Chapter 4
Linking charging objectives and methods

Introduction

There must be clear linkages between policy objectives and charging methods, and consistency with other activities, investments and projects in the sector. The conclusions of the previous chapters are that two issues dominate current priorities - recovering costs to achieve financial sustainability, and limiting demand for water to the environmentally sustainable level. A third issue - providing a mechanism that facilitates transfer of water use from lower to higher value uses (whether ‘value’ is financial, economic, social, or whatever) is of significant concern to planners and policymakers, and can be addressed through tradable water rights. Table 2 summarizes the relationship between different pricing tools and these policy objectives.

The most obvious choice for a system to recover costs is an easy-to-implement, non-volumetric method, such as area pricing or output taxation. Area pricing has little or no effect on demand for water by individuals and its effect on water productivity is negligible. Indeed, imposition of high, non-volumetric, charges may lead to higher water use as users feel entitled to use as much as they want, as it has been paid for. Volumetric methods are intrinsically unsuited to the recovery of costs, because while the costs of system operation are relatively stable, the revenues from uncertain sales - possibly compounded by uncertain prices - provide an operating agency with limited financial security. Tradable water rights separate the issue of revenue generation (the payment by users for the cost of supplying) from pricing the resource (payments among the holders of water rights for the benefit of use).

Where the objective is to control demand, there are three options: quotas (which is not a charging mechanism); volumetric charges at levels sufficient to induce the required reduction in demand; or tradable water rights. Chapter 3 indicates that volumetric charging is not observed as a mechanism to bring demand and supply into balance - this is better achieved through water allocations (quotas). Volumetric pricing can promote some degree of water conservation. However, at the prices normally applied, their influence on the level of demand expressed is slight. Formal water markets require the allocation and monitoring of water allocations in order to function, which in turn provides a mechanism to manage overall demand. When well developed, such markets can also facilitate the redistribution of water between sectors.

This chapter analyses the case studies undertaken for this project, plus the findings from the wider literature review, in order to assess the extent to which the objectives of cost recovery and demand management are addressed and achieved in practice. Table 3 presents a summary of the financial and economic aspects of scheme performance in the case studies. Figures 3 and 4 place the case study data within the wider global range. The world-wide data have been compiled from the extensive sources listed in the bibliography. A tabular summary of the six case studies, assessing the extent to which either cost recovery or demand management objectives are realized, is included in Annex 2. A full presentation of the material which underpins the analysis is to be found in two reports, Bosworth et al (2002), Cornish & Perry (2003).

TABLE 2

Summary of the relationships between irrigation charging objectives and pricing methods

FIGURE 3
Global range of irrigation water charges per hectare

Achieving cost recovery

All the case studies report the objective of recovering O&M costs from beneficiaries, while three (Gujarat, Former Yugoslav Republic of Macedonia and Morocco) also report the objective of securing a substantial element of capital cost recovery. However, in only three of the studies - Haryana, Gujarat and the Tadla system in Morocco - do the fees recovered actually cover or exceed the O&M costs. In the Haouz scheme, Morocco, where the O&M costs per hectare are less than half those at Tadla (a smaller and older scheme), fees are not levied in extensive ‘traditional’ areas and the scheme receives large annual subsidies from central government in compensation.

In the case of the deep tubewells in Gujarat, farmers cover all O&M costs as they arise, but the state provides a large subsidy on the operating costs through the tariff structure applied to electricity for agriculture. If this subsidy were removed, it would prove uneconomic for farmers to continue pumping water to irrigate alfalfa unless the price of the crop increased.

FIGURE 4
Global range of irrigation water prices per cubic metre

Table 3 shows that the tubewell farmers in Gujarat and farmers on the Tadla and Haouz schemes in Morocco pay 37, 17 and 7 percent, respectively, of their net income in water charges. In all three cases, the recovery rates are high - 100 percent in the case of Gujarat. In these cases, the fee collection process is transparent and the fees collected are retained and used on the scheme. Farmers understand, and are content with, the service provided. In Haryana, if the present cross-subsidy from other sectors were removed, farmers would face a threefold increase in the water fee to meet required O&M costs. However, this would still only represent 1.5 percent of net farm income per hectare. Here, farmers’ unwillingness to accept, rather than their inability to pay, a large price increase would be the principal constraint. In the example of the Former Yugoslav Republic(FYR) of Macedonia, farmers would face a doubled water price (of US$0.04/m³) to meet estimated necessary O&M costs, equivalent to 20 percent of average net income. Such a price would be very high by international standards. Given the weakness of the agriculture sector in the FYR of Macedonia and the condition of irrigation infrastructure, the goal of recovering even annual O&M costs from irrigation charges appears unrealistic.

In Nepal, where agricultural productivity is low, the ability of small farmers generally to pay fees high enough to cover annual O&M costs must be questioned. On those schemes which are in better condition and delivering better service, the full O&M fee would represent about 5 percent of net income per hectare, which is affordable. However, on deteriorated surface schemes, where total average farm incomes are lower, the proportion rises to 8 percent and is greater still for pumped schemes. 'Affordability' is not a simple criterion. Where irrigation supplements rainfall, as in Nepal, it is important to assess the net incremental return to irrigation, that is, the net return to irrigation in excess of the income that could be obtained from rain-fed farming. This is because it affects farmers’ willingness to invest the greater money and effort needed for irrigated agriculture. For schemes in arid areas, where crop production cannot occur without irrigation, the net return is effectively the same as the net incremental return. However, in supplementary irrigation, incremental net returns to irrigation are substantially less than net returns to agriculture because rain-fed production may produce significant yields. In the best surface scheme investigated in Nepal, the annual incremental net return to irrigation of the average farm (2.1 ha) was some US$205 -of which full O&M fees would take 11 percent -compared with a net return to irrigated agriculture of US$507. On the worst scheme (Tilawe), the incremental net income was US$82 for the average farm (1.1 ha), of which necessary O&M costs would take some 30 percent. The average incremental income in unfavourable seasons would be substantially less than these figures. For the poorest farmers on a scheme, whose incomes are barely at the subsistence level, raised irrigation charges would be particularly difficult to meet. In such conditions, both the ability and willingness to pay of small farmers must be doubted. Thus, for the general run of schemes in Nepal, it seems that the returns to water would have to rise before a general policy to recover full O&M costs could be considered.

TABLE 3
Summary data from the case studies

Notes:

1. Required O&M cost in Haryana is low because of to cross-subsidy from M&I sectors.
2. Income per hectare after deduction of all production costs except the cost of water
3. Fees shown recovered at Haouz are levied on a part of the area only.

In Sindh Province, the fee would again represent about 4 percent of net income per hectare, although the price per cubic metre is low in comparison with the other case studies outside the Indian subcontinent. In practice, existing fees are currently well below even this low figure, and with fee collection rates at less than 30 percent. Irrigation in Sindh remains entirely dependent on large, annual government subsidies. One factor that makes farmers in Sindh better off than their counterparts in Nepal is the average size of an irrigated farmholding. In Sindh, holdings average 6 ha, while on many Nepali schemes the average is little more than 1 ha. Thus, although average net incomes per hectare in the two cases may be very similar, the larger holdings in the Sindh provide a better household income, making it more likely that farmers could pay to cover annual O&M needs.

With the exception of the studies on Nepal and the Former Yugoslav Republic (FYR) Macedonia studies, Tables 3 and 4 suggest that it should be financially feasible to recover full annual O&M costs from farmers. However, legal, political, social and administrative constraints may prevent the effective recovery of fees. Furthermore, while farm budget data demonstrate that the 'average' farmer on some schemes should be able to pay fees to cover full O&M costs, there will always be poorer farmers obtaining lower net incomes. Therefore, careful analysis is required before drawing general conclusions about 'affordability'.

TABLE 4

Current fee levels and fees required to cover annual O&M costs

Notes:

1. Income per hectare after deduction of all production costs except water. In some cases, average of a range

2. An assumed annual irrigation depth has been used for each case study where field data were not available.

3. Net water value is the farmer’s net income per hectare divided by water used.

4. Based on an estimate of O&M fees in Haryana if cross-subsidy from other sectors were removed.

5. Water consumption based on supplementary irrigation to monsoon crop.

Of the three case studies where it is aimed to secure a substantial element of capital cost recovery (Gujarat, FYR Macedonia and Morocco), only the small, user-funded and user-managed deep tubewells of Gujarat are successful in recovering capital expenditure. Of the Moroccan schemes, management at Tadla regularly recovers more than its O&M expenditure but does not achieve a target of 40 percent of capital costs. The situation at Haouz is unusual: fees would be sufficiently high to recover O&M plus capital costs if universally levied. However, in practice they are only charged to a fraction of the total number of beneficiaries. In FYR, Macedonia, the area irrigated is declining and infrastructure is being abandoned. The reasons are complex but it seems that water management organizations cannot afford to pay the costs of historical bad debt and capital depreciation.

It is difficult to obtain reliable data on the magnitude of charges required to recover the full supply cost, i.e. including capital costs, on most schemes. However, it may be that the steep fall in agricultural commodity prices in recent decades - to the benefit of the non-agricultural poor - means that full recovery of O&M plus capital costs can seldom be achieved. Box 11 describes the changing approach to capital cost recovery adopted by the United States Bureau of Reclamation (USBR) in the United States of America.

Improving cost recovery - the role of irrigation management transfer

The discussion on what price to charge should be secondary to the issue of improving recovery rates for any charges that are set. The summary information presented in Annex 1 shows that recovery rates of water charges are usually below 50 percent. Moreover, in several countries, levels of recovery have declined. In Pakistan, recovery rates dropped from 38 percent in 1994/95 to 26 percent in the following season; in FYR Macedonia, they fell from 88 percent in 1990 to 35 percent in 1995; in India, they decreased from 64 percent in 1974-75 to 8 percent in 1988-89 (Saleth, 1997); and in Suriname, water users stopped paying water charges altogether and water boards practically disappeared (Risseeuw, 1997). In Croatia, no charges are collected for irrigation because the area and income involved are too small to justify the costs of collection (Ostojic and Lukšic, 2001).

Transfer of poorly maintained irrigation systems to users is often seen as a way to improve O&M and to reduce costs to the state and the users. However, where systems are already chronically under-funded, the result will be further deferral of maintenance and general decline. Governments then have to borrow to rehabilitate their irrigation systems, creating a vicious circle. Underassessment of the true cost of service provision can be as great a problem as non-payment of fees by farmers.

Reviews of irrigation management transfer (IMT) policies demonstrate that recovery of water charges does not necessarily cease to be a problem after management transfer. Many transferred schemes struggle to enforce fee collection. For example, in Colombia, the level of cost recovery declined in two out of three case study areas after IMT (Vermillion and Garcés-Restrepo, 1998). These findings reinforce recent appreciation that governments’ support for transfer should not be terminated abruptly following transfer. Institutional, rather than financial, support is likely to be needed for some time post-transfer.

Demand management and water saving

Ray (2002) discusses the assumptions implicit in ‘getting the prices right’ in order to deter wasteful use of water and to achieve irrigation efficiency (Box 12). Ray concludes that these assumptions are invalid for India, and that enforceable and transparent allocation rules may be more effective to curtail water demand. Molle (2001) reaches similar conclusions for Thailand.

'Getting the prices right' is a complex issue, not least because of the multiple uses which may be made of water from irrigation systems. As indicated elsewhere, detailed discussion is not made here of such issues, but they are addressed by e.g. Facon (2002). Some authors suggest that the value of water (or any other resource) is the maximum amount the user is willing to pay for the use of the resource. Where pricing is used as a regulatory measure to ensure efficient allocation of water resources between sectors, all economic costs - including opportunity costs, positive and negative economic externalities - need to be taken into account.

A major obstacle to the introduction of volumetric water pricing is the high cost of the required measurement and billing system. Where farmers are many and farm sizes are small, the process of monitoring water use, billing and collecting fees is difficult and carries high overhead costs. Water measuring devices at farm level and institutional infrastructure are rare and generally unserviceable in the developing world. Perry (1995 and 2001a) shows that in Egypt and the Islamic Republic of Iran, the costs of charging individual farmers are likely to outweigh the projected benefits. Apart from some higher-income countries, the high costs of associated with charging individual farmers present problems for volumetric pricing. A study of OECD countries shows that volumetric pricing in agriculture is not widespread because of practical difficulties (OECD, 1999). As long as the transaction costs remain a high percentage of the revenue collected, or of the value of the production, volumetric measurement is difficult to justify.

Source: Ray, 2002.

In parts of China, volumetric supply and billing to townships (and nowadays some WUAs), coupled with pre-payment for water, appears to be successful. Acknowledging the considerable differences in social and institutional arrangements, there may be lessons here for other parts of SE Asia, as Facon (2002) argues.

In Gujarat, Haryana, Morocco and Pakistan, water is particularly scarce, agricultural demand continues to grow, and there is increasing competition between sectors for a limited resource. Nonetheless, agencies in those states and countries have not explicitly identified water demand management as an objective of pricing, although Morocco does charge volumetrically below an imposed limit.

Case study areas

Gujarat

In the study area in Gujarat, aquifers and groundwater quality are in decline, and the government currently has no mechanisms to address the issue other than the price of electricity. As there is full flexibility and control over water deliveries at farm level, volumetric pricing is feasible and indeed it is practised by the farmers’ groups who own wells. Although the state subsidizes electric power, current prices for irrigation water in Gujarat are high by international standards (Figures 3 and 4). However, the high price has failed to prevent overexploitation of the aquifers. Given the site-specific nature of farmers’ crop selections, varying market prices for crops, different water table depths and the variable yields of different aquifers, it is unlikely that a uniform policy of increased energy prices could reduce irrigation demand to the levels needed to sustain ground water levels.

The twin problems of overexploitation of aquifers and near-bankruptcy of state electricity companies are linked. However, it seems unrealistic to expect a pricing policy for electrical energy to solve both problems. If the flat-rate tariff structure for agriculture is replaced, so that farmers pay for the amount of energy actually consumed, the energy suppliers may be returned to solvency, provided that collection can be enforced. Practically, it is difficult to envisage an energy price alone achieving a balance between water supply and demand, since the price needed to achieve any sort of effective restraint, like the water charge on surface schemes, would need to be raised many times over against vocal public opposition. Without the introduction and implementation of effective abstraction licensing laws, it seems that balance will only be achieved by the draw-down of aquifers to a point where further pumping becomes uneconomic, or the aquifer becomes too saline for productive use.

Haryana

The State of Haryana addresses the issue of water scarcity by a system of water allocation. Allocation decisions between sectors are made at the political level, and that between farmers is based on well-established and fixed rules. First, the water demands for nonagricultural and agricultural use are prioritized at state level. Expected allocations to irrigation are then drawn up, based on reservoir storage levels and projected inflows. Finally, the water allocated to irrigation is distributed proportionally across all areas. This provides a transparent and easy-to-manage distribution system. However, control structures are located in the upper levels of systems. Thus, there is essentially no scope for moving towards more flexible and volumetric delivery with pricing at the farm level. It can be argued that this system of water allocation, both between sectors and among farmers, is an economic instrument in that allocation decisions are based on the perceived value of a scarce resource. However, the system does not rely on the cost to the user of the resource either to control demand or achieve a planned distribution, and, indeed, the present water price is insignificant to the relatively affluent large farmers of the area. The preferred crop choices (cereals) do not provide the farmers with obvious motivation for introducing water-saving irrigation technologies.

Increasingly, the system does not operate as well as formerly, in the face of a number of adaptations to operational practice, of de-facto physical changes and a changed role as farmers increasingly supplement supply with groundwater. Nonetheless, it provides a well-codified and recognised framework for water management.

Morocco

Irrigation authorities in Morocco depend on a combination of relatively high water charges, which discourage waste, and seasonal allocations to ensure that supply and demand are in balance. It is the allocation system that constrains use. The vast majority of farmers would take more water at the prevailing price, and development of more expensive groundwater within irrigated areas is progressing rapidly. Increases in volumetric prices to discourage demand for surface water may lead to unsustainable use of groundwater, which is a serious problem in many areas outside the irrigation schemes. Furthermore, since much of the groundwater comes from the formal irrigation system, reduction in use of surface supply will have corresponding effect on the availability of groundwater.

Pakistan

In Pakistan, water is scarce, and the underlying legal, administrative and system design is the same as in Haryana. However, performance in terms of cost recovery and allocation of water at the farm level is far inferior. Water charges are lower and collection rates are poor. The physical system has not been maintained well (which leads to inequitable distribution), and it has been tampered with. Thus, the design, intended to ensure automatic distribution of the proper amounts of water to each area, is circumvented. Scarcity of water is ‘managed’ by delivering what supply is available - meaning that those with best access to water take as much as they want, while those with poorer access have what may be left.

Volumetric charging in practice

It is helpful to compare the unit value of water (represented by net farm income divided by quantity of water used) with the unit cost of water (indicated by the irrigation charge divided by the quantity of water used) - columns 4 and 5 in Table 4. The higher the value-cost ratio becomes, the greater will be the demand for water, and the larger will be the increase in price required to influence demand. In the case studies, the indicator varies from a value of about 3:1 for tubewell users in Gujarat to about 180:1 in the surface systems of Haryana. Nepal and Pakistan show ratios of 125: 1 and 50:1, respectively. Farmers are generally unlikely to reduce their consumption of water significantly until its cost begins to approach its value to them. As water prices generally still represent only a small percentage of farmers’ net incomes (Box 12), increases of more than an order of magnitude would apparently be needed to influence demand. It would generally be politically and socially unacceptable to enforce change of this magnitude.

The impact of volumetric water pricing and farmers’ response to increased charges depends on a variety of factors. Box 13 provides an overview of factors that influence the price elasticity of demand for irrigation water - that is the change in quantity demanded for given change in price.

Irrigation water demand curves in Spain exhibit a perfectly inelastic (non-responsive) stretch at low prices and become elastic beyond a certain threshold (Varela-Ortega et al., 1998). As prevailing prices are low, this implies that only considerable increases in price, i.e. setting the price above the threshold, will induce the desired efficiency. Existing low prices of water may be the main reason why farmers are not very responsive to price changes. Moreover, factors other than price may have a greater impact on the quantity of water demanded, e.g. climate variation, agriculture policy, product prices, and the reliability of the water supply. Malla and Gopalakrishnan (1995) report that price increases in Hawaii had no significant impact on water use, as climate factors, such as rainfall, primarily determined water use decisions.

Source: USBR (1997).

Ray (2002) stresses that even where volumetric pricing leads farmers to improve their water use efficiency, they can only improve the management of that fraction of diverted or released water that reaches their fields. On many large surface-irrigation schemes this might be as little as 25 percent; the rest of the water released may be lost in conveyance and no pricing policy is likely to address these losses.

Berbel and Gomez-Limon (2000) estimate that farm incomes in Spain will have to decrease by 40 percent before water demand decreases significantly. Perry (1995) estimates that inducing a 15-percent reduction in water demand in Egypt through volumetric pricing would decrease farm incomes by 25 percent. The study by Ray (2002) on water pricing in India uses an analytical model to show that in order to induce the water-conserving response under existing allocation practices, a sixfold price increase would be needed. In the Islamic Republic of Iran, to be effective in curtailing demand, water prices would need to rise by a factor of ten (Perry, 2001a). Price increases of this order of magnitude are quite unlikely in the prevailing political context and thus the political feasibility of volumetric water pricing, as a tool to curtail demand, is questionable. This is particularly valid in most developing countries where existing water prices are very low - well below the threshold where a significant response to price is seen.

Severe impacts on farm income are implicit in using water pricing as a means to limit demand.

Enforceable and transparent allocation rules and abstraction licences may be a more effective way to curtail demand (Perry, 2001a; Ray, 2002). Although there are clear difficulties in enforcing allocations, the previous sections have shown that well-established systems such as warabandi in north India can be effective, and that the effective enforcement of unpopular pricing measures is arguably considerably more problematic.

Theoretical analyses made for the US, Bernardo and Whittlesey (1989), suggest that farmers in Washington State would aim to reduce water use at the cost of increased labour by switching to a more efficient use of their current irrigation technology. Consequently, under restricted supply (rationing), water use apparently could be reduced by up to 35 percent for surface irrigation and 25 percent under centre-pivot schemes, without greatly affecting farmers’ incomes. Hoyt (1984) reaches similar conclusions for groundwater use in the Texas High Plains. However, increased water extraction costs and crop prices appeared to have no significant impact on the efficiency of water use. Hoyt argues that, owing to the inelastic demand for irrigation water, reliance on price mechanisms to conserve water has limited impact in the short run. Only if prices increase dramatically, do capital investments in more efficient irrigation technology become viable - at considerably reduced profits.

It is also important to the issue of saving water that 50% or more of water lost on surface irrigation systems occurs in the main system, upstream of the area traditionally controlled by farmers. Charges levied on farmers can therefore clearly make no impact on a large part of the total system 'loss'.

Summary

All the government agencies reviewed in the case studies state that their primary objective in charging is to recover O&M costs. In practice, the objective was only realized in three of the eight cases. Capital costs were only fully recovered in one case and, partially, in a second case.

In most cases, the benefits of irrigation substantially exceed the basic costs of delivery. Farmers’ unwillingness to pay generally constrains efforts to recover O&M costs. However, in the poorest countries, small farmers may face real difficulties in paying the full O&M cost. Governments may need specific policies to assist the poorest farmers in such circumstances.

Management needs to establish a consistent and reliable water supply, with transparent systems for assessing and collecting fees and sufficient in-built flexibility to meet farmers’ needs. Charging systems need to be supported by clear land and water rights and the effective rule of law. Where irrigation services are deteriorating for lack of proper maintenance, it is in the interests of all beneficiaries to ensure that services are continued, by paying the fees necessary for O&M. However, it may be necessary to bring about considerable political and institutional change to gain the confidence of those who are to pay for the service. Such pressures are encouraged by transparency in accounting and effectiveness of water delivery. From their side, farmers must perceive true benefits in paying for what they have often viewed as a 'free' government service.

For the cases investigated, the unit value of water in agriculture was compared with its unit cost. Apart from the privately-owned wells in Gujarat, high ratios on all case study schemes suggest that the price of water would have to rise by at least an order of magnitude to affect demand. Such price rises would be difficult to enforce and threaten the livelihoods of the poor. The wider review of literature supports the concern that the responsiveness of demand to water pricing is usually low. Models indicate that if price is used to bring about substantial reductions in demand this will usually be at the cost of disproportionately greater reductions in net farm income.

There are very few places in the world where pricing is the prime mechanism for constraining irrigation water demand. High marginal prices for water will prompt some reassessment of water use by farmers and a more conservationist attitude, but moves to balance supply and demand in overexploited basins are led by water allocation. In the case study locations where water is particularly scarce, controls over water allocation (rationing) are used to limit consumption. In Haryana and Pakistan, water is distributed under the warabandi system. In Morocco, farmers order and pay for water on the basis of volume, but their overall use of the resource is limited by a fixed quota.

If water fees were increased to the levels required to bring about significant reductions in demand, there would be a substantial threat to the livelihoods of smaller and poorer farmers. The case studies do not provide evidence on this point because none of the countries concerned has pursued such a policy. Nonetheless, it is concluded that water demand management through bulk water allocations or through a system of tradable water rights would better protect the interests of all farmers, especially the poorest. As with other economic instruments, these approaches also require significant and sustained political support, and the technical infrastructure to measure allocated volumes and permit water transfer between users.