Identifying Examples of Succesful Cost Recovery Approaches in Low Income Urban and Peri-urban Areas

March 2003



South Africa has a history of inequitable distribution of water services, skewed by racially discriminatory policies and severe socio-economic inequality. The advent of democracy not only gave all South Africans the right to participate freely in public affairs, but it also constitutionally entrenched rights to eventual access to basic social and economic services. The approach of the Department of Water Affairs and Forestry (DWAF) - outlined in a 1994 white paper - paid careful attention to the financial dimension of expanding basic water services. Drawing in part on conventional thinking in international development organisations, this approach sought to recover recurrent (operation and maintenance) costs of basic services through user charges.

Difficulties in securing user payments soon emerged as a major threat to the continued expansion of basic water services. In 1998, DWAF admitted that, "many of the Department's RDP water projects are proving unsustainable as cost recovery is not taking place."

The purpose of the study is to identify the main determinants of successful cost recovery for water services in South Africa and to use this information in the development of practical strategies to overcome obstacles to cost recovery.

Project Objectives

The purpose of the project as a whole is to identify the main causes (determinants) of successful cost recovery for water services in South Africa and to use this information in the development of practical strategies to overcome obstacles to cost recovery. As stated in the project contract, the aims are:

All the aforementioned project objectives have been achieved in this Final Project document.

Project Methodology

The project used the following methodology:

  1. Consulting stakeholders such as the Department of Provincial and Local Government, Department of Water Affairs and Forestry, the South African Local Government Association and selected local authorities on the issues, research, design and products required.
  2. Undertaking a national survey of local authorities to identify issues affecting cost recovery. The main objective of the national survey was to collect quantitative data on the cost recovery situation countrywide.
  3. Reviewing literature and existing data.
  4. Identifying four successful local authorities and undertake case studies (in-depth review).
  5. Synthesising findings into a report and identify broad trends and access of interventions

The principal aim of this study is to account for differences in cost recovery outcomes, with particular attention to issues related to low income households. It investigates the interplay between socio-political and institutional context, service infrastructure, and billing and payment policies and procedures in influencing payment rates and debt ratios in diverse municipalities countrywide.

The National Survey

The principal aim of the national survey was to establish a database of experience at the national level with particular attention to issues related to low income households. It investigated the interplay between socio-political and institutional context, service infrastructure, and billing and payment policies and procedures in influencing payment rates and debt ratios in diverse municipalities countrywide. The empirical findings regarding the causes of successful (and unsuccessful) cost recovery are used to offer guidelines for the development of more effective cost recovery strategies at the municipal and national levels. The national survey data supported the kind of descriptive audit that was planned, but it also made it possible to go further - using statistical techniques - to draw causal inferences about the determinants of successful (and unsuccessful) cost recovery. Although it was not initially clear that the national survey data would be suited to such an exercise, the analysis is directly motivated by the aims of the broader project - to explain why cost recovery succeeds in some cases and fails in others, and to use these empirical findings to inform the development of effective cost recovery strategies.

The national survey of cost recovery yields the most up-to-date available information about water service infrastructure, billing and payment, and cost recovery outcomes in South Africa. The questionnaire was designed to capture as much relevant information as possible while minimising time demands on respondents. Both for respondents' convenience and to ensure an adequate response rate, the time required for completion was limited to approximately ten minutes. The final questionnaire contained 30 questions - of which 16 had structured ("multiple-choice") response options. For questions where respondents were asked to enter specific figures, they were instructed to give their best estimates if exact information was unavailable or inaccessible.

310 questionnaires were returned. Responding municipalities served a total of 3.2 million households, probably representing more than 40 percent of all households receiving municipal water services in South Africa. The data is summarised in the Tables below.

Descriptive Data on Cost Recovery Outcomes
Municipal data, unweightedmean, with standard deviation
Municipal data, mean weighted by household consumers, with standard deviation
Multiply-imputed data, unweighted mean, with standard error
Payment rate 60.7
Ratio current to total consumer debt:
Consolidated accounts 21.7
Water only 25.4

Note: Source is national survey of municipal cost recovery, 2000. Measures of dispersion around the mean (standard deviations and errors) are in italics. Numbers of non-missing observations are in parentheses.

The results highlight severe constraints on cost recovery - especially for the provision of basic water services to the poor. Combining consumer poverty, low "RDP" service levels (e.g. public standpipes), and the associated inability to punish non-payment by selectively restricting services is the surest recipe for failed cost recovery. Although case studies have identified a few potentially instructive exceptions to this rule, the national survey data suggest that expectation of recovering the operation and maintenance costs of "RDP" water services through user fees is an unrealistic premise for national policy in the foreseeable future. Meanwhile, the national government's adoption of a "free basic water" policy - itself partly a response to the difficulties of collecting user fees from this segment of consumers - forces careful attention to issues of municipal cross-subsidisation and redistribution through the national "equitable share" mechanism. Neither municipal demarcation nor government commitment to "free basic water" solve the cost recovery challenge, but they substantially alter the institutional and policy environment in which that challenge is addressed.

Analysis: The Causal Model

We constructed a causal model to identify the major factors that explain why some municipalities recover costs for water services while others are less successful. In identifying potential causal relationships, two requirements are paramount. The first is the existence of a logically coherent and compelling account of the process through which the cause is expected to bring about the effect. The second is that the expected effect of a causal factor (or combination of factors) should occur subject to an "all else being equal" condition.

To highlight the concrete implications of this study's findings, the statistical estimates are presented by simulating the effects on cost recovery outcomes of altering features of an otherwise "typical" South African municipality. The analysis was conducted in two stages - the first in which statistical estimates were calculated for data on actual municipalities, and the second in which simulations were run to illustrate their implications. Because the first stage is fairly technical and the "raw" statistical estimates difficult to interpret, it is reported in an appendix. The second (simulation) stage allows presentation of the same findings in a much more accessible format - expressing the expected effects of altering various features of a typical South African municipality as percentage changes in payment rates and debt ratios.

The simulations focus on expected effects likely to be of broad interest. The first examines the impact of upgrading service infrastructure to metered private connections. The second considers the imposition of service restriction as a penalty for non-payment. The third focuses on the effects of other changes in billing and payment practices - the introduction of a progressive tariff, an active Masakhane campaign, and payment at supermarkets. The fourth combines the impact of each of the first three in order to give a general sense of the degree of influence municipalities have over cost recovery outcomes through decisions about service infrastructure and billing and payment practices. The fifth and sixth consider the implications of poverty and, more specifically, the extension of basic services to the poor.

The graph below presents the simulation results. The "before" assumptions are of an otherwise typical municipality in which roughly one in five household consumers lacks a private, metered connection; the tariff schedule is not substantially progressive; payment facilities are not available at supermarkets; and there is no Masakhane campaign. In short, the "before" assumptions are of a municipality pursuing "worst practices" for cost recovery. Then, in the "after" simulations, all of these "worst practices" are replaced by "best practices." Because few actual municipalities approximate the "worst practice" assumptions, the improvements in payment rates and debt ratios available to them by adopting "best practices" will generally be somewhat smaller than the simulated results presented in the figure. The simulations are, however, useful in highlighting the overall sensitivity of cost recovery outcomes to factors within municipalities' control.

The graph confirms that for a typical municipality the combined effects of moving from "worst" to "best practices" are substantial. The upper graph shows that eight of ten simulations fall in the range of 28 to 43 percent improvements in payment rates, with an average improvement of 36 percent. For perspective, the average simulated payment rate for a typical "before" municipality is 47 percent, while for an "after" municipality it is 83 percent. The lower graph shows that the combined effects on debt ratios is also substantial - with the shaded range stretching from 16 to 37 percent, and an average increase of 27 percent. The corresponding average "before" and "after" debt ratios are 10 and 37 percent, respectively. Municipalities may clearly exert heavy influence on cost recovery outcomes through their own choices and actions.

An unfortunate implication of these results is that extending basic services to the poor shows a strong tendency toward harming cost recovery outcomes. Achieving the goal of financing basic services through improved cost recovery thus appears unlikely in light of the empirical patterns illustrated in the simulations. As pessimistic as this conclusion may seem, it should not be surprising given the findings presented in the analytical sections of the report.

Combined Effects of Possible Cost Recovery Interventions
<Simulation results Shaded areas contain 80 percent of simulated outcomes


BEGIN with an otherwise typical municipality where:

Roughly 20% of household consumers lack private, metered connections;

Non-payment is not penalised through service restrictions;

There is not a progressive tariff, payment at supermarkets, nor an active Maskhane campaign.


Upgrade existing consumers to private connections, introduce penalties for non-payment and other billing and payment "best practices"

The Case Studies


Four municipalities that have had relatively good experience with cost recovery were selected: Mogale (Krugersdorp), eThekwini (Durban), Middelburg, and Nxuba (Alice). The municipalities exhibit differences in terms of size, population, resources etc. and it is recognised that communities present a range of situations related to their specific histories, experience with development projects, access to natural and human resources, models of governance and social organisation and cost recovery systems. The case study approach gives a longitudinal view of the process, which complements the snapshot view provided by the national study. Furthermore, it was hoped that the in-depth review, which the case study methodology format allows, would provide insights by examining both the negative and the positive aspects of the experience that the municipalities had gained.

More specifically, the case studies aimed to do the following:

The sample includes population groups ranging from 3,800 households to approximately 64,108 households. Variety can be seen in the roles and influence of water boards, community representatives and local government in each community. The study examined both bulk water supply projects and those drawing upon local groundwater resources.


A number of very interesting similarities have emerged from the case studies which suggest a basis for a code of best practice. These are:

Preliminary phase
Community perceptions with regard to cost recovery are a very important factor in successful cost recovery. These perceptions will be strongly influenced by the conduct of the municipality over a period of time.

This may include combining increasingly strict enforcement of regulations regarding non-payment with demonstrations of good faith by making investments in infrastructure and upgrading plumbing.

Community involvement
Such a carrot and stick approach must be supported by community interaction. Municipalities must provide dedicated and well-trained staff to communicate with the communities, and act as a go-between between them and the municipality. They should assist residents to be aware of how a council operates: what rules are being used, how water consumption is measured, the different ways of paying for water, budgeting and the like.

Where pre-paid meters are not being used it is generally considered desirable for electricity to be linked to the water account, and for both services to be billed together. If so, electricity can be cut off instead of water. If this is not possible a link should be made between electricity and water so that a defaulter runs the risk of both being terminated.

The quality of administration makes an important impact. Good administration must be:

Treatment of consumers, especially in the case of default
The following are important:

The problem of vandalism and illegal connections

Inevitably, one of the responses by consumers is to tamper with meters or bypass them. How the municipality responds to this behaviour is important.

Vandalism is a different situation. If it can be reasonably deducted that meters have been vandalised, then the householder must pay a replacement fee.

In this section we discuss two aspects: pre-paid meters and flow limiters.

As systems become more and more sophisticated, data collection becomes easier. This allows monitoring of individual accounts as well as overall water use and income levels etc. While there is room for congratulation in terms of the recovery of current debts which is an important step forward for many municipalities, there remains a substantial minority of non-payers. Detailed age analysis of the accounts helps to focus the attention of the municipality on such cases, and target resources at those cases. This allows a better definition of which households are truly indigent and which are simply evading payment.


There can be no shortcut to successful cost recovery. It must be a broad-brushed approach involving technical, financial and community aspects. It must be implemented with a unity of purpose between the political and administrative arms of the municipality. Finally it must be seen as a sustainable means by which the standard of service delivery to the public can be enhanced.

There is a close fit between the empirical observations from the case study and the result of the simulations from the National Survey. Predictors of success are individual water meters (in whatever form), prompt and firm response to non-payment, progressive tariffs, good community relations and convenient payment facilities.