F. van Steenbergen, Sociologist, Euroconsult, Arnhem, The Netherlands
This paper discusses water delivery schedules in farmer managed irrigation systems, using examples from Baluchistan (Pakistan). The assumption is sometimes made that, when resource management is user controlled, changes towards more efficient use will come about automatically. The examples from Baluchistan demonstrate that such changes do not necessarily occur and correlates the lack of change to the specificity of the schedules, the different gains from the change and the importance of objectives other than improved irrigation water management.
To this day water delivery schedules continue to be a 'big unknown' in irrigation management. This is unfortunate, because their potential for increasing agricultural production from ever scarcer water is substantial. One reference (Reidinger, 1974) has described how agricultural production losses occurred as a result of the unpredictability of water supplies under a particular system of canal rotation in Hissar District in Haryana (India). The erratic surface water deliveries were at variance with the cultivation of modern varieties and prompted farmers to develop private tubewells. Similarly, farmers had recourse to conjunctive irrigation in Leyyah District in Punjab Province (Pakistan), after a restricted canal rotation schedule was introduced in the 1970s. Yet in contrast to Hissar, in Leyyah the reduction in canal supplies and the development of tubewells had a beneficial effect, lowering groundwater tables and removing waterlogging.
Water delivery schedules deserve more attention than they receive at present. Regrettably, the discussion on irrigation management in the last few years has been overshadowed by other themes, in particular privatization, irrigation management transfer and water markets. With regard to the latter, several authors, such as Shah et al. (1993), argue that commercial water transactions will contribute to efficient water use by facilitating the reallocation of water to the user that is willing to offer the highest price. This is often assumed to be the most efficient producer. To make the transactions possible, non-land-based water rights are considered essential (Rosengrant and Binswanger, 1992). For permanent water transfers this may indeed be most appropriate. (This statement overlooks the informal temporary exchanges of water turns, common throughout the world of irrigation, that take place in the absence of a formalized water rights system. Similarly, informal permanent water transactions take place independently of alienable water rights and occur even in systems where water transactions are legally forbidden.) Yet water transactions are similarly facilitated or frustrated by water delivery schedules. The scope for water sales in the rice-based irrigation systems of east Asia for instance is limited, as no one can extract himself from the rigid field-to-field delivery schedule.
The relevance of water delivery schedules goes beyond whether or not they facilitate water transactions. Reliability and appropriateness of water deliveries in terms of timing, volume and flexibility are important assets. They represent social capital. In this paper, water delivery schedules in three small-scale farmer-managed systems in Baluchistan Province (Pakistan) are discussed. In each of these systems water is very scarce, the rights to it are individually owned and the systems are entirely user controlled and managed. The water delivery schedules, unlike in large scale systems, are determined by the users themselves. With compelling simplicity, several authors such as Korten (1990) and Pretty and Guyt (1992), have argued that such local control of productive resources facilitates efficient use. The same theme surfaces in the discussion on irrigation management transfer: under farmer management, water utilization will supposedly be more efficient. Yet, is this true and to what extent? Do local institutions capture the opportunities for improvement under local control? This paper discusses institutional change in water delivery schedules in the selected farmer-managed systems. It reviews, if there is obvious scope for improvement, acknowledged by the users, whether these improvements do or do not occur. It explores what factors encourage adjustments and what factors cause, to use a term introduced by North (1990), institutional 'lock-ins'.
FARMER-MANAGED WATER DELIVERY SCHEDULES
The attributes of water delivery schedules encountered in Baluchistan resemble those in small scale systems in other arid parts of the world (Geertz, 1972; Wilkinson, 1977; Mahdi, 1986; Martin and Yoder, 1986). The private property right to the shared source of water prevails in these systems and water delivery schedules are essentially coordinating institutions between different individual water users. Individual ownership in a collective resource is symbolized by the water share, i.e., the time unit that describes the duration of a person's water turn in the total flow or in part of the flow.
In these systems three parameters describe local water delivery schedules: the duration of the irrigation cycle, the delivery pattern and the method of water distribution. The irrigation cycle is the period over which a water turn returns. Ideally, the duration of a cycle is in line with crop requirements. Irrigation cycles that are too long complicate the cultivation of crops with short irrigation intervals.
The second parameter, the delivery pattern, arranges the sequence in which each water user will receive his turn within the irrigation cycle. Ideally, the movement of water through the command area is systematic, going from head to tail or reverse rather than haphazard. This reduces losses due to dead storage and perimeter wetting. Yet where water users have scattered holdings or exchange water turns a systematic schedule may be difficult to achieve.
The final parameter that describes water delivery schedules is the method of water distribution. Water distribution directly affects discharges at field level. It divides the flow in different proportions. For this reason it is an important determinant of field efficiency as flows at field level may be too large to handle or too small to distribute across a plot. It similarly affects conveyance efficiency, since losses from small flows are proportionally larger. Another element of local water distribution is storage. Large reservoirs act as buffers. They make system discharge independent of micro-variations in inflow. Small buffers facilitate night storage and affect the timing of the supplies. They also concentrate flows, combining day and night discharges and thus help to reduce conveyance losses.
While the irrigation cycle, the delivery pattern and the distribution system are the determinants, water delivery schedules also vary in degree of specificity. In this respect there are two opposite and interchangeable principles, coordination of water deliveries by rule or by management. In the first case the irrigation cycle, flow distribution system and delivery pattern are rigidly described and are followed nearly automatically, each water user knowing his entitlement from the system. In other systems, on the other hand, an irrigation manager is in place and the discretion given to this specialist (or panel of specialists) reduces the weight and specificity of pre-set rules. These different degrees of specificity of water delivery schedules raise a number of intriguing issues. Posner (1973) and Libecap (1989) have argued that, when the value of a resource increases, institutions become more specific and access rules will be narrowly described1. This leads to a paradox: with increased specificity of rules, resource allocation becomes almost mechanistic. There will be less need for a coordinating organization. The system that comes into being is one of minimal transaction costs. However, without an actively coordinating organization, modifications will be more difficult to initiate and sub-optimal allocations are a likely consequence. This is reminiscent of a point raised by Langlois (1986) that institutions should not only be judged by their efficiency in terms of minimal transaction costs, but also by their capacity to change and be changed. A very rigid and specific schedule will not be changed easily; instead of a revision the schedules will be kept acceptable by muddling through, primarily inter-individual adjustments, such as exchanges and transfers of water turns between two users or within a subgroup of users.
1 In this article specificity is defined as the level of detail of resource allocation rules. Management by rules is considered the 'highly specific' extreme of a continuum that starts with active management. Alternatively, one may also argue that within the category of active management there are different poles, with the degree of specificity increasing as the decision-making powers are more narrowly defined.
In comparing water delivery schedules, the grounds on which to judge their performance need to be explained. There are several facets: the degree to which the water delivery schedule facilitates exchanges to high productivity users; the degree to which it optimizes the overall water utilization; and, in detail, the extent to which the water delivery schedule contributes to high field efficiency by making appropriate deliveries available; and the extent to which it minimizes conveyance and seepage losses.
Can the different facets be covered by one delivery schedule? The answer to this is dependent on the type of irrigation infrastructure in place and illustrates the importance of the artefact underneath the institution. In a system with unlined or partly lined irrigation channels, with significant conveyance losses, such as described in the three case studies below, two objectives contradict as flexibility occurs at the cost of conveyance efficiency. The frequent rearrangement of the distribution pattern that follows from commercial transactions will translate in more seepage loss as channels will be filled and emptied frequently, involving the repeated wetting of the channel perimeters.
This situation can be contrasted with a system, designed by Merriam (Davis and Merriam 1990; Merriam, 1990; Merriam, 1993), which consists of a network of semi-closed pipes, that distribute water that is stored in a reservoir. From a valve on their lands, farmers can take water whenever they need. Because of the pipelines, there is no conveyance loss. Similarly, flexibility is served as each farmer can take water when desired. Refreshingly, Merriam postulates that the unconstrained availability of the resource is essential for the efficient use of land and water. This contradicts the economist's wisdom that efficient use comes with scarcity. The weak spot in Merriam's system, however, is field irrigation efficiency, as the unrestrained availability of water does not encourage efficient field water use and, apart from self-restraint, there is no incentive not to overirrigate2. Even though it is difficult to combine optimum flexibility, optimum conveyance and field efficiency in one water delivery schedule, it is still not impossible to identify unambiguous improvements in most water delivery schedules.
2 To overcome this problem, Merriam suggested intensive farmer training in field water management. Even then a second problem needs to be resolved which is the definition of the area under irrigation, as when each farmer can take water at will, conflicts are unavoidable. Surveillance of the designated command area, however, means higher transaction costs and a loss of productivity. Another criticism of Merriam's design is the high physical investment cost involved.
CHANGING WATER DELIVERY SCHEDULES: THREE CASES FROM BALUCHISTAN
The three case studies presented here come from a training given to selected farmers from small irrigation systems in highland Baluchistan. In this semi-arid region water is scarce and extremely precious, as it sustains a farming system of high value horticulture. The high value attached to water was the background to the participatory training in water management. In this training key farmers were asked to analyse their water delivery schedule and identify improvements in dialogue with trainer/engineers (van Steenbergen, 1995). This was preceded by the preparation of maps and an extensive investigation into the rationale of the water delivery schedule in place.
The first case concerns Appozai village in Zhob District, located at the tail end of the Saliaza system. The water is perennial and diverted from the Kapip Lora River at a place 20 km upstream of the village. Though Appozai is located at an altitude that allows high-value horticulture and, moreover, is at a short distance from the market town of Zhob, the command area looks forlorn, with low-value maize and underirrigated almond trees dominating the farming system.
FIGURE 1 - Water delivery system in Appozai
The poor crops reflect the inconvenient water allocation system in force in Saliaza Appozai. There are three major secondary channels in the command area (Figure 1). Between the three channels a time division is in place. The right channel receives water first for four days; next the central channel is served for two days; and finally the full supply goes to the left channel for six days. An additional complication is that in the critical summer season a separate command area located close to the source is entitled to the entire Saliaza system flow for four days every 12 days, resulting in repeated four-day gaps in the. Appozai water supply. Consequently, the entitlements of the different Appozai secondary channels stretch out over 16 or 20-day periods in the summer season. This cycle is too long to enable the cultivation of important cash crops, such as tomatoes, summer vegetables and even apple trees.
A second remarkable feature of the water delivery schedule concerns the distribution of water within the branch channels. In each of the three secondary channels the flow is proportionally divided over a large number of tertiaries, in which the water flows permanently for the full duration of the secondary's turn. The size of the intakes of each tertiary is equivalent to the water rights of the command areas served by them. The base summer flow of the Saliaza channel was 75 l/s. Yet, as this is flow-divided over a large number of tertiary channels, a very large part of the flow is lost in seepage. The inconvenient irrigation cycle and the unnecessary conveyance losses in Appozai demonstrate that local resource management does not necessarily conform to the best pattern, even within the physical constraints of the system.
Several solutions to improve water availability and reduce conveyance losses were discussed with village leaders in Appozai, including the local water master. The options all revolved around reversing the order of time and flow division, namely to introduce flow division at the highest level in the system and using time shares instead of proportions of the flow to divide water over the tertiary channel. This would reduce steady state conveyance losses and make water available in each of the secondary command areas for most of the time. It would facilitate local exchange of water turns, improving flexibility and convenience. Interestingly, the key farmers acknowledged the serious shortcomings of the present pattern and the relevance of the suggested improvements. Still, they were reluctant to correct the system, basically because they considered the costs of convincing fellow water users prohibitive. A number of reasons were given as to why it would be difficult to make innovations: the large number of water users (500), the low dependence of many shareholders on agriculture, since most landowners had ventured successfully into the construction industry, the weakness of traditional leadership and slumbering resentments between the four constituent clans of Appozai village. A final and very important obstacle was the fact that only few shareholders understood the entire system of rights and allocation rules, so that most farmers were wary of any dramatic rearrangement of the water allocation system.
Whereas in Appozai the lack of interest of any group to take the initiative explained the institutional stagnation, in the second case of Uriagai (Lorelai District) small groups of people, benefitting from a suboptimal status quo, blocked improvements in the water allocation system.
Uriagai is dependent on the flow of a vertical well (qanat or kareze) on the right bank of the Kohar Manda River. The water used to cross the bed of the river through an unlined farmer-built conduit. It served two separate command areas, both on the left bank (Figure 2). The first command area was under almond trees and high value vegetables. The entire summer flow used to be devoted to this area and did not in fact suffice to serve it adequately. The second downstream command area only received supplies in the winter season, when water supply was high and demand was low and was dedicated to low-value winter wheat. Almost all farmers owned land in both command areas.
FIGURE 2 - Water delivery system in Uriagai
With the help of public investments the farmer-built conduit was replaced by a concrete syphon, significantly reducing main system conveyance losses. The amount of water available to the two left bank command areas went up by 70%. The available summer base flow increased from 80 to 135 litres per second. This changed the relative scarcity of land and water in the upper command area significantly. The holding of a farmer with the fictitious name of Mohammad Siddiqui can serve as an example: before the improvement to the river crossing his 12-hour water share just about sufficed to irrigate his orchard. In his own assessment he was in fact slightly underirrigating. The increase in water supply changed this and eight hours of the augmented flow now sufficed. It created an excess flow of four hours in the summer. To utilize this, Mohammad Siddiqui would have to take it to the lower command area, since all his land in the upper command area was under cultivation. This, however, made little sense, as most of the water would be lost in wetting the long earthen conveyance channel to this downstream command area. Instead, therefore he continued to utilize the water in the upper command area: overirrigating his own plot and giving away some of the excess flows, that still had land for expansion. Several farmers were in a similar situation to Mohammad Siddiqui. The solution would be to coordinate their excess flows. They could avoid excessive transient losses for instance by splitting the irrigation cycle in a component serving the upper command area and a component serving the lower command area. Resistance to this solution came from landlords with sufficient land in the upper command area. A change in the irrigation cycle would bring them no advantage, but instead would deprive them of the windfall extra water deliveries.
Zum Shah Murad
The village of Zum Shah Murad (Qila Saifullah District) is dependent on a highland horticulture farming system, that is comparable to Uriagai. Its source of irrigation water is a small spring (20 litres per second) in the hills overlooking the settlement. A water cycle of 15 days is applied throughout the year. An extra day is added that is 'sold' by the community to people with very small water shares. The transaction takes place at a price below the commercial value of water, in a gesture of solidarity. Apart from this, collective water share selling and purchasing of water is unusual. There are, however, intense exchanges of (parts of) water turns between farmers to meet crop irrigation intervals. While functional in this respect, the exchanges may also be expected to result in an unsystematic water schedule and concomitant high transient losses. The water schedule is, however, already chaotic without the exchanges, as principally different parts of the command area are served during a single day. This is a deliberate choice, related to the absence of an alternative source of domestic water, in contrast to Uriagai and Appozai, where piped systems were in operation. By scattering water turns throughout the command area in Zum Shah Murad, people living in small clusters in different parts of the command area are able to collect domestic water at a place close to their residence. This underlines an aspect of irrigation systems that is often ignored, but is nevertheless of relevance in water delivery schedules: the trade-off with the additional function of domestic water supply1, for which, as in the case of Zum Shah Murad, irrigation efficiency may be sacrificed.
1 An as yet unresolved issue is the impact of separating domestic and agricultural water supplies for irrigation efficiency. A variation on the chaotic schedule in Zum Shah Murad is to let the major secondary channels 'bleed' with a small flow to enable the collection of surface water. The total volume of water 'lost' in this manner may be significant, which poses the question of whether domestic water supply schemes may already be justified from the point of view of irrigation efficiency.
The example of Zum Shah Murad reminds us of the danger of overemphasizing economic and technical efficiency while ignoring the importance of non-economic objectives in local resource systems, in this case domestic water supply and local solidarity. Another irrigation system, where users took part in the participatory training on water allocation in Baluchistan, illustrates the same point. In Thal (Lorelai District), public investment in the long main channel improved the reliability of summer supplies and increased the available discharge in the command area. The expectation of those responsible for the public investment was that this would result in an expansion of the highly profitable fruit orchards. This required, however, that the traditional practice of block rotation be abolished. Under this practice a different tract of land was cultivated every year and a lottery determined the tract of land that a particular shareholder in the system would cultivate. Obviously, block rotation and land allocation by lottery was incompatible with the cultivation of perennial crops. The water users of Thal hence faced a dilemma, but eventually made a conscious choice not to give up their traditional method of block rotation. The reason was that they feared that permanent individual rights to the land would facilitate transactions, which could set off an invasion by a neighbouring clan and undermine the relative harmony of the community at Thal.
Similarly, in Appozai and Uriagai gains in irrigation efficiency, identified as such by the water users themselves, were foregone. Instead of the requisite adjustments in the water delivery schedules, 'muddling through' prevailed. The reasons behind the lack of significant institutional change in the two systems differed and offer a view on the mechanics of indigenous institutional change, the interest of the different stakeholders in a resource system and their level of organization.
In Appozai the discussed improvements in the water distribution were Pareto-neutral. They would not have harmed anyone, but would have brought benefit to at least those farmers that wanted to adopt high value horticulture. The case of Appozai underlines the importance of the costs of institutional change. The water delivery schedule was highly specific and complex. Though each water user knew when to expect his turn from the system, very few farmers understood the Appozai water delivery schedule in its entirety. This general lack of comprehensive system knowledge may also be called 'tradition. The Appozai water users were the captives of this incomplete knowledge. To change the water delivery schedule, this obstacle had to be overcome, yet no one in Appozai was willing to take the initiative and bear the cost of explaining the present system and the proposed modifications and convincing a multitude of co-users, probably as for no one in Appozai the additional individual benefit derived from the new schedule compensated for the efforts in convincing fellow farmers. Water shares in the system were generally small and the larger shareholders, who stood to gain most, derived most income from outside agriculture.
By contrast in Uriagai there were clear winners and losers. If the irrigation cycle had been adjusted to accommodate summer supplies to the second downstream command area, the persons with ample land in the first command area would have lost their additional supplies. The changes in other words were Pareto-biased and as both parties were equally strong, side-payments would have been necessary to overcome resistance from those that were negatively affected1. Such compensations were however not acceptable. What was at stake was a windfall gain of excess water to farmers with ample land resources upstream, not a hard entitlement that could be negotiated about. In addition to complexity of conflicting interests, the cost of institutional change played a role in Uriagai as well. The water delivery schedule was highly specific and rule-oriented and few farmers understood the full system. Furthermore, there were the many interindividual adjustments and water deliveries which were hard to disentangle. Like Appozai, the water delivery schedule in Uriagai exemplifies the paradox of a specific, but immutable institution.
1 This raises the issue of inequity and institutional change. In case the of polarized control of power one would obviously expect that institutional change would be blocked if the party in command were to loose from the change; and would succeed if the ruling party stood to benefit.
What is evident is that the step from local economic opportunities to local institutional change, as often casually assumed to take place, is not always obvious. Non-economic considerations, conflicting interests and the question of who is going to bear the costs of the change process may stand in the way. In this respect the theory of critical mass in collective action as formulated by Marwell and Oliver (1993) is illuminating. Marwell and Oliver have postulated that one-time collective action (as required in bringing about change) is more likely to take place in the case of heterogenity. In the case of changing a water delivery schedule, this could be translated as the presence of a group of individuals with more than average interest in the improved schedule and more than average resources to bring about the change. This group would be willing to bear the initial costs of effecting the change, thus lowering the threshold for the others to follow. The consequence of this micro-social theory is that changes that bring a small benefit to all people may not necessarily materialize in contrast to changes that bring significant benefit to few.
It is here that the degree of organization comes in. Where individual users are already organized, the cost of institutional change will be lower and it will be easier to capture the marginal overall gain, or even, in the case of Pareto-biases, to negotiate side-payments. The problem with highly specific institutions is that precisely such an actively coordinating organization is missing. Water users become captives of their own schedule. In such cases it may not be wise to rely on user-driven indigenous change, but to consider outside intervention, though while respecting the sovereignty of the local users, for instance, by participatory training, awareness creation and reinforcing local organization.
Davis, G.G. and Merriam, J.L. 1990. Sri Lanka upgrades social, agricultural and economic conditions with a demand irrigation schedule. Paper at 14th International Congress ICID, Rio de Janeiro.
Geertz, C.T. 1972. The wet and the dry: traditional irrigation in Bali and Morrocco. Human Ecology 1 (1): 23-39.
Korten, D.C. 1990. Voluntary organisations and the challenge of sustainable development. Australian Development Studies Network Paper 15. Canberra: ANU.
Langlois, R.N. 1986. The new institutional economics: an introductory essay. In: Economics as a Process: Essays in the New Institutional Economics. R.N. Langlois (ed), Cambridge: Cambridge University Press. pp. 1-25.
Libecap, G. D. 1989. Contracting for Property Rights. Cambridge: Cambridge University Press.
Mahdi, M. 1986. Private rights and collective management in a High Atlas Berber tribe. In: BOSTID Proceedings of the Conference on Common Property Resource Management. Washington DC: National Academy Press. pp 181-197.
Martin, E.D. and Yoder, R. 1986. Institutions for irrigation management in farmer-managed systems: examples from the Hills of Nepal. IIMI Research Paper 5, Digana: IIMI.
Marwell, G. and Oliver, P. 1993. The Critical Mass in Collective Action: A Micro-social Theory. Cambridge: Cambridge University Press.
Merriam, J.L. 1990. Gadigaltar tank irrigation pilot project. Khargone, Madhya Pradesh. Mimeo.
Merriam, J.L. 1983. Flexible water supply system introduction facilitated by the use of pilot projects. In: ICID, Transactions of Fifteenth Congress, 6-11 September 1993. New Delhi: ICID. pp. 617-630.
North, D.C. 1990. Institutions, Institutional Change and Economic Performance. Cambridge University Press.
Posner, R. 1973. An economic approach to legal procedure and judicial administration. Journal of Legal Studies 3: 257-286.
Pretty, J. and Guyt, I. 1992. Primary environmental care: an alternative for development assistance. Environmental and Urbanization 4 (1): 22-36.
Reidinger, R.B. 1974. Institutional rationing of canal water in Northern India: conflict between traditional patterns and modem needs. Economic Development and Cultural Change 23 (1): 79-104.
Rosengrant, M.W. and Binswanger, H.P. 1992. Markets in tradable water rights: potential for efficiency gains in developed-country irrigation. Paper at World Bank Ninth Annual Irrigation and Drainage Seminar at Annapolis, December 8-10 1993.
Shah, F., Zilberman, D. and Chakravorty, U. 1993. Water rights doctrines and technology adoption. In: The Economics of Rural Organization: Theory, Practice, and Policy. K. Hoff, A. Braverman and J.E. Stiglitz (eds.). New York: Oxford University Press. pp. 478-499.
van Steenbergen, F. 1995. Participatory farmers in water allocation in Balochistan (Pakistan). FMIS Newsletter, September 1995. pp. 30-32.
Wilkinson, J.C. 1977. Water and Tribal Settlement in Southeast Arabia: A Study of the Aflej of Oman. Oxford: Clarendon Press.