2.1 Organizational structure at the project level
2.2 Organizational structure at the small group level
2.3 Organization at the government level
The success of an organization depends partly on its structure (the way in which tasks and responsibilities are formally allocated among its members) and partly on its management process (the way in which decisions are taken within the existing structure). This chapter and the next are concerned with questions of organizational structure. The management process will be considered in Chapter 4.
In the planning of an irrigation organization the most important questions about structure arise at the project level itself and also - particularly in small-farmer agriculture - at the level of the small local group.1 In each case, the choice of an appropriate structure will depend on a large number of factors specific to the particular objectives and context of the organization concerned. In other words, there is no single "ideal" structure which is best for all organizations in all circumstances. The principal factor affecting choice of structure at the project level will be discussed first and then those affecting choice at the small local group level.
1 The expression "small local group" is used here to designate those groups of farmers (informal associations) which are formed on the basis of the watercourses and are characteristic of many irrigation projects in the Far East and Asia.
2.1.1 The organization of water management activities
2.1.2 The distribution of responsibility
Organizational structure has a horizontal and a vertical dimension. The horizontal dimension is concerned with the way in which the various activities essential to the achievement of an organization's objectives can best be differentiated (in accordance with the specialist skills required for each activity) and then coordinated in order to produce the necessary unity of effort among the resulting specialist units. The vertical dimension is concerned with the way in which responsibilities are distributed among members working at different levels of the organization and its component units, from the project to the small group level. We will deal first with horizontal dimension and later with the vertical dimension.
i. Large projects: alternative forms and functions
Two of the most "important factors affecting the choice of an appropriate form of horizontal organization at the project level are the size of the project area and its level of economic development. Other important factors include the objectives of government and the character of existing institutions in the area concerned.
Large projects (usually greater than 200 ha and/or more than 500 farm units) can be defined as projects which require a substantial professional staff to manage them. A survey of existing large project organizations would show that all of them concern themselves with certain central water management activities. These activities include water distribution and system maintenance (always); assessment and collection of water charges or similar fees and taxes (nearly always); and assistance and extension to farmers on water management at the local level (sometimes). Certain general management services - finance, personnel, planning and monitoring - are also provided.
However, many irrigation organizations also embrace other sets of activities: agricultural services (agricultural extension, possibly with research support); commercial services (input supplies, credit, marketing); and basic infrastructure and social services (housing, roads, schools, health services, etc.).
The case for including commercial services within the project organization is usually very strong in new settlement schemes. It is also very important that, on new projects in already settled areas, the project management should ensure satisfactory provision of commercial services. If other existing agencies in the area are poorly developed, the project should be invested with the responsibility for these services, at least initially until the other agencies acquire greater strength. The provision of adequate marketing facilities is likely to be particularly needed in many new projects, because the introduction of irrigation will create a potential for cultivating crops which have previously been unknown in the area; if farmers lack reliable marketing outlets for these crops they will not be prepared to take the risks of adopting them.
In certain types of project (usually settlement projects) the government may have to pay particular attention to making commercial services an integral, indeed central, part of the project organization. The key to these "integrated management" schemes is the project management's ability to control a single-channel marketing outlet. Such an organizational pattern is usually feasible only when the project has been established for the primary purpose of producing high-value cash crops, often for export. The farmers' obligation to sell their produce through the project organization makes it possible for the management to deduct a sufficient amount from the farmers' final payment to cover the cost of the other services it provides - water distribution, agricultural extension, credit, input supplies, in some cases even mechanized land preparation - and management's direct control over these activities can be used to ensure greater timeliness of throughput as well as quantity and quality of production.
Where conditions for an "integrated management" approach do not exist, i.e. in areas where farmers are allowed a relatively free choice as to the crops they cultivate, the reasons for wishing to include commercial services as one of the continuing activities of project management are much less compelling. The establishment of single-channel control is extremely difficult in conditions of multichoice cropping, especially if food crops predominate. Over time, as farmers become increasingly familiar with irrigated agriculture and alternative commercial service institutions develop in response to the new production patterns, the need for the project organization to concern itself directly with commercial services will decline.. This does not mean that it can afford to abandon its responsibilities in this area altogether. To give one example, it is an important function of irrigation management at all stages of small-farmer development to ensure the timely supply of other inputs besides water. However, once other specialist agencies are in a position to take direct responsibility for such services, the best results may be achieved through good coordination between them and the project organization.
The inclusion of responsibility for basic infrastructure, social services and commercial services within the irrigation organization (in addition to water management and agricultural services) is usually necessary where the area to be irrigated has hitherto been little developed and there are no other institutions in the area capable of performing these tasks satisfactorily.
The project organization must almost inevitably take direct responsibility for providing the necessary infrastructure and social services on new settlement schemes. In the case of projects whose purpose is to supply irrigation water to already settled areas, there will usually be less need for this. There may well be a need for the project management to participate in decisions concerning infrastructural development (particularly roads) but it should often be possible for direct responsibility for planning and implementation in these fields to be left to other agencies which are already in existence.
This raises the question as to when water management and agricultural extension activities should be combined within one organization and when a specialized organization should be established for water management activities alone, with agricultural extension and all other service activities being performed by separate agencies.
Again, this is partly a question of the level of economic development of the area concerned and of the farmers' experience with irrigated agriculture. Where the farmers' level of education and technical knowledge is low, there are usually considerable advantages in combining agricultural extension and water management activities within the same project organization. There are additional arguments in favour of a combined approach in the case of countries where there has traditionally been a very sharp division of responsibilities between the Department of Irrigation and Public Works, usually headed by civil engineers, and the Department of Agriculture. Very often this divided pattern of organization has had most unsatisfactory results. One of the principal reasons is that the civil engineers are experts in design and construction work, but have no specialist training in irrigation management, which requires a detailed understanding of agricultural processes and the farming community. Moreover, the agriculturalists, whose department is often poorly endowed with resources and has to cover both rainfed and irrigated areas, are often very short of specialist expertise in irrigated agriculture, especially water management extension.
For these and other reasons, several governments of countries with previous experience of a divided administration are showing an increasing tendency to favour an integrated project authority with direct responsibility for coordinating both water management and agricultural activities. Countries in which this unified approach has recently prevailed include India (The Command Area Development Programme), the Philippines (Upper Pampanga and other major irrigation projects), Malaysia (Muda and Kemsuku projects), the Irrigation Districts of Mexico, El Salvador and others.
The need for a combined organization is much less strong where farmers already have a long experience of irrigated agriculture, the agricultural extension service is of a high calibre, and the officials in charge of water distribution are specialists in irrigation and are well-versed in agricultural matters. Here there are often advantages in having one organization exclusively concerned with water management activities. The degree of coordination within the project organization is reduced and one of the possible risks of bringing agriculture into a command area structure - the comparative neglect of rainfed areas - is reduced.
Examples of successful organizations of this type are the Irrigation Associations of Taiwan Province (Republic of China) where responsibility for agricultural extension is vested in Farmers' Associations, which also have credit, input supply and marketing functions, the Irrigation Associations of Spain, and the "Asociaciones de Canaleros", Chile. Where specialized water management organizations are adopted, close coordination with agricultural agencies is clearly essential. It is also desirable that the offices from which they operate at the local level should be at the same place and that, where farmer groups are being employed as a point of contact, they should use the same groups.
ii. Medium-sized and small projects
Medium-sized projects may be defined as those which are too small to allow consideration of unified or multipurpose project agencies (owing to their relatively high administrative costs) but which nevertheless require a substantial input of professional staff to operate and maintain them. In general, a specialized water management organization is likely to be most feasible on medium-sized projects. However, there may be other possibilities in areas where there are numerous medium-sized or relatively small, discontinuous irrigated commands interspaced with rainfed land. Here the choice appears to lie between a specialized water management agency strongly linked to agriculture through a coordinating committee and the more radical alternative of an area-based pattern of administration in which the units of management would be defined not on the basis of irrigation commands but on complete watershed and catchment areas.
Small projects are definable as those in which official staff are either not required or not available to operate and maintain the main water delivery system, except perhaps for the headworks. In many parts of the world, particularly in hill-stream areas, small indigenous systems can be found, i.e. ones constructed by the farmers themselves. These continue to be communally managed by the farmers with little or no external government assistance. They are typically in the 50-200 ha size range, with between 40 and 100 members and depend on simple technology (e.g. simple diversion weirs of boulders or brushwood). Their internal organization is often impressive, the most remarkable being the Balinese subaks, which have memberships of up to 700 and very complex rules and methods of water distribution. In indigenous systems the farmers themselves are entirely responsible for the activities of water distribution, repairs and maintenance and financing, though they are of course dependent on outside support for extension advice and input supplies. Group members are often willing to accept stringent discipline from their selected leaders, who are accountable to them.
Despite their often excellent organizational features, there is usually considerable scope for improving the performance of most indigenous schemes at relatively little cost, not only through better agricultural extension but through advice and assistance on various aspects of system design. For example, major benefits can frequently be obtained from installing permanent diversion structures in place of traditional temporary weirs, particularly in areas of heavy seasonal rainfall where valuable farming time tends to be lost at the end of each rainy season on repair and reconstruction work. However, intervention by government in the affairs of indigenous organizations should always be done with caution to avoid disrupting the relationships from which the traditional groups have derived their organizational strength. Unfortunately, it has been a common experience that when governments have attempted to influence the development of indigenous systems, the old groupings have tended to disintegrate, because they lose many of their old responsibilities and are given no new ones in their place.
Small schemes which have been constructed by government usually lack the strong internal organization which is characteristic of the indigenous systems and often perform very poorly as a result. The main reason appears to be that after construction, operational responsibility is handed over almost immediately to the farmers without sufficient time being devoted to technical or management training. As in the case of indigenous systems, service support is often inadequate; when new small schemes are to be constructed it is essential that the planners should obtain a good prior understanding of the capabilities and needs of the farmers in the areas concerned, and in the initial stages of implementation emphasis should be given to group formation and training.1 Thought also needs to be given to the most effective way of organizing supporting services for small irrigation systems of all kinds.
1 Most of the points made with regard to the formation of groups within larger projects also apply here (see section 2.2).
It appears to be generally true that outside the larger command areas, resources for the development and support of irrigated agriculture - both financial and personnel - tend to be spread particularly thinly. An obvious explanation for this is that it is more difficult to organize services to a large number of scattered small irrigation schemes than to a single large one. Nevertheless, it seems probable that in many cases very high returns could be obtained from improving administrative support to these small schemes. An increase in the number of staff may often be required for this purpose and questions of organizational structure will also need to be considered. In many cases, the provision of specialist advice on operation and maintenance may best be provided through a rural engineering service, with junior field staff based at local offices of Farmers' Service Centres offering a range of other services (agricultural extension, credit, inputs, etc.). They would need to be in regular contact with senior engineers for technical advice and support. Much of their work should be concerned with low-cost improvements in design and with the development of improved water distribution procedures. It is extremely important in the latter case that a clear division of responsibility be established, with the farmers' organization being encouraged to undertake all functions except those that require specialist knowledge.
The points made here apply equally to small surface schemes and to those groundwater schemes which are not integrated into a large surface irrigation system. In the case of groundwater development, the choice of technology is particularly important with regard to the question of appropriate group size. It has often been found that large tubewells, though expected to be more efficient than smaller wells on the basis of purely technical and economic calculations, have in fact proved much less efficient owing to problems associated with their management. Many of these problems have arisen because the command area of the tubewells is too large and includes too many farm units. Apart from the physical difficulties of conveying water over much larger distances than would be necessary with a greater number of smaller wells, there are many more grounds for social conflict and it is much more difficult to establish communal responsibility for high standards of operation and maintenance.
With larger tubewells the choice between operation by a publicly-owned operator and a supervised farmers' group may be difficult to make. Unless subjected to close management control, public tubewell operators have often taken advantage of their position of local monopoly to exploit the farmers who are dependent on the water from their wells; but group operation may also cause problems, particularly if the numbers involved are large. Essential functions of government, irrespective of patterns of ownership and operation, are the organization of comprehensive studies and surveys in advance of any major groundwater development, continuous monitoring of groundwater conditions as extraction proceeds, and the imposition of controls over extraction rates where necessary. Where groundwater is developed privately, it will be particularly important to provide credit and other supporting services, although it may be possible eventually to transfer many of these service functions to the private sector as new businesses are set up to undertake repair and construction work.
Once a decision has been reached on the most appropriate form of horizontal organization at the project level, there are further questions to be considered regarding the pattern on which specialist units should be set up within the organization. In particular, there are two important questions concerning the division of responsibilities for water management. The first is whether water distribution and maintenance work should be undertaken by the same unit or two different ones; and the second is whether there should be an additional unit with specific responsibilities for assisting farmers in matters related to the water management at the watercourse or small group level, which is discussed further on in the text.
With regard to the first question, there are strong grounds for arguing that, except on very small projects where the two functions may have to be combined, there should be separate units for water distribution and maintenance, each with its own specialist cadre. The reason for this is that the skills required in each case are very different. Water distribution requires detailed knowledge of agricultural processes, whereas maintenance work is best carried out by civil engineers in the case of canal and drainage systems and by mechanical engineers in the case of wells and low-lift pumps. If functions are to be differentiated on the basis of specialization of skills, this would imply that canal systems should have two units - one for operation and one for maintenance - and systems with additional groundwater or pumped supplies three units: (i) canal and well operation (for conjunctive use of surface and groundwater); (ii) canal and drainage maintenance; and (iii) well and pump maintenance. This pattern, which is used by the Irrigation Associations in Taiwan Province (Republic of China), has obvious merits over the following division of responsibilities which is not uncommon elsewhere: (i) canal operation and maintenance; (ii) well/pump operation and maintenance; and (iii) drainage operation and maintenance.
An alternative solution would be to give water distribution specialists responsibility for maintenance as well. The feasibility of such a solution would depend on the scope and complexity of the maintenance function of the project concerned. On some projects, for example, mechanical maintenance is a complex and time-consuming activity and requires the undivided attention of a specialist cadre; on others, maintenance work may include a substantial amount of minor design and construction work, in which case the full-time attention of a separate cadre is again required. In all cases, the important principle to be observed is that water distribution should always be done by people with appropriate specialist training, and the quality of their work should not be impaired by overloading them with excessively heavy maintenance responsibilities. The qualifications and training required by water distribution and maintenance staff are described in Chapters 5 and 6.
i. Officials and farmers
Subject to certain limitations, it is generally desirable in the administration of any kind of agricultural project or programme that as much responsibility as possible should be delegated by government from its own officials to the farmers themselves. This is not only because there are obvious advantages to government if some administrative responsibilities can be transferred from officials to farmers, but also because in agriculture there are nearly always significant differences in physical conditions and farmers' needs between one small locality and another. The quality of executive decisions is therefore likely to be substantially improved if farmers have an effective means of informing officials about their local circumstances and discussing with them alternative solutions to their problems. There is also the point, particularly important in the case of rural communities in the early stages of social and economic transformation, that one of the long-term objectives of rural development - the strengthening of the position and status of poorer and smaller farmers - will be enhanced by their active participation in the management of their own development process.
There are, however, limits to the amount of responsibility that can profitably be delegated to the farmers in any particular set of circumstances (not only in terms of benefits to government but also of benefits to the farmers themselves). The most important limiting factor is the farmers' level of education and the length of their experience of irrigated agriculture. Where their level of education and experience is low, the form of official-farmer relationship likely to produce the best results is one involving a high degree of supervision and support and a relatively low degree of delegation. However, as the farmers' experience and management capabilities increase over time, increasing responsibilities can be delegated to them and the need for the officials to act as educators and supervisors is correspondingly reduced.
Three other factors place an upper limit on the extent to which it is feasible for farmers to participate in irrigation management decisions. The first is the technical complexity of the water distribution activity. The second is the need on all irrigation systems except the very smallest for an executive agency which can implement all water distribution decisions impartially and act as an arbiter in the case of major water disputes between groups of farmers. The combined effect of these two factors is that the farmers' responsibilities with regard to water distribution (whatever the level of their management capabilities) are restricted to participating in strategic decisions about planned patterns of cropping and water scheduling before each major crop season, monitoring the pattern of scheduling actually adopted by the officials within each crop season, and taking direct responsibility for operation and maintenance decisions below the watercourse outlet. The officials must retain responsibility for the execution of day-to-day operations of the main distribution system and the settlement of major water disputes above the watercourse outlet.
The final limiting factor is the degree to which the project is financially dependent on government funds. Because of their very high capital costs, there are few irrigation projects in any part of the world to which government has not been a net financial contributor. To the extent that it has contributed funds towards the capital and recurrent costs of a project, government clearly has a right to take part in decisions as to how they shall be spent. But at the same time, the greater the farmers' own financial contributions, whether their payments go into general government revenue or are partly or fully retained by the project organization, the greater their entitlement to a share in decisions about the project's programme of expenditure.
Certain types of projects tend to be subject to a relatively high degree of government control at all stages of their development owing to the nature of their objectives. For example, specialized cash-crop schemes (which are usually settlement schemes by origin) frequently make use of various measures which restrict farmers' freedom of choice for the purpose of standardizing the production process and stimulating increased output.
At the opposite end of the organizational spectrum to the specialized cash-crop settlement schemes in terms of their potential for decentralized management are the Irrigation Associations, which make formal provision for farmers' representative bodies at the project level and sometimes at various primary and secondary levels within the project. However, although IAs are intended to be highly democratic bodies, in the final stages of their development farmers require high levels of education and experience if they are to be capable of assuming the full responsibilities of self-management. It is therefore common practice in the early years of an IA for officials to be given very substantial responsibilities with regard to management decisions and farmers' training.
ii. Within the project organization
Within each unit of the project organization, the most appropriate distribution of responsibilities among different levels of staff will depend largely on three sets of factors:
- the nature of the activity or activities with which each unit is chiefly concerned;
- the technology used to perform the activity, whose appropriateness will depend in turn on the relative scarcity of capital and labour;
the availability of skills (management skills, technical skills, skills in communication) at different levels of the agency.
The ways in which the nature and technology of different activities combine to affect the pattern of vertical organization required for them can be illustrated by the following examples:
a. In the case of any work concerned with farmer extension and training, the principal need will be for a large number of junior field staff with skills in communication. Their level of technical knowledge need not be high. A small number of higher-level specialists will however be needed to provide them with the necessary technical advice and knowledge.
b. The mixture of skills required for water distribution may vary considerably according to the technology of the water delivery system (number of control structures; whether manually or automatically operated) and the complexity of the water distribution method: simple proportional distribution methods require less computing skill at all levels and less communication skill at the lower levels than an 'on-demand' method or a method which calls for differential allocations of water to areas with different designated cropping patterns. The greater the availability of relatively unskilled manpower, the more labour-intensive the technology should be.
c. Where the technology of the irrigation system is simple and labour abundant, the maintenance of civil works can be done by a large body of relatively unskilled labour, directed and supervised by a small number of technically skilled senior officials. Systems with more complex technologies, particularly those requiring a lot of mechanical maintenance (as in the case of tubewells), will have a greater need for qualified technicians at the lower and middle levels of management.
In the case of all these activities, good communications and transport can either enable fewer staff to achieve the same level of effectiveness or else increase the effectiveness of a given number of staff. The first of these is likely to be the more important objective where staff costs are high, the second where these are low.
With information about these three factors - the nature of each activity to be performed by the organization, the administrative manpower available, and the technology appropriate to that manpower - it should be possible for the planners of new irrigation projects to build up a detailed picture of the organizational structure required. Chapters 5,6, 7 and 8 provide information regarding manpower and machinery requirements for each main activity. A very important point to note here is that the availability of manpower and skills should be one of the principal determinants of choice of technology and not vice versa. The implication of this is that investigations into the organization and management aspects of irrigation projects should begin at an early stage in the planning process and not, as so often, at the end of the process (if they are carried out at all).
Once a basic structure has been established, there is still the question as to how much responsibility should be delegated from more senior to more junior staff within that structure. This will depend on the degree of management skill attained by staff at each level. As in the case of official-farmer relationships, the distribution of skills in the early stages of economic transformation is likely to favour an initial pattern of high supervision/low delegation. But with the passage of time, as junior staff increase in experience, increasing responsibilities can be delegated to them. This process of change can be greatly assisted by appropriate in-service training of junior staff, which should be combined with promotion policies stimulating their motivation to perform well.1 The adoption of policies that will help to promote as rapid a transfer of responsibilities as possible to lower levels of the administration is particularly important in those countries where higher level agricultural managers are themselves in short supply; otherwise senior staff will be continually burdened with excessively heavy workloads.
1 See Chapter 4.
There is a particular need to establish farmer groups at the local level where the number of small farm units is so high that it is impossible for each one to be provided with services such as agricultural extension. On large and medium-sized irrigation projects, this generally leads to the need for water users' groups to be formed for the purpose of organizing water distribution and maintenance work below the watercourse outlet. The project's organizers can hardly ever afford the staff which would be required for the direct management of numerous small channels and field outlets. Moreover, if well-trained, water users' groups are likely to be the best possible agencies for the purpose owing to their members' intimate knowledge of local conditions.
The need for a special unit within the project organization to educate farmers about watercourse operation and maintenance depends on the level of the farmers' previous experience of irrigated agriculture and the degree of social cohesion within the farming community. In small-farmer agriculture, there will almost always be a need for intensive extension work on watercourse management in the early stages of an irrigation project. This should have the dual purpose of educating farmers about the technical aspects of good operation and maintenance within the watercourse area for which they are communally responsible and also of fostering the development of small group institutions to which increasing responsibilities can be delegated over time. Initially, a separate cadre of field staff is likely to be required for the purpose, but as time passes and the farmers' capacity to manage their own watercourse affairs increases, it should be possible to transfer these extension and supervisory responsibilites to the water distribution staff.
One point which must be strongly emphasized is that agricultural extension staff should on no account be asked to take on this work. Not only would this leave them less time for their principal work at the farm level which is of particularly vital importance in the early stage of irrigation development, but there are certain aspects of the work (involving the resolution of local conflicts over water and the imposition of discipline) which are incompatible with the agricultural extensionists' effective performance of their central role as the farmers' counsellor and friend.
However, the work of these special units is likely to be a far from easy task. Not only will individual farmers be unfamiliar with good water management practices but there may be little natural social cohesion among those whose land happens to fall within a particular watercourse command and who are therefore expected to cooperate with one another in a mutually advantageous way. Friction is particularly likely to occur in societies with religious differences and significant disparities in power between larger and smaller farmers, manifesting itself in refusals to share water equitably or to collaborate in watercourse maintenance. There is also almost always a potential conflict between the interest of head-reach and tail-reach farmers within the watercourse.
Under these circumstances group formation will often require much patient effort, backed by a capacity to use powerful sanctions against persistently uncooperative individuals or groups, and an essential condition for success is a reliable and predictable water distribution service to the watercourse outlet. However, once farmers have been persuaded of the long-term private benefits of collaborative action, at least two other major advantages can follow from the existence of effective water users' groups besides the primary one of improved watercourse management:
a. They can help to provide a point of contact for other services - agricultural extension, credit, input supplies - and may prove particularly valuable for this purpose in societies where larger village-level cooperative organizations have failed because of domination by large farmers.
b. They should also provide the best possible foundation on which to build a system of representative farmer participation in decision-making at the project level. Indeed, unless such local groups can be formed, it is unlikely in many societies that the farmers chosen to participate at the project level will be genuinely representative of the mass of farmer opinion. Local group formation is thus a key to a long-term process of social transformation within the project area.
Because of the great potential contribution of watercourse groups to improved project performance, the manner in which a project's watercourses are designed at the planning stage takes on added importance. It is generally agreed from observation of spontaneously formed farmer groups (including those formed for the purpose of managing small "indigenous" irrigation systems), that there will be the greatest chance of cooperative action and the least chance of major conflict or large farmer domination where the groups are very small - usually between 10 and 30 farmers.1 This factor should therefore be very carefully considered by project planners. Decisions based entirely on topographical considerations and cost of survey are unlikely to be satisfactory. Questions also need to be asked about the number of farm units within each watercourse and the cohesiveness of the local society.2 This is another example of the importance of basing technological decisions on social science information and hence of the need to consider management issues early in the planning process.
1 See literature references (Coward).
2 Further questions also need to be asked on issues which affect water management practices at farm level (e.g. what is the rate of water flow which an individual farmer, with his existing technology, can most easily handle?).
A few words also need to be said about organizational structure at the national and provincial government levels, since this too can often have a very significant influence on an irrigation project's performance.
As far as horizontal relationships between departments are concerned, it is clearly desirable that collaboration between irrigation and agriculture be very close even if they are not in fact part of the same Ministry. This is particularly important for the purposes of achieving a balanced and integrated approach to project planning. Invaluable support can also be given to the improvement of management practices at the project level through a central research agency with responsibilities for directing and coordinating field experiments on crop water requirements, optimum irrigation intervals, etc., in different agro-climatic regions of the country. The results of these experiments can then be tested on each irrigation project and adapted to its own particular circumstances.
The nature of the vertical relationship between the higher level of government administration and project management is of great importance. If the project manager is to perform his function well it is essential that so long as his appointment is in the hands of government, very substantial authority and autonomy of decision-making should be delegated to him. At the later stages of a project's development the manager may automatically acquire greater autonomy if a policy is followed (as in many Irrigation Associations) of retaining the revenue from farmers' fees within the project agency itself. This process may eventually lead to the manager becoming an appointee of the farmers themselves, with government providing only a limited amount of supporting finance. In the absence of such a policy of devolution of powers to an Association or a project authority, however, the government must ensure that the project manager and his senior staff are in a position to act as relatively independent managers, not simply as passive administrators of policies laid down for them at higher levels.