Community Management of groundwater

Empirical Findings From Canal Irrigation

The design principles are best illustrated by two irrigation studies that included multiple farmer managed and government managed systems. Because irrigation systems are common-pool resources, irrigators experience dilemmas in the operation and maintenance of such systems. In many instances, water is insufficient to meet all irrigators’ needs all of the time. Water allocation rules must be established to share water across irrigators. As water becomes increasingly scarce, and as water demands are not satisfied, irrigators face increasing temptation to violate the rules. Furthermore, irrigation structures must be built and maintained. Rules must be established governing irrigators’ contributions. Again, irrigators face temptations to avoid their contributions because it is difficult to exclude irrigators from enjoying the benefits of a system, even if they did not contribute to it (Tang 1992, 1994). Irrigators face numerous challenges in creating and maintaining a well-functioning irrigation system.

Tang (1989, 1992, 1994) studied 47 irrigation systems – 29 farmer-owned and governed systems and 18 government owned systems. Among the government owned systems, 14 were governed by national or regional governments and four by local governments. Three performance measures captured the extent to which the governing systems mitigated the multiple dilemmas experienced by irrigators – system maintenance, adequacy of water supply, and rule-following behavior by irrigators. System maintenance is directly tied to how well farmers’ contributions to construction and maintenance are elicited and coordinated. It measures the extent to which provision dilemmas have been mitigated. The adequacy of water supply depends on the rules used to allocate water, and the level of maintenance. System maintenance and water allocation depends on well-crafted rules and the extent to which irrigators follow those rules.

In comparing the performance of farmer-owned systems to government-owned systems, Tang divided the cases into high performing and low performing systems. High performing cases were ranked positively on both rule conformance and maintenance, whereas low performing cases were ranked negatively on rule conformance, or maintenance, or both. Among the 14 government owned systems, 6 performed well and 8 did not. (Because the number of cases governed by local governments was so small, Tang excluded them from the analysis). Among the 29 farmer owned irrigation systems, only 25 had sufficient information concerning the three performance measures. Of those 25, 18 performed well and 7 did not. Thus, 43% of the government owned systems performed well whereas 72% of the irrigator owned systems performed well.

What accounts for the differences between high performing and low performing irrigation systems? Tang (1994) argues that among irrigation systems that perform well, rules that govern water allocation and maintenance activities are better crafted to the specific conditions of each irrigation system. Among those that perform poorly, rules do not match the particular situations, or no rules whatsoever were adopted. Since the most complete set of data relates to water allocation, only that data will be examined here.

Water allocation is affected by the rules governing access and distribution. Among the cases examined, four different types of access rules were used -- ownership of land within a specified location; ownership of shares, transferable independently of land, to a certain proportion of the water flow or water delivery facilities; membership in an organization; and payment of an entry fee each time before appropriating water (Tang 1994:231). Among all 14 government owned irrigation systems only one rule of access was used – ownership of land. Among the 25 farmer owned irrigation systems, 9 used the single rule of land ownership and 16 used a combination of the 4 different types of access rules. In terms of performance, 14 of the 15 low performing irrigation systems used just the single rule of land ownership, whereas 9 of the 24 high performing irrigation systems used just the single rule. In other words, in terms of access rules, low performing systems look a lot alike, no matter if they are farmer or government owned, they fail to adequately limit access to the available water supplies. As Tang (1994) argues, the requirement of land ownership is a blunt means of limiting access. Unless land is carefully matched to the capacity of the water source for the irrigation system, water scarcity and allocation problems may become acute (design principle one).

A very similar story can be told about allocation, or distribution rules. Among the cases, three different types of rules were used: dividing the flow of water into fixed proportions by a physical devise; assigning each individual a fixed time slot during which he can take water; and each individual takes turns in appropriating water. Each of the rules may be based on different premises. For instance, a fixed time slot may be defined by the location of an irrigator’s field, whereas a proportion of water may be based on the number of shares of water owned. Among the government owned systems for which there was information about allocation rules, 10 used a single rule – fixed time slot, whereas 2 used a combination of rules. Among the 25 farmer owned systems, 3 used no rules, 9 used the single rule of a fixed time slot, and 13 used combinations of rules. Again the low performing systems were similar, regardless of who owned them. Of the 14 poor performing systems, only 2 used more than one allocation rule. While fixed time slot rules are relatively easy to define and enforce (presumably irrigators will protect their time slots or risk losing their water), they are fraught with uncertainty. A fixed time slot only guarantees a farmer a certain amount of time, not a certain amount of water. If the fixed time slots are not carefully coordinated with water availability, or if water availability is unpredictable, a fixed time slot will have little value (design principle two).

Tang (1994) also noted differences concerning monitoring and enforcement arrangements. Government owned systems were more likely to have formal monitoring systems in place. Of the 15 government owned systems, 12 had guards, 9 of the 12 were government employees and were paid. Among the 25 farmer owned systems, 17 had guards who were also local farmers and 12 had no guards. The farmer guards were much less likely to be paid, and if they received compensation, it typically took the form of reduced obligations. Farmer guards, however, were much more likely to impose sanctions on rule breakers than were government guards and rule following behavior was much higher in farmer owned systems. Furthermore, rule following behavior was much higher in farmer owned systems without guards than with government owned systems without guards. Having formal monitors present is not enough, monitors, to be effective, must also enforce (design principles four and five).

While low performing irrigation systems are very similar in the rules of access and allocation that are used; such systems are much more likely to be government owned rather than farmer owned. Government owned irrigation systems tend to perform poorly relative to farmer owned irrigation systems in part because a single simple rule set is applied to every situation, whether or not those rules are appropriate or likely to be effective. Tang (1994) suggests two reasons for government irrigation officials to consistently adopt land ownership and fixed time slots as the access and allocation rules for governing irrigation systems. First, government officials face powerful incentives to include as much land as is possible in an irrigation system, with little thought given to the capacity of the water source. The benefit-cost ratio of the projects are more attractive, and more potential recipients means more political support (Ostrom 1992; Tang 1994: 232). Second, a fixed time slot rule for allocating water lightens the burden of administration. Not only is such a rule, if meaningful, self-enforcing, but it requires little engineering. It does not require installing, operating, and maintaining water allocation devices, for instance.

Farmers face a very different set of incentives than do government officials. Since they directly experience the consequences of their rule-making decisions, they confront incentives to carefully craft the rules to the particular situation that they face. Not only might this account for the varied rule systems found among farmer-owned irrigation systems, but it may account for the better performance of such systems.

Like Tang (1992, 1994), Lam (1998), in a study of over 100 irrigation systems in Nepal, found that farmer managed irrigation systems performed significantly better than did government managed irrigation systems. Irrigators in farmer managed systems exhibited significantly higher levels of 1) entrepreneurial activities in attempting to coordinate irrigation activities, 2) information and understanding of the irrigation system, and 3) mutual trust (Lam 1998: 126-133). Irrigators in farmer managed systems also use more varied and complex sets of rules for governing their activities. Also, irrigators in farmer managed systems are significantly more likely to be monitored and sanctioned for rule violations than are irrigators in government managed systems. Rule-following behavior is significantly higher in farmer managed systems (Lam 1998: 131).

Most of Lam’s findings support the theory of common pool resources, particularly the conditions supporting robust institutional arrangements. Similar to Tang (1994), Lam found that irrigators paid close attention to boundaries and to exclusion in farmer-managed irrigation systems, attempting to more nearly match water supply with demand (design principle 1). Furthermore, irrigators in farmer managed systems devised more rich and complex sets of rules to govern access, water allocation, and contributions to maintenance (design principle 2). Irrigators in farmer managed systems have a better understanding of their systems, and are more likely to engage in attempts to revise the rules (design principle 3). Irrigators in farmer managed irrigation systems have devised active monitoring systems, and those irrigators are more likely to be sanctioned if caught violating the rules (design principles 4 and 5).

Evidence from groundwater irrigation is suggestive, but to my knowledge, no systematic comparative institutional studies of different forms of well governing arrangements have been conducted. Shah (1993) cites to a study conducted by Lowdermilk et al. (1978) in Pakistan of crop yields under different levels of control of water sources. Among groundwater users, crop yields were highest among farmers who owned their own wells and lowest among farmers who depended on public tubewells (Shah 1993:29). Shah (1993:29) states that a number of studies have been conducted in India that suggest that farmers prefer water from privately owned tubewells over publicly owned tubewells. This is so, Shah (1993:29) argues, because water service from state tubewells is inferior to that of private tubewells. State tubewells suffer from poor maintenance, long shutdown periods, erratic power supplies, and so forth. The root of the problem lies in management. “A state tubewell operator is in reality accountable to no one, for he can neither be punished nor rewarded by the community he is meant to serve…” (Shah 1993:30).

A case study by Singh (1991) of the construction, operation, and maintenance of a public tubewell used for irrigation in Uttar Pradesh, India, clearly illustrates Shah’s arguments. The well and its associated infrastructure was designed and built by the government irrigation department. The department is to operate and maintain the well. Water allocation and distribution was turned over to farmers’ committees formed by the government irrigation department. The well and its infrastructure are not well matched to the patterns of landownership. According to Singh (1991), government officials face few incentives to operate and maintain the well appropriately, water service is erratic, and the farmers’ organizations have slowly fallen apart.²

The evidence from studies of well ownership and operation appears to coincide with the evidence from canal irrigation systems. Government operated canal systems and wells perform poorly relative to farmer operated canal systems and wells. What is not well understood in relation to wells, and consequently needs more study, is the relative performance of different types of farmer based ownership and management structures. For instance, Dubash (2000), in a case study of a groundwater market in a village in Gujarat, India, carefully details the roles that different types of well ownership groups play in setting village wide groundwater prices. Dubash (2000) notes that it is not the individual well owners who set prices, rather prices emerge from the collective decision making processes of a rather large well partnership group. The author speculates that the large well partnership sets the prices because it is viewed as legitimate. In order to set an acceptable price, the well partners must first sort out their own differences. Those who engage in farming as their primary source of income want water prices for partners to be set low, whereas, partners who engage in little agriculture want prices set high to increase the earnings of their well shares.

Substantive and Theoretical Issues

While considerable theoretical and empirical work has been undertaken over the past two decades on local level governance of irrigation systems, there remains considerable work left to do. Substantively a major issue among studies of irrigation systems is the effects of heterogeneity on the performance of farmer governed systems (Tang 1992, Lam 1998, Bardhan and Dayton-Johnson 2002, Ruttan 2004). Bardhan and Dayton-Johnson (2002) reviewed the findings of several large-n studies of farmer managed irrigation systems that devoted attention to heterogeneity. Across all of the studies, the effects of heterogeneity were consistently negative (Bardhan and Dayton-Johnson 2002:104-105). Income inequality and asymmetries between head-enders and tail-enders was associated with rule breaking, poor system maintenance, and poor water delivery performance. Landholding inequalities were associated with poor canal maintenance. Differential earning opportunities among irrigators were associated with lower rule conformance and system maintenance.

Ruttan (2004) carefully re-analyzed the data collected by Tang (1989, 1992) to explicitly examine the effects of different forms of heterogeneity on the performance of irrigation systems. In addition to the findings reported by Bardhan and Dayton-Johnson (2002), she found that variation in income had a negative effect on the likelihood that sanctions for rule breaking would be applied (Ruttan 2004:28). Ruttan (2004:35) also found that socio-cultural heterogeneity had a negative effect on rule conformance and system maintenance. These studies suggest that as community based approaches to irrigation management become more widely embraced and adopted, careful attention must be paid to the effects of different forms of heterogeneity. What “careful attention” should consist of is not entirely clear and probably requires more study. In general, explicit and direct redistribution programs, such as land reform, have met with very mixed results. What are needed are careful studies of the mechanisms that local communities have devised and adopted to limit or dampen the effects of heterogeneity.

As Agrawal (2002) has so forcefully argued, considerable theoretical work remains to be done concerning the conditions under which self-governing arrangements are likely to emerge and persist. Agrawal (2002) argues that much of what has passed for theory consists of a listing of factors that affect cooperation gleaned from a single for a few case studies. According to Agrawal (2002), scholars should stop developing lists of factors that they believe affect self-governance and instead develop and conduct carefully designed research projects that will allow for the identification and empirical testing of relationships among the various factors.

The Groundwater Setting

As is apparent from the previous section, much of the empirical work on local, self-governing institutional arrangements has centered on canal irrigation systems. Less attention has been devoted to groundwater irrigation, per se, although a growing body of work focuses on groundwater use and governance in the western U.S. (Ostrom 1965; Blomquist 1992; Blomquist, Schlager, and Heikkila 2004). Nevertheless, the emerging theory of common pool resource governance provides a consistent set of concepts and analytical tools to diagnose problems, provide a deeper understanding of the conditions under which local governance of groundwater resources is likely to occur, identify promising policy alternatives, and shed light on the shape and form of productive relations between local level governance arrangements and regional and national governments.

The governance challenges groundwater irrigators commonly face differ considerably from those faced by canal irrigators. The differences result from the distinct physical structures of canal irrigation systems compared with groundwater basins and, consequently, the different water development paths that unfold between the two types of water systems.

To construct and operate a canal irrigation system requires considerable upfront production and transactions costs. Using Ostrom, Gardner, and Walker (1994) terminology, irrigators immediately confront provision problems. At a minimum production costs entail building a diversion structure, a distribution system, and field outlets and channels. A single person or family cannot meet such production costs; rather a collective effort is necessary, involving many people, their resources, and their participation. The transaction costs of organizing people, developing information about the physical setting, negotiating over the location and design of the irrigation system, organizing labor, and monitoring and enforcing agreements concerning contributions and work are significant. Furthermore, considerable effort and attention must be paid to developing governing arrangements to manage the irrigation system once it is operating. Developing and modifying water schedules; devising, adopting, and modifying water allocation rules; monitoring and enforcing rules; and organizing maintenance activities are all ongoing transactions costs that must be met to realize the benefits of the irrigation system. For canal irrigation systems, devising a governing system well matched to the physical system, or vice versa, is critical, it determines the quality of service. In general, then, before an irrigator is likely to enjoy water flowing into his fields, he and his fellow irrigators collectively must make significant upfront investments in production and governance.

The water development path is considerably different in relation to groundwater. Farmers using groundwater are much more likely to face appropriation problems initially and provision problems later, however, both types of problems are only likely to emerge after farmers have made substantial personal investments in wells and in productive activities that depend on the well water. The governance challenges that well irrigators face are substantially different than those that canal irrigators face.

One of the most striking aspects of groundwater development is how rapidly it unfolds once a minimum level of technology and energy become widely available. Entry to groundwater basins is minimally restricted, with land ownership or leasing the only requirement for access. Depending on the setting, such as water table levels, even relatively poor farmers may access groundwater through inexpensive technologies. Even if farmers do not invest in their own wells, either because they do not have the necessary capital or their landholdings are too fragmented to justify a well, they may gain access to groundwater through markets (Shah 1993).

Groundwater is widely adopted because of its high value. For some farmers it may be the only source of irrigation water, either because they do not have access to canal irrigation, or even if they are within the command area of a canal system they may not receive water. For many farmers, groundwater is more reliable, timely, and adequate than the water they receive from canal systems. For other farmers, groundwater may be more “convenient” than canal water, even if canal water is reliable and timely. A farmer who owns a well that provides enough water for his irrigation needs may opt out of a communal system and its various requirements and responsibilities, such as contributing labor and materials for canal maintenance.

Compared to canal irrigation, developing groundwater entails substantially lower upfront production and transaction costs. Nature has provided a reservoir that at least initially, and in many cases, is very easily accessed, through a well. Consequently, production costs may be born by a single individual or family. Transaction costs are also low. Farmers need not organize bargain and negotiate over the development of an irrigation system and system design, or monitor and enforce commitments. Some farmers may form partnerships to raise the capital necessary to build a well, however, the transaction costs they face are substantially lower than those faced by farmers attempting to develop and build a surface irrigation system.

The types of governance arrangements that farmers are more likely to create during the early stages of groundwater development are those that center on providing farmers access to groundwater, as discussed at the end of the last section, than they are to govern groundwater basins as common pool resources. Gaining access is of immediate concern, commons dilemmas are likely to arise later.

Appropriation Problems

Appropriation problems are highly local. They stem from actions and choices of appropriators.³ Assignment problems, for instance, occur because people compete to use the most productive patches of a common pool resource and in the process they interfere with one another’s harvesting activities. People may place wells to closely together reducing the productive capacity of each of the wells. Or, technological externalities occur because the different harvesting techniques that people use interfere with one another. A high capacity well may create a cone of depression that dries up surrounding shallow tubewells.

Effectively and equitably addressing assignment problems and technological externalities requires considerable time and place information. Working knowledge of the types of technologies used, the location of wells, the uses made of the water, landholding patterns, the actions causing the harvesting conflicts, and so forth are necessary if rules that match a specific setting are to be devised. Such local knowledge resides with water users and not regulators. Shah (1993:129-132) notes the numerous difficulties regulators external to local communities have in devising effective rules. A common approach to address assignment problems is to impose well-spacing rules. The rules only apply to more modern technologies, such as electric and diesel pumps, thus failing to afford any protection for more traditional technologies. Also, well spacing rules are enforced through banks that will not provide capital for the purchase of pumps unless well spacing rules are followed. Farmers who can raise sufficient capital without relying on a bank can avoid well spacing rules.

Groundwater users can determine the causes and effects of spacing wells too close together or of allowing high capacity wells to be situated among traditional water lifting devices. Well owners and others who are dependent on those wells for water face incentives to problem solve in order to protect their water sources. Depending on the social ties among groundwater users and experiences that they have had in engaging in other collective efforts, they may pursue strategies or undertake collective efforts to address assignment problems and technological externalities. For instance, Shah (1993) cites several examples of groundwater users effectively addressing such problems among themselves.

The owners of grape orchards in Karnataka and Andhra Pradesh, for instance, are known to buy up neighbouring lots at premium prices to solve the problem of interference … in many parts of Gujarat, where localized water markets have assumed highly sophisticated forms, it is common for a well owner to lay underground pipelines through neighbours’ fields at his own cost, and dissuade them from establishing their own wems by informal long-term contracts for the supply of water at mutually agreed prices”(Shah 1993:7).
Appropriation externalities, while closely related, are conceptually distinct from provision problems (Ostrom, Gardner, Walker 1994). Appropriation externalities result from over use of common pool resources in the short term. Provision problems may result from repeated over use, producing a long term decline in the productive capacity of the resource; however, provision problems emerge from other sources as well, such as lack of maintenance that affects the productive capacity of the resource. Findings from studies of common pool resources such as fisheries, suggest that resource users find appropriation externalities to be more challenging to address than are assignment problems and technological externalities. In the case of fisheries, fish populations fluctuate unpredictably and fishermen find it difficult to relate their harvesting activities with fish abundance or scarcity (Schlager 1990, 1994). The “noise” of fish population dynamics drowns out the effects of harvesting on fish stocks. While local fishing communities do a relatively good job of addressing assignment problems and technological externalities, they rarely attempt to directly address production externalities (Schlager 1994).

Groundwater users may find appropriation externalities less challenging to address than fishermen because the interaction between pumping and water tables is more direct and observable than is the interaction between fishing and fish populations. Appropriation externalities in groundwater may often be spatially and temporally confined, allowing closely situated groundwater users to learn about the effects of pumping on water tables and on one another’s pumping activities. That learning can form the basis for developing locally devised solutions to appropriation externalities. For instance, Sadeque (2000) examines the development of water access and allocation rules to address appropriation externalities that emerge during the dry season in Bangledesh. Domestic water uses are satisfied through shallow hand pumps. During the dry season, when groundwater demand is quite high, especially to irrigate the winter rice crop, the hand pumps dry up, leaving many households without a reliable and convenient source of water. As Sadeque (2000: 277) notes, “In the competition for groundwater, simple, low-cost technologies like hand tubewells, used mostly for drinking and other domestic users, lose out. The perception of affected people of the low water table areas as victims of water deprivation is becoming marked, with acrimony towards irrigation.” In a study of two villages in northwestern Bangladesh, Sadeque (2000) found conflict between domestic users and irrigators to be widespread during the dry season. However, the author also found instances of cooperation and coordination emerging to address such conflicts. For instance, a series of shallow wells installed by an international non-profit development agency for domestic water uses, are carefully governed by the households who participated in their development and who are responsible for their maintenance. During the dry season, the households impose restrictions on water use to tide families over. These restrictions also affect households who did not participate in the well project. While during the wet season non-participating families are not restricted in their access to the wells, during the dry season, their access and use is strictly limited. They are allowed water after the households who govern the wells have their needs met. In addition, cooperation is emerging between villagers and owners of irrigation wells. Irrigation well owners allow villagers to take water from wells to meet basic consumption and cooking needs. Also, some well owners operate wells during early morning hours for the express use of villagers domestic water needs. Sadeque (2000:286) argues that such cooperation has emerged as a means of avoiding government regulation. “People realized that negotiation was better than having controls imposed by central and distant authorities which might not be in the interest of either party. Additionally, regulations would result in bureaucratic control and therefore encourage corruption.”

In general, appropriation problems tend to be local in nature. Furthermore, the specific types of problems that emerge and their causes tend to be highly dependent on configurations of factors unique to each situation. Consequently, workable solutions are usually those grounded in specific time and place information, information that is readily available to groundwater users, but not to regulators. In addition, groundwater users often face incentives to invest in collaborative attempts to resolve such problems. Coordination may yield substantial benefits. Thus, it is not uncommon in the emerging literature on groundwater and irrigation to find instances of groundwater users addressing appropriation problems or having the capacity to address such problems. For instance, Shah (1993) describes a village in Junagadh District, Gujarat, in which numerous irrigation wells dry up during the dry season. Shah (1993) notes that farmers have a good understanding of how their wells function, and they pursue a variety of strategies to ensure water throughout the dry period, but with mixed success. Some farmers are more innovative than others and appear to have developed approaches that are relatively successful. Shah (1993:164-165) argues that with a little assistance, primarily in the form of information, such as location and productivity of wells over time, various successful strategies that some farmers pursue, and so forth, that the farmers could develop collective strategies to address appropriation problems and thereby increase agriculture productivity.