have piped connections-and given the oft-stated premise of the PSP that the public water sector in Ghana is unreformable and congenitally unable to provide quality service, it may seem strange to leave expansion responsibilities to the GWC.

However, all parties agree that the system itself will not generate revenues to pay for expansion, and that the only viable source of external funding is donor support. Donor support is generally more available for government-controlled projects, and so GWC is left with expansion responsibility.

A consultant to the government, W.S. Atkins, estimated the overall expense of rehabilitation, renewal, improvement and extension of the system as approximately $1.25 billion.

At a November 2000 donors' conference, donors - including the World Bank and donor agencies from the United Kingdom, France Japan, Denmark, Germany and the European Union-suggested they were prepared to provide approximately $400 million over a 5-10 year period These funds will mostly be in the form of concessional (low or no interest) loans, but except for some small grant monies, they will need to be paid back.

The donors indicated a variety of funding preferences Most prioritize expansion, some rehabilitation.

Some of the donor money might therefore be allocated to the lessees for renewal and rehabilitation, through mechanisms that may not yet be determined. Determining the mechanism in advance of entering into the contract is vital It would run against the logic of the deal for the lessees to get a return on the donor investment, but there are no clear safeguards to prevent this. One possible means to avoid such an outcome would be to set aside revenues generated as a result of investments, at least in part, to repay donor loans. In the 2001 Information Memorandum, it appears that the lessees would have first claim on these revenues.

GWC's skeletal staff will be responsible for allocating these monies for expansion projects. They will rely on private contractors for the physical work of installing pipes and extending the system. Whether the lessees may bid as contractors is not apparent, and should be clarified.

Because the lessees will have the duty to service new customers connected to the pipe system, it is unavoidable that they will seek to be involved in the decision on where to expand the pipe system, and pacing of new installation.

Lessee involvement in planning of the expansion and more generally in the determination of use of donor funds raises important concerns about prioritization in investment: What communities will be served? Which consumers will benefit?

The public health, human rights and social equity imperative is to provide service to those unserved and underserved, with priority given to low-income consumers.

This imperative may not match with the lessees' interests, however.

Because the lessees will have an interest in maximizing volume of water delivered and revenue generated, they will tend to prioritize upper-income and industrial consumers over lower-income consumers who will be lower-volume users.

This structural problem of lessee bias against low-income consumers is worsened by the investment prioritization proposed for the PSP and included in the 2001 Information Memorandum.

A prioritization exercise undertaken by W.S. Atkins and reported in the 2001 Information Memorandum assigns weights to different factors in determining investment priorities. Economic merit (meaning the extent to which investments generate new revenues) received the heaviest weighting, weighted as two times as important as the other service criteria factors combined

(An advisory committee had recommended including public health considerations as a criterion, but because "sufficient system level health data was not available," public health considerations were excluded from the prioritization exercise. It is unclear whether this data problem has been or will be remedied in time to incorporate health considerations into investment priorities.)

The overall formula is weighted heavily toward investments that will increase volume as distinct from increasing the number of connected consumers. Furthermore, the formula will favor expansion and renewal in areas closer to already serviced areas (a category heavily skewed to upper-income communities), where investment costs for extension will be lower.

Thus the investment priorities will work strongly to favor expansion and rehabilitation for upper-income and industrial users, and as a disincentive for expanding the system to reach low-volume, low-income consumers.


The 2001 Information Memorandum includes boilerplate language that "a major impetus for the PSP is to improve the utility's ability to consistently provide water of good quality to the largest number of citizens"-with the caveat that this should occur "while moving toward full cost recovery." It adds, "the Government of Ghana felt strongly that LICs [low-income consumers], together with other consumers, needed to reap social, economic and health benefits from the PSP process. The World Bank, DfID [the UK aid agency], and other donors likely to invest in the sector have reinforced this as a priority. Bidders will therefore need to demonstrate their ability to serve these consumers."

At a minimum, it would be important to include clear contractual obligations and performance targets related to providing expanded access and improved service to low-income consumers. The government has plans to establish an Urban Low Income Group Water Unit that would have oversight of the contractual provisions related to service of low-income consumers, though the exact role of this unit is unclear.

However, the 2001 Information Memorandum actually suggests that the lessees will not be expected to take any measures to improve access and service for low-income consumers: "It is fair that the operator will need to devote considerable time in the early stage of a contract to rehabilitation of an existing system and other priorities rather than expansion to LICs."

Instead, low-income consumers are expected to continue to rely on status quo arrangements, specifically to continue reliance on tanker-supplied water.

Urban consumers who are not connected to the piped system, or those who receive irregular supply, are forced to rely on private water suppliers. These suppliers procure their water from the GWC, and then sell at prices five to 10 times the rate for piped water-meaning the poorest Ghanaians pay much more for drinking water than those who are better off. Women and children are mostly responsible for purchasing water, and may carry buckets as far as a mile on a daily basis. Poor families may spend as much as 20 percent of their income on clean water from the tankers.

The 2001 Information Memorandum contains general language on the obligation assumed from the GWC to work with private tankers, and perhaps to take measures such as providing the tankers with more geographically diverse access to water supplies (so as to reduce transportation expenses). "In the case of Ghana, an indigenous system already exist[s] for reaching LICs outside the network so the Lease does not attempt to displace local private providers," the 2001 Information Memorandum states.

However, the existing tanker system is plagued by price gouging and market failures. Some of the worst abuses could be easily remedied by regulation or contractual terms between the water supplier and the tankers that implemented price controls. Either of these approaches will require cooperation from the lessees. Astonishingly, there is no evidence that such terms are contemplated for the contract.

Also apparently given short shrift in the leasing arrangement is the tariff rate for lower-income consumers.

Ghana currently maintains a lifeline tariff, so that the first small quantity of water used by consumers is provided at a much lower rate than larger volumes. The government is debating whether to continue with this bloc tariff arrangement.

While the existence of the lifeline tariff is mentioned in the 2001 Information Memorandum, and while the lessees will be required to charge whatever rates the PURC establishes, it is notable that there seem to be no contractual obligations for the lessee regarding provision to lifeline consumers.

This contractual hole may be important for several reasons.

First, lifeline service is an absolutely vital element in ensuring the broadest access to piped water and promoting service to the poor. Because it is the contract terms above all that determine the lessees' obligations, such a fundamental component of the water system-if the government wishes to maintain it-should be included as a contractual responsibility to eliminate any uncertainty and to emphasize the government's priority for relying on the lifeline to expand access.

Second, enshrining the lifeline in the contract would remove it from political pressures from the lessees. The lessees may have financial interest in eliminating the lifeline, lessening its subsidy component or moving from a bloc tariff arrangement. To the extent that a highly discounted, bloc tariff lifeline reduces revenues, it may harm the lessees' interests. Once implanted into Ghanaian society, the lessees will be powerful political and economic actors, certainly the most powerful force operating in the regulatory arena. If the lifeline tariff is not established as a contractual obligation, it is easy to image the lessees working to roll it back.

Third, the lower revenue stream from lifeline consumers, and the fact that their tariffs will not cover the marginal cost of servicing them, will give the lessees a disincentive to service these lower-income consumers. Absent countervailing provisions embedded in the contract, the lessees are likely to make servicing lifeline consumers a low priority, and particularly to disfavor new connections for lifeline consumers as against higher volume consumers.

Fourth, and for the same reasons that the lessees have a disincentive to service lower-income consumers, they will have a strong incentive to disconnect non-paying lifeline consumers. Money losers even when paying their bills, the lessees will have an incentive to eliminate lifeline consumers from their customer list.

It is striking that completely absent from the performance multiplier, or other portions of the contract arrangement that are public, is any incentive to expand access.

If expanding access is in fact a primary goal of the PSP, and a fundamental public health and human rights imperative, it is inexplicable that there are no contractual incentives to achieve this goal, that this goal does not appear as a contractual obligation, and that no measures are included to offset the apparent contractual incentives to disfavor lifeline, low-income consumers.


The full report of the International Fact-Finding Mission on which this article is based is available at: < Robert Weissman was a member of the fact-finding mission.

Water page

International Monetary Fund (IMF) & World Bank,

Index of Website

Home Page