Farmers' Willingness to Pay for Irrigation Water: a Case of Tank Irrigation Systems in South India

The economic value of tank irrigation water was determined through Contingency Valuation Method by analyzing farmers' willingness to pay for irrigation water under improved water supply conditions during wet and dry seasons of paddy cultivation. Quadratic production function was also used to determine the value of irrigation water. The comparison of the economic value of water estimated using different methods strongly suggests that the present water use pattern will not lead to sustainable use of the resource in the tank command areas. Policy options for sustainable use of irrigation water and management of tanks in India were suggested.


Introduction
Water has unique characteristics that determine both its allocation and use as a resource in agriculture.Irrigation is a vital component of agricultural production in many developing countries.Over the years, many researchers have examined the valuation of water as an instrument for improving water allocation, reducing water consumption and management of the irrigation systems [1][2][3][4][5][6][7][8][9][10][11][12].The Fourth Principle of the 1992 Dublin statements defines water as an economic good in order to achieve efficient and equitable use, and encourages conservation and protection of water resources.The Fourth Dublin Principle denoted a landmark shift in emphasis to the economic dimensions of water use in general, and irrigation development in particular [13].
Willingness to Pay (WTP) is an economic concept, which aims to determine the amount of money a consumer is willing to pay for the supply of water.The consumers' WTP is becoming increasingly popular and is one of the standard approaches that is used by market researchers and economists to place a value on goods or services for which no market-based pricing mechanism exists [14,15].Experiences show that very high level of WTP for water is observed in developing countries [16][17][18][19][20].
Literature suggests that two approaches are being used to analyze the consumers' WTP.The direct approach, involves taking a survey through a structured questionnaire of consumers' WTP specified prices for hypothetical services, also referred as Contingent Valuation Method (CVM).The direct approach used in CVM has been to directly ask survey respondents to state their exact maximum WTP for the particular use or non-use value of the water.The WTP is defined as the amount that must be taken away from person's income while keeping his utility constant [21].The CVM still have serious methodological and theoretical shortcomings when used to assess WTP for non-market based goods and services, such as format bias, embedding effect, ordering problem, starting bid effects, strategic bias, information bias, non-response bias, payment vehicle, free rider problem, warm glow effect [22][23][24][25].However, CVM is still useful tool for water resource management in developing countries [19,20,24].The indirect approach involves observing consumers' behavior and modeling of behavior based on the approximate expenditure in terms of time and money to obtain the goods or services and infer about WTP through measurement of revealed preference [26,18,27].The revealed preferences approach derives WTP values indirectly from the actual market behavior of individuals.In this study, both direct (CVM) and indirect (crop-water production function analysis) methods were used to determine the economic value of irrigation water considering tank irrigation system as a case in South India.

Background
The Tamil Nadu State in India is deficient in water resources.The annual available water resource per capita in Tamil Nadu is estimated at 600 M 3 , which is quite small when compared to 4,000 M 3 of the national average.Hence, it becomes necessary to utilize the limited water resources efficiently in the State.Since total surface water sources in the State is estimated about 340 million M 3 and the developed surface water is 333 million M 3 , it is difficult to develop new water resources for irrigation.Wells are the major source of irrigation in the State accounting for 46.4 % of the net irrigated area followed by canals (29.1 %) and tanks (23.9 %).Over years, the area irrigated by tanks is decreasing while the area irrigated by wells is increasing [28].The current Water Policy of Tamil Nadu State stressed the importance of equitable use of scarce water resources, and in the planning and operation systems, water allocation priorities were given for drinking purposes followed by irrigation, hydropower, industrial and other uses.Hence, it is imperative that optimal and sustainable patterns of water use be established to meet the requirements of a growing population and need for basic agricultural foodstuffs.
In many parts of the world both the free distribution and under pricing of water have led to inefficient allocation of the scarce resource.Both under pricing of water and lack of cost recovery mechanisms in government managed irrigation systems had resulted in poor O & M [29].Actions are necessary to use the water sustainably and manage the tank irrigation systems in South India.One strategy is to reduce water demand by adopting water conservation programs and improving water use efficiency, while another strategy involves a water pricing policy.This policy has the advantage that the income could be used to finance developments like the O & M of irrigation system.Pricing of water can also be considered as a pre-requisite for sustainable use of water resources.The underlying principle of irrigation water pricing in relation to sustainability concerns is that it should reflect the benefits forgone in the future from using a unit of water today which refers to the opportunity cost of irrigation water.[30].The economic sustainability criteria or the socially optimal rule for water use can then be assessed by comparing farmer's WTP and the opportunity cost of water.From the stand point of economic efficiency, water prices should relate to the marginal value product or the opportunity costs [31][32][33].From the government's viewpoint, water price should at least cover capital costs as well as O & M expenses [34].From the standpoint of feasible revenue collection, tank irrigation water charges depend highly on farmer's WTP.The objective of this paper is to determine the value of tank irrigation water, which farmers would be willing to pay under dry and wet seasons and thereby draw policy implications for sustainable use and management of the tank irrigation systems.

The Study Area
The Tamil Nadu state covers a total area of 130,069 km 2 .Its climate is basically tropical, benefited both by Southwest and Northeast monsoons.The average number of rainy days in the State is 50 per year with an annual rainfall of 925 mm.The total population of the State is 55.60 million.Agriculture is the traditional and major industry of the State, employing about 60 % of its labor force and it contributes for about 25 % of the Net State Domestic Product.The net cultivated area of the State is 5.85 million hectares which accounts for about 45 % of the State's land area.The average size of operational holdings in the State is 0.93 hectares [28].
An irrigation tank is a small reservoir constructed across the slope of a valley to catch and store water during rainy season and use it mainly for irrigation besides rearing fishes, growing trees and domestic purposes like washing and bathing as well.The tanks have existed in India from time immemorial, and have been an important source of irrigation especially in Southern India.They account for more than one-third of the total irrigated area in South India.The tank irrigation system has a special significance to the resource poor marginal and small farmers who depend on tank for irrigation.There were more than 39,000 tanks in the State, with varying sizes and types.The tanks are classified into system tanks (which receive supplemental water from major streams or reservoirs in addition to the yield of their own catchments area) and non-system / rain fed tanks which depend on the rainfall in their own catchments area and are not connected to major streams / reservoirs.The tanks are also classified into Panchayat Union (PU) tanks and Water Resources Organisation (WRO) tanks based upon the management authority.This study concentrates on rain fed tanks that are managed by WRO and PU authorities.

Socio-economic characteristics of farmers
The socio-economic profile of the sample respondents revealed that 76.81 % were male and 23.19 % were female.A majority of the respondents (62.31 %) were between 41 to 60 years old followed by 36.23 % in 20-40 years category.Educational status of the respondents was found to be at high school level (46.37 %) followed by primary level (24.63 %).About one-fifth of them were illiterates.Three-fourth of the respondents lived in joint family system and had more than five members in their family followed by one-fourth of them adopted nuclear family system with less than five members.
Majority of the respondents (68.12 %) were engaged in agriculture followed by both agriculture and business (29 %).As regards the size of the holding, a majority of the respondents (68.12 %) operated less than one ha farm followed by 23.19 % with farms ranging between 1.1-3.0ha and 7.25 % had 3.1-4.0ha.Almost all the farmers belonged to either middle or lower socio-economic class.This group of farmers will have significant impact on the average WTP value as their WTP for water is usually expected to be low.

Farmer's perception on availability and pricing irrigation water
Farmers' perception about irrigation water availability assumes important for their WTP.Of the total command area of 5927 ha, only 2691.69 ha was found to be cultivated using the tank water.Rice was cultivated in 2691.69 ha using tank water supply during dry season.Of the total respondents, only 79.73 % cultivated rice during dry season and the rest could not cultivate due to unavailability of irrigation water from tanks.A majority of these respondents (63.77 %) revealed that the water received by them was insufficient to harvest a successful crop and the rest were not certain about whether water was sufficient or not.The perception of the farmers on availability of irrigation water during wet season was analyzed.Majority (60.87 %) of them perceived that the water in the tank would be sufficient to raise single rice crop successfully during the wet season.
The reasons attributed to the insufficient availability of water from the tank to irrigate their crop were: seasonal failure of monsoon (60.32 %), poor maintenance of tank and channels (53.67 %), inequitable water distribution (35.33 %), and increased number of wells leading to poor water storage in tanks (25.43 %).The suggestions were elicited from the respondents for possible improvement in the tank system to overcome the present situation.A majority of them suggested for regular maintenance of tanks in terms of desilting, cleaning the supply and distributor channels (50.42 %) followed by need for the effective functioning of the Water Users Association (WUA) to distribute the water equitably (43.23 %) and eviction of encroachments (19.32 %).
The farmers were asked to react to the opinion about pricing the irrigation water as a way to recover the O & M cost of the tank system.Majority of the respondents (60 %) reacted positively stating that pricing of irrigation water might enhance the functioning of the tank system and/or enhance the water use effectively by the farmers.It is expected that the farmers' perception on water availability and pricing might have a bearing on their WTP for water.

Rice yield, per unit production cost, and existing O & M costs
The average yield of rice in the study area during wet season was 4748.03 kgs/ha and its average cost of cultivation was INR 2216.51/ha.The land revenue including water cess paid by farmers according to the three land grade classification was known to be INR 5. 27

Farmers' WTP for irrigation water
The results of the survey revealed that in general, a majority of the farmers were willing to pay for the irrigation water under improved water supply conditions from the tank system.Nearly 50 % of the farmers were willing to pay for the irrigation water during the wet season.Among these willing farmers, two-third of the farmers was willing to pay INR 200/ha/year and one-third at INR 250/ha/year.The average WTP value during wet season was INR 212.50/ha/year.In the case of dry season, two-third of the farmers responded positively for the WTP.Of the willing farmers, nearly 64 % were willing to pay to an extent of INR 200/ha/year followed by INR 250 /ha/year by 24 % and the higher extent of INR 300/ha/year by 12 % of the farmers.The average WTP value during dry season was INR 224.50/ha/year.The overall mean WTP value across seasons was INR 218.50/ha/year.
The negative response for the WTP for irrigation water by the rest of the respondents was due to the belief that tanks do not get filled to its full capacity either due to failure of monsoon or due to poor maintenance bestowed by the management authorities over the common property resource in the past.Some also felt that tanks are common property resource for open access by farmers in the command area.

What affects farmers WTP for Irrigation water?
The estimated results on factors affecting the farmers' WTP for irrigation water across seasons are presented in Table 1.It could be seen that the family labor force (FLABOUR), area under rice cultivation (AREA) and the water requirement (WREHACM) found to be significant factors influencing farmers WTP in the wet season.While in dry season, the variables AREA and WREHACM are found to be significantly influencing the farmers' WTP for irrigation water.Area under rice cultivation had significant bearing on WTP by farmers.The small and marginal farmers had relatively higher WTP for water when compared to medium and large farmers.Generally this is true, as large farmers use to own wells and their dependency over tank water is relatively less when compared to small and marginal farmers who solely depended on tank water for irrigation.It is also evident that as family labor force increases the WTP decreases.The decision making is easier in the case of a small/nuclear family compared to joint family.Besides, allocation of money for irrigation water is also more feasible as the family expenditure would also be relatively less.The positive relationship between water requirement and WTP might be due to farmers' perceived fact that water requirement at critical growth stages of the rice would severely affect the yield and hence farmers were willing to pay for the irrigation water irrespective of the season.Electricity subsidy for agriculture has been implemented in few selected states of India including Tamil Nadu.The surface irrigation sources namely, canals and tanks account for about 55 % of the total irrigated area in Tamil Nadu state.In the case of these surface irrigation systems, the contribution of wells was very much limited as compared to the areas that solely depended on well irrigation.Hence, the WTP was recorded among the farmers who depended solely on the tanks which were not directly competing with the electricity subsidy in the well irrigation systems, as already the surface water charges were also subsidized (very low water charges as compared to the O & M charges).
Hence the farmers in the tank systems were willing to pay more than the actual water charges being paid by them.In addition to this fact, it was also observed that only very few farmers who operated large sized farms in the tank command area owned wells unlike small and marginal farmers who depended on tanks.The study also revealed that farm size had negative influence on the WTP behavior of farmers for tank irrigation water.It could be stated that the small and marginal farmers (without wells) expressed their interest towards WTP as compared to their counterparts who may or may not owned a well.Hence the result obtained on WTP is justified in the context of free electricity policy of the sate.
It is evident from the analysis that irrespective of seasons, the significant and most influencing factors that determine the farmers' WTP for irrigation water from tank were found to be the area under rice and water requirement.Most of the CVM studies carried out in developing countries were limited to the measurement of users WTP under improved water supply conditions for drinking water supply.As the respondents' need for drinking water is somewhat different compared to the tank irrigation water the results cannot be directly compared.However, comparison with some of the earlier studies can help to provide useful insight on the contribution of this study to a limited works in this field with special reference to tank irrigation.Studies on the WTP for improved drinking water services show a positive relationship with the respondents' attitude, education, income and distance for water source [35,36] and negative relations with the age, sex, water quality index, distance and family size [35][36][37][38].

Determining the economic value of irrigation water
The economic value of irrigation water was determined by employing production function approach [39].The marginal value of water of each ha.cm. is the marginal physical product times the output price.A quadratic production function was estimated with Yield (kgs/ha.) as dependent variable and volume of irrigation water used in ha.cm (WATER) as independent variable.The estimated production function is as follows: The estimated values of quadratic production function were used to derive the marginal value product of water across seasons.It is evident that the marginal productivity of water worked out to INR 243.07 in dry season while it is INR 236.94 in wet season.It is lucid from the analysis that the marginal value productivity of water in dry season is little higher than in wet season.This may be true due to the fact that farmers faced water scarcity in dry season and produce their crop with inadequate water.

Comparison of economic values of water
The opportunity cost of irrigation water was calculated on the basis of ground water extraction costs, i.e., assuming that the only source of irrigation for the farmers is ground water and if the existing tank source was not available.The cost of ground water sold by the well owners in the tank command area ranged between INR 30 and INR 40 per hour of extraction using 5HP motor.It is also worth to note that the farmers in Tamil Nadu enjoyed free electricity for agricultural purpose.To irrigate one ha cm of land by well water it needs an average of 3 hours/irrigation.On an average 15 irrigations were needed for raising a successful rice crop in the study area.Hence, for rice cultivation involving 45 hours of irrigation with the average cost of INR 35/hr works out to INR 1575/Ha.cm.Compared to the value of irrigation water estimated from other methods the opportunity cost of irrigation water appears to be very high.
The comparison of the economic value of water estimated using different methods strongly suggests that the present water use pattern and ultimately the dominant rice based cropping pattern will not lead to sustainable resource use pattern in the tank command area.Although the indirect methods of valuation has resulted in a higher value compared to the mean value of WTP, it is difficult to arrive at definite conclusions as the way in which the maximum WTP is supposed to vary due to water scarcity during the dry season.However, the WTP is found to be higher than the O & M costs of the state average and almost close to the marginal value product of water.This is a positive sign to make appropriate policy decisions on cost recovery for meeting out the O & M expenditures of the tank.
Hence, based on the results of this study, policy implications for achieving the sustainability of water use for irrigation from the tank system are proposed.

Data
The CVM method was employed in this study to measure farmers' WTP for irrigation water.For the purpose, thirty one tanks were randomly selected from the tank intensive districts of Tamil Nadu state.Of the total number of selected tanks, 26 tanks were managed by WRO and 5 by PU.The command area of these tanks was divided into two zones: head and tail for the purpose of conducting survey.Totally 62 respondents who depended on tanks as the sole source of irrigation were drawn from the selected 31 tanks covering both head and tail for the WTP survey.Among the respondents, 54 depended on WRO and 15 on PU tanks.The purpose of a WTP survey is to elicit farmers' WTP for the tank water drawn for irrigation use.The WTP interview schedule designed for the survey consisted of both close-ended and open-ended questions.In the case of close-ended, farmers were asked whether or not they would be willing to pay a specific amount under improved levels of water supply from tanks for irrigation.In the case of open-ended questions, farmers were asked on how much they would be willing to pay at the improved water supply conditions both in the dry (when water is too scarce) and wet seasons (when water is insufficient for few farmers).Survey was conducted during 2008, which was a normal year in terms of water availability in the tanks.The monetary unit used is Indian Rupees (INR) and area in ha.

Factors affecting farmers' WTP
Our objective here is to identify the factors which influenced the WTP behaviour of farmers.The responses were discrete and therefore, a logit model was developed to examine the WTP behaviour of farmers for tank irrigation water.The logit model which is based on cumulative logistic probability function has the advantage to predict the probability of farmers' WTP for the irrigation water.
The Logit model: The Logit model assumes that the random variable Zi predicts the probability of farmers WTP.Thus, Thus for the ith observation, The probability of farmers' WTP for irrigation water is modeled as a function of various individual, household level and farm level factors.The model is represented as follows: WTP = f ( EDN, FSIZE,AGE,FLABOUR,AREA,WREHACM ) The dependent variable is the farmer's decision on WTP for the irrigation water.It assumes 1 if the farmers is willing to pay for irrigation water and 0, otherwise.Logit model was used to describe the farmer's decision on whether or not they agreed to pay for existing supply of irrigation water as well as under improved water supply conditions.The model was estimated by LIMDEP 7.0.In the case of close-ended questions, the probability of obtaining a 'Yes' or 'No' responses was explained through limited dependent variable analysis [40].

Empirical relevance
The farmers' willingness to pay for tank water is expected to influence by educational level of the head of the household (EDN), family size (FSIZE), age of the respondent (AGE), family labour force (FLABOUR), area under rice cultivation (AREA) and water requirement at farm level (WREHACM).
One may expect that the socio-economic characteristics of the respondents play crucial role in farmer's WTP for tank water.Education can have two different types of effects on WTP for tank water.Education some times offers exit options and this is likely to reduce WTP.However, educated farmers can be influential in the household and can participate in the tank management activities.If this happens, then farmers' WTP will be positively influenced by education.Similarly, the age of the respondents is critical for decisions regarding family and farm operations.Age generally reflects the experience of the farmers and it has influence on farm household's decisions.However, the exact sign of this variable is uncertain.The farm family labour force implies the income sources of the farm family.Different income sources positively influences the farmers participation in tank management activities and hence the WTP.As income from different sources increases, the farmers likely to pay more for the tank water.The area under rice cultivation is expected to influence negatively the willingness to pay for tank water.Generally the small and marginal farmers who depend on tank water for irrigation may have higher WTP when compared to large farmers who depend less on tank water for irrigation.A priori, it is expected that the area under rice cultivation will have an inverse relationship with WTP.The water requirement at farm level may increase the farmers WTP for tank water.Since the water requirement at different stages is critical for increasing productivity of crop, it is expected to have positive relationship with WTP for tank water.

Conclusions
The results revealed that the mean WTP value of farmers for irrigation water was The study also indicated that farmers were willing to pay considerably more than the average O&M costs incurred by the state on tanks and were also willing to pay almost equal or slightly lesser amount than the marginal value product of water.The average value of WTP irrespective of seasons for the tank irrigation water was found to be considerably less than the opportunity cost of the irrigation water, indicating the unsustainable use of irrigation water from the tank system at present.Hence it can be concluded that charging water depends highly on farmer's WTP from the standpoint of feasible revenue collection.
Sustainability of irrigation systems is very important from both farmers' and government perspectives.Conversely, developing countries like India are facing tremendous budgetary pressure arising from the need to defray irrigation costs.Quite often, farmers do not receive adequate service owing on to an insufficient O & M budget.This undoubtedly affects crop productivity and farming income.It is therefore important to decrease the budgetary burdens of government through local control and support.The evidence assembled from the Philippines suggests that there are significant financial, economic and social benefits generated from irrigation charges.If the charging system is appropriate, it will result in improved irrigation performance [41,42].
Pricing water is important not only for generating revenues but also for promoting efficient use of water resource [43].A free or very low water charge encourages overuse, reduces the incentive for farmers to cooperate or participate in irrigation originations, and may result in low system productivity and poor conservation [44].The charges could also bring an ownership feeling to the farmers [45], which will ultimately lead to better use of available water and increased crop production.Of course, collecting irrigation fees should not create any disincentive for farmers to irrigate, which means that the cost recovery mechanism should be compatible with resource use.This can be achieved if the fees are treated as payment for the service rendered and not as tax.Experience from Taiwan suggests the use of an institutional mechanism for promoting managerial performance of irrigation systems [46].
Our results reveal that farmers were willing to pay for the irrigation water drawn from the tanks.As the marginal value productivity of water is positive in both the dry and wet seasons, providing assured irrigation water through improved maintenance of the tanks will help farmers to achieve increased productivity.Hence the study recommends strengthening the existing WUA by empowering them to fix rational water charges for irrigation and collect it from the farmers to meet out the O & M activities of the tanks.Later based on the performance of the empowered WUA, turning over the responsibility of managing the tank irrigation system and its maintenance, authorizing WUA completely to take over the control of other benefits of tank namely auctioning fishing rights and trees rights can be thought of by the Government in future.This policy option would promote the sustainable use of irrigation water in the tank command and manage the tanks effectively.

((
Wet season: The estimated equation was obtained as: Yield = 1807.93+ 31.97WATER *** -0.01 WATER 2 in parentheses indicate estimated 't' ratio.*** : significant at 1 % level.The price of output (paddy) is INR 8/kg in both wet and dry seasons.The quantity of water use is 117.89 ha cm in wet season and 110.87 ha cm in dry season.The value of the marginal product of water (VMP) is evaluated at mean values of water use.Marginal Value Product = Marginal Physical Product * Price of one unit of Paddy (INR/kg) VMP = [31.975-0.02 * (117.89)]* 8 VMP = INR 236.94 Dry season: The estimated equation was obtained as: Yield = 1247.75+ 39.25 WATER ** -0.04 WATER 2 in parentheses indicate estimated 't' ratio.** : significant at 5 % level.Marginal Value Product = Marginal Physical Product * Price of one unit of Paddy (INR/kg) VMP = [39.254-0.08*(110.87)]* 8 VMP = INR 243.07 INR 218.50/Ha/Year.The marginal productivity of water was INR 243.07 in dry season while it was INR 236.94 in wet season.The state average O & M expenditure on tanks was INR 55/year/ha for tanks managed by PU and INR78 per year per ha for PWD tanks.The opportunity cost of tank irrigation water was INR1575/Ha cm.
, INR 5.01 and INR 6.00 per ha of the land in the tank command.For every INR of land revenue, the water cess works out to INR 0.33.The state average O & M expenditure on tanks considering the past 10 years data (1989-1999) was about INR 55/year/ha for PU tanks and INR 78 per year per ha for WRO tanks.Normally the O & M expenditures are met from the local irrigation grant which is made available to the local Panchayat (local governing body at village level) to maintain the tanks in a five-year repair cycle [28].

Table 1 .
Logit estimation of factors influencing farmers' WTP for irrigation water.