rainwater harvesting and rural livelihoods in nepal · rainwater harvesting and rural livelihoods...

Rainwater Harvesting and Rural Livelihoods in NepalRishi Ram Kattel

Working Paper No. 102–15

Published by the South Asian Network for Development and Environmental Economics (SANDEE)PO Box 8975, EPC 1056, Kathmandu, Nepal.Tel: 977-1-5003222 Fax: 977-1-5003299

SANDEE research reports are the output of research projects supported by the SouthAsian Network for Development and Environmental Economics. The reports have beenpeer reviewed and edited. A summary of the findings of SANDEE reports are alsoavailable as SANDEE Policy Briefs.

National Library of Nepal Catalogue Service:

Rishi Ram KattelRainwater Harvesting and Rural Livelihoods in Nepal

(SANDEE Working Papers, ISSN 1893-1891; WP 102–15)

ISBN: 978-9937-596-31-2

Keywords Rainwater harvesting Rainfed agriculture Technology adoption Nepal

SANDEE Working Paper No. 102–15

Rainwater Harvesting and Rural

Livelihoods in Nepal

Rishi Ram KattelDepartment of Agricultural Economics and Agribusiness Management Agriculture and Forestry University (AFU) Rampur, Chitwan, Nepal

October 2015

South Asian Network for Development and Environmental Economics (SANDEE) PO Box 8975, EPC 1056, Kathmandu, Nepal

SANDEE Working Paper No. 102–15

The South Asian Network for Development and Environmental Economics

The South Asian Network for Development and Environmental Economics (SANDEE) is a regional network that brings together analysts from different countries in South Asia to address environment-development problems. SANDEE’s activities include research support, training, and information dissemination. Please see www.sandeeonline.org for further information about SANDEE.

SANDEE is financially supported by the International Development Research Center (IDRC), The Swedish International Development Cooperation Agency (SIDA), the World Bank and the Norwegian Agency for Development Cooperation (NORAD). The opinions expressed in this paper are the author’s and do not necessarily represent those of SANDEE’s donors.

The Working Paper series is based on research funded by SANDEE and supported with technical assistance from network members, SANDEE staff and advisors.

AdvisorCeline Nauges

Technical EditorMani Nepal English EditorCarmen Wickramagamage

Comments should be sent to

Rishi Ram KattelDepartment of Agricultural Economics and Agribusiness Management, Agriculture and Forestry University,Rampur, Chitwan, NepalEmail: [email protected]

Contents

Abstract

1. Introduction 1

2. Technology Adoption in Agriculture 2

3. Study Area and Sampling 3

4. Methods and Variables 4

4.1 Modeling of Farmers' Adoption Decision 4

4.2 Income Model 5

4.3 Controlling for self-selection 5

5. Results and Discussion 6

5.1 Descriptive Statistics 6

5.1.1 Characteristics of Rainwater Harvesting Technology 6

5.1.2 Socio-demographic and Economic Characteristics 6

5.2 Determination of Farm Income and Farmers’ Decision to Adopt RWH Technology 7

5.3 Cost Benefit Analysis of RWH Pond Adoption 8

6. Conclusions and Policy Implications 9

Acknowledgements 9

References 10

Tables 14

Table 1: Description of main variables 14

Table 2: Characteristics of rainwater harvesting (RWH) ponds (n=141) 14

Table 3: RWH pond location, water source, distribution, purpose and status (n=141) 15

Table 4: Socio-demographic and economic characteristics of RWH adopters and non-adopters 15

Table 5: Determinants of household income and RWH technology adoption (Maximum likelihood estimates): Treatment-effects models 17

Table 6: Cost benefit analysis of RWH Plastic Pond for Upland Crops Production 18

Table 7: Rainwater harvesting benefits from district level training 19

Figures 20

Figure1: Location of the four districts 20

Figure 2: Number of vegetable and cereal crop producers among RWH adopters and non-adopters 20

Figure 3: Households farm income by training received and RWH technology adoption 21

Figure 4: Total annual average benefits and costs from RWH plastic pond 21

Appendices 22

Appendix 1: Average monthly rainfall in study area 22

Appendix 2: Determinants of household income from agriculture and livestock sector (Using OLS and instrumental variable models) 23

South Asian Network for Development and Environmental Economics6

Abstract

In this paper we examine the adoption of rainwater harvesting (RWH),

a technology that has relatively recently been introduced to farmers in

Nepal. Using data from 282 farmers in four districts, the study employs

a treatment-effects model to identify factors that influence the adoption

of RWH and the impact of adoption on farm income. We find that the

adoption of rain water harvesting is mostly driven by farmer training.

Further, adoption of this technology more than doubles household

agricultural and livestock income. With incremental annual benefits

of NRs. 69,456 (USD 700), this technology is viable from a household

perspective. Adopters benefit from an increased supply of irrigation

water, which allows them to diversify their crops from cereal production

to high-value vegetable crops. Our analyses suggests that if 10 percent

of households (7000 households) in an average rainfed district receive

farmer training, the net benefits from training in the district would be

approximately NRs. 134,907,710 (USD 1.3 million) per year from adoption

of RWH technology. Given the many weather-related uncertainties faced

by rainfed farmers in Nepal, rain water harvesting is potentially a very

useful climate adaptation strategy.

Keywords

Rainwater harvesting, Rainfed agriculture, Technology adoption, Nepal

JEL Classification: C31, C36, D61, O13, Q16

1

Rainwater Harvesting and Rural Livelihoods in Nepal


1. Introduction

The phenomenon of climate change is well established, which makes it incumbent upon developing countries, in particular, to prepare for a future that includes adaptation to various climatic effects (World Bank, 2006; UNFCCC, 2007; Cosbey, 2009; IPCC, 2014; Zomer et al., 2014). Agriculture is one of the most sensitive sectors to climate change (Cline, 2007a). In many countries in South Asia, a decline is crop yield is already linked to rising temperature, rainfall variability and extreme weather events (Cruz et al., 2007; IPCC, 2007a; IFAD, 2008). Water scarcity is expected to grow due to changes in water availability, increased water demand and lack of good management. Further, heat stress is expected to contribute to reductions in area available for high yielding wheat production in the Indo-Gangetic Plains (IPCC, 2014). If no adaptation strategies are implemented, agricultural productivity could decline by as much as 10 and 25 percent by 2080. For some countries, the decline in yield in rainfed agriculture could be as much as 50 percent (IPCC, 2007a).

Climate change will affect water resources, with implications for agricultural systems and food security (Bartlett et al., 2010). Adverse impacts will be especially critical in a country such as Nepal, where about 66 percent of the population is directly dependent on agriculture for livelihoods. This is particularly worrying because water is the major limiting resource for crop production in rainfed agricultural systems (APN, 2004; UNFCCC, 2007; Wani et al., 2009). Thus, it is important to try and understand how water can be better conserved and more effectively used for agriculture. In this paper, we address this issue in the context of a specific water conservation technology, i.e., rainwater harvesting, which is being increasingly used in Nepal.

Rainfed agricultural land accounts for 65 percent of the total cultivable land area in Nepal. The sector is highly dependent on the weather (Regmi, 2007; Malla, 2008). Since only 24 percent of the arable land is irrigated, crop productivity is significantly lower in Nepal in comparison with the rest of South Asia, leading the country to rely heavily on food imports from India (Bartlett et al., 2010). Agriculture consumes around 96 percent of all water withdrawn in the country (CIA, 2010) and contributes about 34 percent to the GDP (World Bank, 2012).1

In Nepal, more than 80 percent of precipitation occurs in the monsoon during June to September. This rainfall is characteristically heavy (see Appendix 1), leading to flooding, landslides and loss of top soil (HMGN, 2004; Malla, 2008). These occurrences contribute to crop failure and increased food and livelihood insecurity (Lohani, 2007; Malla, 2008; Kohler et al., 2010; Gentle and Maraseni, 2012). Uncertain variation in rainfall also makes rainfed cropping risky, especially for crops that have a reasonable market value such as rice, maize, wheat and vegetables (Gurung and Bhandari, 2009).2 People in the hilly areas of Nepal are subject to even more livelihood risks as water sources are increasingly scarce and located beyond a reasonable distance (NDRI, 2009).

In this context, RWH technologies have become increasingly important as adaptive strategies to manage water problems posed by climate change and rainfall variability (Getnet and MacAlister, 2012; Mutekwa and Kusangaya, 2006; Hatibu et al., 2006; Moges et al., 2011). RWH technology generally involves harvesting rainwater and

1 Nepal’s agricultural growth rate has been between 2.5 and 3 percent as against a 1.35 percent annual growth in population (CBS, 2012). 2 Though livestock of buffalo, dairy cattle, goat and sheep area feasible enterprise in the hilly areas of Nepal, the lack of a reliable water

supply can restrict the extensive use of grazing lands and create pressure on scarce water resources (Zomer et al., 2014).


diverting rainwater to reservoirs for the purpose of coping with rainfall variation and drought. The aim is to enhance soil infiltration at the site and improving rainfed cultivation through water storage. There are two commonly deployed RWH practices: surface rainwater harvesting and rooftop rainwater harvesting, both of which are equally important. The focus of this study is the individually-managed plastic or cemented RWH ponds, which were recently introduced in the hilly regions of Nepal.

The Government of Nepal is promoting non-conventional irrigation systems like RWH, drip irrigation and treadle pumps as part of its 2013 Irrigation Policy (MOI, 2014). However, to date, only about 5 percent farmers have adopted RWH for crop production, at least in the areas examined by this study. Little is known regarding why certain farmers have adopted this technology while others have not. Thus, if RWH technologies are to be scaled up in Nepal, it is important to understand farmer's adoption decisions. Consequently, in this study, we ask two related questions: 1) what are the factors that determine the adoption of RWH technology by rural farmers? and 2) what is the impact of RWH technology on farm income?

In order to answer these questions, we use farm and household-level data from 282 farmers in four districts of Nepal and a treatment-effects model to identify factors that influence the adoption of RWH and its impact on farm income. Instrumental variables models are used to check the robustness of the estimates. We find that the training received by farmers regarding agriculture and livestock production is a strong determinant of RWH adoption. After controlling for different variables such as livestock, agricultural equipment, cultivated land, share of upland areas and socio-demographic variables, we find that RWH technology adoption significantly increases annual household income from the agriculture and livestock. Benefit-cost analysis suggests that the RWH technology is viable from a household perspective in the rainfed agricultural systems because adopters can diversify from cereal crops to vegetable crops.

2. Technology Adoption in Agriculture

Technology adoption models are generally based on the theory that farmers make decisions in order to maximize their expected profits or utility (Feder et al., 1985). Accordingly, the general assumption is that there is a desire to maximize expected utility from adopting new technologies (Bekele and Drake, 2003; Asfawa and Admassie, 2004; He et al., 2007). The utility maximizing objective of individual farmers might be the same for famers everywhere -- however, the specific attributes influencing the utility of farmers and their decisions to adopt technologies are far from uniform. Further, a farmer is both a producer and a consumer (Sadoulet and de Janvry, 1995) and may maximize utility but not necessarily maximize profits at the same time.3 Following this literature, we also use a utility maximization framework in this study. We assume that farmers are able to order practices according to certain preferences, subject to the constraints given by the availability of resources.

Besley and Case (1993) have extensively reviewed existing empirical technological adoption models with regard to their consistency with underlying theoretical models of optimizing behavior. Several studies identify socio-economic, demographic, and structural factors that determine a farmer’s choice and the patterns of diffusion of the innovation over time (Dinar and Yaron, 1990; Dinar et al., 1992; Dridi and Khanna, 2005; Getnet and MacAlister, 2012). Factors such as experience, farmer’s level of education, availability of extension services and information, and supervision are likely to affect the capability of a producer to adopt and utilize available water-efficient technologies.4 However, such results are often contradictory with regard to the relative importance and influence of any given variable on farmer decision-making (Koundouri et al., 2006).

3 Despite its failure to identify the psychological processes that determine preferences, the framework is considered to be less restrictive

than profit maximization approach (Lynne et al., 1988).4 A recent study conducted by Murgor et al. (2013) in Kenya indicates that lack of access to credit, involvement in off-farm activities and lack

of capital negatively influence adoption of rainwater harvesting technologies, while the level of education of the respondents and involvement

in social responsibilities were positive drivers of adoption. Empirical results from Ethiopia suggest that economic, institutional, and ecological

factors, in addition to technical factors, are significant drivers of rainwater management technology adoption and implementation (Getnet

and MacAlister, 2012).

3


Risk is generally viewed as a major factor that influences the rate of adoption of any kind of innovation (Jensen, 1982; Just and Zilberman, 1983).5 Uncertainty generally has two features: the perceived risk associated with farm yield after adoption and production or price uncertainty related to farming itself. Koundouri et al. (2006) propose that farmers adopt new technology in order to hedge against production risk and that the human capital plays a significant role in the decision to adopt more efficient irrigation technology. In this context, Adesina and Zinnah (1993) and Getnet and MacAlister (2012) emphasize the impact of farmer’s perception of the innovation-related characteristics of the technology in making a decision on adoption.

The literature on the determinants of the adoption of water-efficient technologies by small farmers in Asia is mixed. He et al. (2007) found that Chinese farmers’ educational background, active labor force size, contact with extension services, access to credit and other assistance, technical training received and diversity of irrigated crops grown are positive and significant factors affecting adoption while farmers’ age and the distance from water storage tanks to farmers’ dwellings were significant and negative factors correlated with rainwater harvesting and supplemental irrigation technology. Namara et al. (2007), on the other hand, identified access to ground water, cropping pattern, availability of cash, level of education, and the social and poverty status of the farmers as the most important determinants of micro-irrigation technology adoption in India.

The available literature provides some evidence on the technical and financial viability of RWH technologies, particularly in Africa (Oweis et al., 2001; Ngigi, 2003; Yuan et al., 2003; Oweis and Hachum, 2006; Sharma et al., 2010). Moges et al. (2011) indicate that the potential of RWH systems to increase agricultural production in Ethiopia is site-specific, depending on bio-physical, institutional and socio-economic conditions as well as household-specific conditions. A cost-benefit analysis of on-farm rainwater storage systems for supplemental irrigation in Kenya has shown that it provides a valuable solution to crop failure in semi-arid areas (Ngigi et al., 2005). Similarly, Malesu et al. (2006) illustrates how rainwater-harvesting technologies help farmers adjust to land use change and mitigate water scarcity. Authors such as Mutekwa and Kusangaya (2006) reinforce the idea that RWH technologies are suitable for semi-arid areas if these technologies are tailored to local conditions. We draw from these studies in examining the impact of RWH adoption on farm incomes in Nepal because this is a relatively less researched issue in South Asia.

3. Study Area and Sampling

We conducted the study in four mid-hill districts of Nepal, namely, Makwanpur from the Central Development Region and Palpa, Gulmi and Syangja from the Western Development Region. We chose these four districts for two main reasons: firstly, they have recorded the highest rates6 of individually-managed RWH technology adoption and, secondly, they allow us to capture variation in rainfall and elevation across the hilly districts of Nepal. Rainfed agriculture is predominantly practiced in these areas and is associated with the cultivation of major staple crops such as maize, wheat, rice, millet and some vegetables. We selected six Village Development Committees (VDCs)7 from Makwanpur, four VDCs from Palpa, three VDCs from Gulmi and two VDCs from the Syangja district. Figure 1 gives the location of the study sites.

We carried out the study primarily between August and November, 2012 through a semi-structured survey interview of 282 farm households comprising 141 RWH adopters and 141 non-adopters.8 We used a multistage sampling

5 Optimizing utility may also include considerations such as health benefits, environmental concerns, food security and risk (Ribaudo, 1998;

Napier et al., 2000).6 We used the National Population and Housing Census Report (2011), District Profile (2011) and the District Agriculture Development

Office Report (2011) to gather secondary information, including information on the RWH adopters’ list, household population size, and the

occupational diversity of the households living in the selected villages in order to develop the sampling frame. We obtained the list of villages

and adopters from the 2011 Census of Nepal.7 A Village Development Committee (VDC) is the lowest-level administrative unit in Nepal comprising small villages (wards).8 We define a farm household as one where a group of individuals related by blood or marriage live on the same premises, share a kitchen

and practice agriculture farming system.


technique where we first selected four districts from two regions. We then selected VDCs from each district based on the RWH technology adoption rates. Secondly, we stratified farmers in each VDC into two groups, namely, adopters and non-adopters. We identified adopters in each sampled village with the help of the District Agriculture Development Office (DADO). In each sampled village, there were fewer numbers of RWH technology adopters than non-adopters.9 We oversampled the RWH technology adopters such that the proportion of adopters and non-adopters is the same in our sample and we applied the probability proportion to size technique to ensure that farmers in the large village clusters had the same probability of getting into the sample as those in the smaller village clusters. We followed sampling procedure discussed in Qaim et al. (2006) and Sadashivappa and Qaim (2009) in Bt cotton sampling process in India. We sampled at least 10 to 15 households10 from each VDC among the population of individually-managed RWH adopters and non-adopters.

We also conducted three community-level Focus Group Discussions (FGDs) in each district. Around 5 to 10 RWH adopters/non-adopters participated in each FGD (including farmers from different castes, genders and economic backgrounds). The FGDs provided qualitative information for understanding farmers’ perspective on adopting the RWH technology, and we consider those information while designing the survey instrument.

4. Methods and Variables

In order to better understand farmers' adoption decisions with regard to the RWH technology and to assess its impact on farmers' income, we specify and estimate a set of econometric equations. Household income and adoption may affect each other, where household with higher income may adopt RWH technology and adoption of RWH technology may also increase farm income. The problem can be resolved technically using either a treatment-effects model as in Maddala (1983) and Stata Corporation (2011) or an instrumental variable (IV) approach with instruments for an endogenous RWH adoption dummy, as in Angrist (2000). We estimate the equations describing farmers' adoption decision and farm income simultaneously using a treatment-effects model to control for self-selection (Heckman 1978; Heckman, 1979; Heckman and Navarro-Lozano, 2003).

4.1 Modeling of Farmers' Adoption DecisionThe decision to adopt an available agricultural technology depends on a variety of factors including farm households’ asset bundles and socio-economic characteristics, characteristics of the technology, perception of need, and the risk-bearing capacity of the household (Wiersum, 1994; Mendola, 2005; He et al., 2005; Calatrava-Leyva et al., 2005; Namara et al., 2007; Adeoti, 2009; Abdulai et al., 2011; Getnet and MacAlister, 2012). An ‘asset bundle’ comprises physical, natural, human, social, institutional, technical and financial assets.

In this study, we assume that farmers are risk-neutral and that their adoption decision is based on the comparison of their expected profit with and without the adoption of the RWH technology:

d*i E(P1

i) - E(P0i)>0

Where the latent variable, , is not observed, and P1i and P0

i are the profit with and without the adoption of the RWH technology respectively. We assume that farm i’s expected net benefit from adopting the RWH technology can be modeled as follows: d*

i= X'ig + ei , where the vector Xi includes characteristics of the farmer and the farm’s

environment. The decision model for farmer i is thus written as

d*i= X'

ig + ei , .

We use the probability that farmer i adopts the RWH technology using the following Probit model:

di= F + vi ,

9 We found 20 to 120 RWH adopters and 100 to 410 non-adopters in each sample village.10 We selected 10 to 20 households randomly from each VDC where 5-10 were RWH adopting HHs (Treatment HHs) and 5-7 were RWH non-

adopters/participants (Control HHs).

( ) ( )* 1 0 0i i id E E≡ P - P >

*id

* ' 0i i id g e= + ≥X (2)

( )'i i id F g ν= +X (3)

(1)

5


Where, di equals 1 if the expected net benefit di* is positive and 0 otherwise. Function F is the cumulative

distribution of the ei error term, assumed to be standard normal.

Based on the review of relevant literature, we use the following explanatory variables in the adoption model: age, gender and education of household head, number of economically-active members in household, share of upland plots in total cultivated land, agriculture and livestock production-related training received by the household, annual household off-farm income, poverty status11, and district fixed effects (see Table 1 for a detailed description of these variables).

4.2 Income ModelWe use the following model to estimate the impact of the RWH technology on annual household income from the agriculture and livestock sectors:

In INCi = b0 + b1 .Livestock _ Uniti + b2 .In Spade + b3 .In Totallandi + b4 .In Shareuplandi b5 .RWH_Technologyi + ld + wi

where In INCi is the log of annual household income from agriculture and livestock sectors12 (measured in Nepalese rupees), Livestock_unit is Livestock Standard Unit (LSU), lnSpade is the number of spades at home (in natural log), lnTotalland is the total cultivated land (ropani13; in natural log), lnshupland is the share of upland in total cultivated land (percentage; in natural log), RWH_Technology refers to the adoption of the RWH pond (see Table1for more details on these variables), and ld refers to district fixed effects.

We assumed that The RWH technology adoption is determined by gender of the household head, education of the household head, percent share of upland plots in total cultivated land, agriculture and livestock production-related training received by the household whereas age of the household head, annual household off-farm income and poverty status are assumed negatively determined on RWH adoption. District fixed effects is used to capture topographic features in the study area.

Nine explanatory variables are used to explain the adoption of RWH technology: Age (of the household head in years), Gender (of the household head (1 if male, 0-otherwise), Education (years of schooling of household head), Economically active (total household members aged 15 to 60), Shareupland (percent of upland in total cultivated land), Training (received on agriculture and livestock production, 1 if yes, 0 otherwise), lnIncomeofffarm (annual off-farm income (NRs.) in natural log), and Poor (a dummy variable to indicate if the household is poor).

4.3 Controlling for self-selectionWe need to control for self-selection of farmers into the group of adopters because the decision to adopt the RWH technology is not random. Equation (3) which describes the adoption decision and Equation (4) which describes farmer income are estimated simultaneously. We estimate the treatment-effects model using the Maximum Likelihood approach.14 In order to correct the bias induced by over-sampling one group of farmers, we estimate the model using the weighted endogenous sampling maximum likelihood estimator (see Greene, 2003 for more details).15

(4)

11 Poverty status is based on household income per capita per day.12 The Income variable used as the dependent variable in the main model included agricultural and livestock income, but did not include

off-farm income.13 1 ropani = 0.05 hectare (19.67 ropani = 1 hectare)14 The treatment-effects model estimates the effect of an endogenous binary treatment, RWH Technology here, on a continuous variable

lnINC, (i.e., annual income from the agriculture and livestock sector), conditional on a set of independent variables.15 We have used appropriate weights when estimating the likelihood function. The weights correct for the bias induced by the oversampling

of adopters. We calculate them by combining the proportion of adopters in the sample and the true proportion of adopters in each village.


Since household income and adoption may affect each other, where household with higher income may adopt RWH technology and adoption of RWH technology may also increase farm income, resulting into endogeneity. We, therefore, estimate the income equation using an instrumental variable (IV) approach. We use training received on agriculture and livestock production as an instrument for RWH technology adoption. The main identification assumption here is that the adoption of RWH may depend on any training that farmers receive but training may not be dependent on annual farm income of the household. This IV approach is to corroborate the results obtained from the treatment-effects model.

5. Results and Discussion

5.1 Descriptive Statistics

5.1.1 Characteristics of Rainwater Harvesting Technology

Among the 141 RWH adopters, approximately 82 percent had constructed a plastic pond (Kachi) with only 18 percent having cement ponds. The water-holding capacity of each pond ranged from 1000 liters to 75,000 liters. Nearly half of the RWH respondents had received a subsidy from the government (i.e., DADO) and/or non-governmental organizations (NGOs) (mainly materials like plastic, pipe or cement) ranging between 30 and 50 percent. Table 2 gives the main characteristics of the RWH ponds.

Approximately 94 percent of the RWH ponds are located near the homestead of the adopter. While a majority of the adopters used only rainfall (45.4 percent), 30.7 percent used both rainfall and stream water in their RWH ponds. Approximately 97 percent of adopters stated that they installed a RWH pond for producing vegetable and high-value crops. We found most (93 percent) of the RWH ponds to be actively used though 7 percent were inactive due to seepage and weeds management problems in plastic ponds. A majority (85 percent) of RWH adopters reported that their cropping pattern had changed after RWH pond construction. About 25 percent of respondents reported problems with RWH pond management and water holding capacity of pond due to seepage and weeds. 75 percent of adopters used RWH pond water in the field for irrigation purposes through a pipe-line connection. The age of the RWH ponds varied from 1 to 14 years. Table 3 gives details of RWH pond location, water source and distribution, and purpose and status of the technology.

5.1.2 Socio-demographic and Economic Characteristics

Table 4 presents the socio-demographic and economic characteristics of the sampled farmers. The average age of the household head is approximately 50 years. Though RWH adopters are on average younger than non-adopters, the difference (between the two means) is not statistically significant. Approximately 74 percent of the household heads are male with average education of 4.5 years of schooling. The level of education is higher for adopters than for the non-adopters. Among other variables, the total number of spades in the house, the training received with regard to agriculture and livestock production and the membership of the household head in any group, organization or cooperative were found to be greater among RWH adopters than among non-adopters. There was a greater proportion of higher caste households and those with knowledge of climate change among the RWH adopters than among the non-adopters. The percentage of people below poverty line is lower for adopters (34 percent) than for non-adopters (45 percent). This indicates that there is a positive associate between RWH Technology adoption and households poverty. The difference between the sample means in the adopter and non-adopter groups is not statistically significant for the following variables: the livestock standard unit, the availability of extension services at the farm, access to credit, total cultivated land, and percent of upland in total cultivated land. The latter indicates that the land holdings are similar in the two groups of farmers.

7


We found approximately 27 percent of the respondents to grow cauliflower and cabbage. The proportion of farmers growing cauliflower and cabbage16 was higher among the RWH adopters (39 percent) than among the non-adopters (22 percent). The same pattern was discernible for tomatoes, which were grown more frequently by adopters (55 percent) than by non-adopters (27 percent) as well as for other vegetable such as bean, pea, broadleaf mustard, and guard. For the latter, the proportion of farmers growing them in the adopter group was 71 percent while it was 61 percent in the non-adopter group. We also found approximately 43 percent and 78 percent respondents to grow rice and maize/wheat at their farms respectively. On the other hand, cereal production was more common among the non-adopters than among the adopters. We thus noticed that most RWH adopters were growing a high-value crop (i.e., vegetable) and that they had diversified their cropping pattern towards vegetables from cereals in contrast with the non-adopters. Figure 2 gives the frequency of different types of vegetables and cereals produced by the RWH adopters and non-adopters.

Table 4 also presents the major crops produced and the revenue from the marketed crops in both the lowland and the upland during the different cropping seasons. The production of cereal crops (mainly, rice, maize, wheat) in upland and lowland areas is not statistically significant. With regard to farm revenue from different crops, revenue from tomato and other vegetables is significantly higher among the RWH adopters.

We found income from vegetable and fruit, and from agriculture and livestock sectors as well as total annual household income to be significantly higher among RWH adopters than non-adopters. However, income from off-farm activities was significantly higher for non-adopters. The annual income from the agriculture and livestock sectors for RWH adopters was NRs. 104,969 (USD 1,049)17 while it was just NRs. 53,876 (USD 538) for non-adopters. The RWH technology adopters thus appeared to benefit from an increased supply of irrigation water which allowed them to diversify their cropping system from cereal crops to high-value vegetable crops.

5.2 Determination of Farm Income and Farmers’ Decision to Adopt RWH TechnologyWe estimated six different models to identify the impact of RWH adoption on farm income: three treatment-effects models (see Table 5) two instrumental variable (IV) models, and one OLS model (see Appendix 2). The first treatment-effects model is the most basic model. Poor and off-farm income variables are included in second model whereas district fixed effects are included in third model. In case of IV approach, model 2 is basic model. We include the district fixed effects in third model. The IV and treatment-effects models produce results that are very similar. The OLS model underestimates the effect of RWH adoption on farm income (45.8 percent) while the treatment-effects models and IV models indicate that RWH adoption results into 111 percent to 138 percent increase in an average household’s farm income. We comment on the estimated coefficients of the third treatment-effects model that also controls for district fixed effects (see Table 5).

Results indicate that RWH adoption (RWH Technology) significantly increases household income from agriculture and livestock. RWH technology adopters have earned around 131 percent more annual household income from the agriculture and livestock sectors than the non-adopters in the study area. The livestock standard unit (Livestock standard unit) and the total cultivated land (lnTotalland) also contribute positively to household income.

The most important factor affecting farmers’ adoption decision is agriculture and livestock training (Training). The age of the household head (Age), annual household income from off-farm activities (lnIncomeofffarm) and poverty status (Poor) have significant but negative impacts on RWH adoption whereas the training received by farmer and the gender of the household head (i.e., being a male) have significantly positive impacts on the RWH adoption decision. However, other variables like the economically active members in the household (Economic active HH),

16 Cabbage and cauliflower require regular irrigation, particularly on sandy soils. If leaves begin to wilt mid-day, plants are moisture stressed.

Plants that wilt intermittently produce smaller yield while plants that wilt frequently (Delahaut and Kewenhouse, 1997). The estimated daily

irrigation water requirement per plant of cauliflower is 0.74 liter at early stage and 1.35 liters at the development stage whereas irrigation

water requirement per plant of cabbage is 1.17 liters during early stage and 1.66 liters during peak growth stage (NCPAH, 2014).17 USD 1 = NRs. 100


the percent shared by upland in total cultivated land (Shareupland), and education of the household head had no impact on the adoption decision.

All other factors remaining constant, if a farmer receives agriculture and livestock production related training, the adoption decision vis-à-vis RWH technology will increase by 28.5 percent. The training provides knowledge and skill to farmers to adopt innovative technology at the farm. Off-farm income too has a significant and negative impact on the RWH adoption decision. The probability of a farmer who is considered as poor adopting RWH technology is 18.7 percent lower than that for a relatively better off farmer adopting such technology because, in Nepal, poverty is a major hurdle in the way of adopting new technology at the farm. More than 25 percent of the people live below the poverty line in Nepal (see Model 3 in Table 5).

5.3 Cost Benefit Analysis of RWH Pond AdoptionTable 6 presents the cost-benefit analysis for the RWH plastic pond. We chose plastic ponds as they are cheaper and more popular than cement-ponds. We calculate the annual benefit and cash flow for a 10-year period a plastic pond functions well up to 10 years. We estimate the Benefit-Cost (B/C)18 ratio, Net Present Value (NPV), Internal Rate of Return (IRR) and Pay Back Period (PBP) using the standard discount rate in Nepal of 12 percent19 in financial analyses. We perform sensitivity analysis to check the robustness of the results.

Cost calculations are based on the total investment cost including labor, plastic, other equipment and the opportunity cost of the land on which the pond is constructed. The initial average investment cost to construct a RWH plastic pond on 3 ropani of upland area (with a water-holding capacity of approximately 45,000 liters) is NRs. 55,372 (USD 543). In addition, there is a maintenance cost of NRs. 913 per year per household based on survey information on labor costs required.

We estimate the incremental income from RWH pond construction and crop diversification from the second year onwards. We use the benefits estimates of using RWH technology from our model (Table 5, Model 3). Thus, we estimate that the average farmer using RWH technology obtains an annual incremental benefit of NRs. 69,456 (USD 700). This number reflects the benefits a farmer receives from using this technology relative to farmers who do not.

Our calculations suggest that the NPV of investing in a RWH pond with a carrying capacity of 45000 liters to a farmer is NRs. 276,649 (USD 2,766), assuming a12 percent discount rate. The B/C ratio is 6.1 and IRR is 34 percent. The Pay Back Period is approximately two years, which indicates the time required for the repayment of the initial investment is rather low. Sensitivity analysis indicates that the investment is viable even if the investment costs increase by 20 percent or benefits decrease by 20 percent.

Although RWH Technology is very profitable, findings from stakeholder meetings show that a majority of the farmers are not adopting this technology. This is because of lack of technical knowledge, large start-up cost of NRs. 55,000 (28 percent of total annual household income), limited subsidies and lack of labor in the villages to construct pond (due to massive out-migration of adult population for better job opportunity). Additionally, farming is not commercialized and most farmers do not like to shift from cereal based farming system to high value crops due to production and market risks.

Our analysis indicates that the training received by farmers on farm management, agriculture and livestock production helps to increase RWH adoption. The cost of a three-day village level training package is about NRs.

18 We calculate the B/C ratio and the NPV using the following equations:

19 ADB (2013) proposes a higher discount rate (about 12%) for cost benefit analysis in developing countries considering higher production and

market risks and uncertainty to introduce new technologies at farm.

∑+

∑+

=

=

t

tt

t

t

tt

t

r

rC

B

1

1

)1(

)1(

∑+=

-T

tt

tt

rCostBenefit

1 )1(

B/C Ratio =

NPV =

9


3,000 (USD 30) per farmer (including 30 percent organizational overhead costs). With training, the probability of RWH technology adoption increases by approximately 29 percent. Thus, in any district in our study area, if approximately 10% of households, i.e. some 7000 households, are trained, then we can expect 2030 trainees to adopt RWH (see Table 7). The net annual public benefits from training are NRs. 66,457 per farmer. Thus, per district the annual public benefits from providing training to10 percent of farm households is expected to be approximately NRs. 130,907,710 or about 1.3 million USD as a result of some subset of farmers adopting RWH. These benefits, however, would require substantial investment in building the tanks (Table 7).

6. Conclusions and Policy Implications

In this study, we examine the role of RWH technology adoption in four districts of Nepal. Our findings indicate that adoption of RWG increases annual household income from the agriculture and livestock sectors by over 130 percent. The technology allows the farmers to diversify their cropping system from cereal crops to high-value vegetables. We estimate that the net present value to farmers from investing in RWH technology is NRs. 276,649 (USD 2,766), using a discount rate of 12 percent. Sensitivity analysis indicates that this investment is viable even if investment costs increase by 20 percent or benefits decrease by 20 percent.

Although RWH is profitable in rainfed agriculture systems, a majority of farmers seem reluctant to adopt this technology. This is because of lack of technical knowledge, large start-up costs of NRs. 55,000 (USD 550), which amount to 28 percent of annual household income and lack of labor in the villages due to out-migration. Furthermore, without diversifying farm production, this technology is less beneficial. Thus, risk-averse farmers may be less willing to adopt RWH.

Our study suggests that at least some of the constraints to adoption can be reduced with farmer training. If a farmer receives farm management training from agricultural extension services, the probability of RWH technology adoption increases by approximately 29 percent. In an average district, if some ten percent of farmers or 7000 farm households receive training, we estimate that the net annual public benefits from RWH adoption would be approximately NRs. 130,907,710 or about 1.3 million USD. Thus, policy makers and extension services need to play a more proactive role in promoting RWH technology in the rainfed hilly regions of Nepal.

Acknowledgements

The research was supported financially and technically by the South Asian Network for Development and Environmental Economics (SANDEE), Nepal. I am grateful to Celine Nauges, Mani Nepal, Priya Shyamsundar, Pranab Mukhopadhyay, Subhrendu Pattanayak, Enamul Haque, Punya Prasad Regmi and Jay Prakash Dutta for their critical comments and encouragement. I would also like to acknowledge the comments and suggestions that I received from SANDEE advisors and associates at various SANDEE Research and Training Workshops. I am thankful to Anuradha Kafle, Neesha Pradhan and Malvika Joshi for providing the necessary administrative support during the study. I would like to thank the editors and anonymous reviewers for their valuable comments and suggestions that have improved both the analysis and this working paper. Last but not least, I am deeply indebted to the farmers in my study area, who are too numerous to mention individually, but without whose cooperation this study would not have been possible. Errors, if any, are entirely my own.


References

Abdulai, A., V. Owusu, and John-Eudes A. Bakang (2011), ‘Adoption of safer irrigation technologies and cropping patterns: Evidence from Southern Ghana’, Ecological Economics 70(7): 1415-1423.

ADB (2013), ‘Cost benefit analysis for development: a practical guideline’, Mandaluyong City, Philippines, Asian Development Bank, 2013.

Adeoti, A. I. (2009), ‘Factors influencing irrigation technology adoption and its impact on household poverty in Ghana’, Journal of Agriculture and Rural Development in the Tropics and Subtropics 109(1): 51-63.

Adesina, A. and M.M. Zinnah (1993), ‘Technology characteristics, farmers’ perceptions and adoption decisions: A Tobit model application in Sierra Leone’, Agricultural Economics 9: 297-311.

Angrist, J. (2000), ‘Estimating of limited dependent variable models with dummy endogenous regression: simple strategies for empirical practices’, Working Paper No. 248, National Bureau of Economic Research, Massachusetts Avenue Cambridge, United States.

APN (2004), ‘Water resources in South Asia: an assessment of climate change-associated vulnerabilities and coping mechanisms’, Asia-Pacific Network for Global Change Research (APN), [Available at] http://www.apn-gcr.org/resources/index.php/items/show/1494 [Accessed on: 15 January, 2012].

Asfawa, A. and A. Admassie (2004), ‘The role of education on the adoption of chemical fertilizer under different socio-economic environments in Ethiopia’, Agricultural Economics 30: 215-228.

Bartlett, R., L. Bharati, D. Pant, H. Hosterman, and P. McCormick (2010), ‘Climate change impacts and adaptation in Nepal’, Working Paper 139, International Water Management Institute (IWMI), Colombo, Sri Lanka. [Available at] http://www.iwmi.cgiar.org/Publications/Working_Papers/working/WOR139.pdf, [Accessed on: 10 December, 2012].

Bekele, W. and L. Drake (2003), ‘Soil and water conservation decision behavior of subsistence farmers in the Eastern Highlands of Ethiopia: a case study of the Hunde-Lafto area’, Ecological Economics 46: 437-451.

Besley, T. and A. Case (1993), ‘Modeling technology adoption in developing countries,’ American Economic Review 83: 396–402.

Calatrava-Leyva, J., J. Augustin Franco, and C. Gonzalez-Roa (2005), ‘Adoption of soil conservation practices in olive groves: The case of Spanish mountainous areas’, Paper prepared for presentation at the XIthInternational Congress of the European Association of Agricultural Economists.

CBS (2012), Nepal Population and Housing Census 2011 (National Report), Government of Nepal, National Planning Commission, Central Bureau of Statistics, Kathmandu, Nepal.

CIA (2010), The World Factbook, Central Intelligence Agency, [Available at] http://www.cia.gov/library/publications/the-world-factbook/goes/np.html [Accessed on: 15 February, 2013].

Cline, W. R. (2007a), ‘Global warming and agriculture: impact estimates by country’, Centre for Global Development and the Peterson Institute for International Economics, Washington DC.

Cosbey, A. (2009), ‘Developing countries’ interests in climate change action and the implications for the post-2012 climate change regime’, United Nations Conference on Trade and Development (UNCTAD), United Nations' Organization, New York and Geneva, 2009.

Cruz, R.V., H. Harasawa, M. Lal, S. Wu, Y. Anokhin, B. Punsalmaa, Y. Honda, M. Jafari, C. Li and N. HuuNinh (2007), ‘Asia, climate change 2007: impacts, adaptation and vulnerability’, Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, in M. L. Parry, O. F. Canziani, J. P. Palutikof, P.J. van der Linden and C.E. Hanson (eds.), Cambridge, UK: Cambridge University Press, pp. 469–506.

Delahaut, A. K. and A. C. Kewenhouse (1997), ‘Growing Broccoli, Cauliflower, Cabbage and Other Crops in Wisconsin: A Guide for Fresh-Market Growers’, University of Wisconsin-System Broad of Regents and University of Wisconsin-Extension, Cooperative Extension, pp. 9.

Dinar, A. and D. Yaron (1990), ‘Influence of quality and scarcity of inputs on the adoption of modern irrigation technologies’, Western Journal of Agricultural Economics 15(2): 224-233.

11


Dinar, A., M. B. Campbell, and D. Zilberman (1992),‘Adoption of improved irrigation and drainage reduction technologies under limiting environmental conditions’, Environmental and Resource Economics 2:360-73.

Dridi, C. and M. Khanna (2005), ‘Irrigation technology adoption and gains from water trading under asymmetric information’, American Journal of Agricultural Economics 87: 289-301.

Feder, G., R.E. Just and D. Zilberman (1985), ‘Adoption of agricultural innovations in developing countries: a survey’, Economic Development and Cultural Change 33: 255-298.

Gentle, P. and T. N. Maraseni, (2012), ‘Climate change, poverty and livelihoods: adaptation practices by rural mountain communities in Nepal’, Environmental Science and Policy 21: 24-34.

Getnet, K. and C. MacAlister (2012), ‘Integrated innovations and recommendation domains: Paradigm for developing, scaling-out, and targeting rainwater management innovations’, Ecological Economics 76: 34-41.

Gardebroek, C. (2002), ‘Farm specific factors affecting the choice between conventional and organic dairy farming’, Proceedings of the 10th EAAE Congress on Exploring Diversity in the European Agriculture Food System, August 28-31, 2002, Zaragoza (Spain).

Greene, W.H. (2003), Econometric Analysis, 5th Edition, Upper Saddle River, New Jersey, USA: Prentice Hall Publication.

Gurung, G.B. and D. Bhandari (2009), ‘Integrated approach to climate change adoption’, Journal of Forest and Livelihood 8(1): 91-99.

Heckman, J. J. (1978). Dummy endogenous variables in a simultaneous equation system, Econometrica 46(6): 931–959.

Heckman, J. (1979). Sample selection bias as a specification error, Econometrica 47: 153-61.

Heckman, J. and Navarro-Lozano, and S. (2003). Using matching, instrumental variables and control functions to estimate economic choice models, Working Paper, IFAU - Institute for Labour Market Policy Evaluation, No. 2003:4.

Hatibu, N., K. Mutabazi, E.M. Senkondo, and A.S.K. Msangi (2006), ‘Economics of rainwater harvesting for crop enterprises in semi-arid areas of East Africa’, Agricultural Water Management 80(1-3): 74-86.

Hattam, C. (2006), ‘Adopting organic agriculture: an Investigation using the theory of planned behaviour’, Proceedings of the International Association of Agricultural Economics Conference, August 12-19, 2006, Gold Coast, Australia.

He, Xue-Feng, H. Cao, and Feng-Min Li (2007), ‘Economics analysis of the determinants of adoption of rainwater harvesting and supplementary irrigation technology (RHSIT) in the semi-arid Loess Plateau of China’, Agricultural Water Management 89: 243-250.

HMGN (2004), ‘Initial national communication to the COP of UNFCCC’, Ministry of Population and Environment, Government of Nepal, Kathmandu.

IFAD (2008), ‘Climate change and the future of smallholder agriculture; How can rural poor people be a part of the solution to climate change?’, Discussion paper prepared for the Round Table on Climate Change at the 31st session of IFAD’s Governing Council, 14 February 2008, IFAD Policy Reference Group on Climate Change, International Fund for Agricultural Development, Rome, Italy.

IPCC (2007a), ‘Summary for Policymakers in Climate Change 2007: Impacts, Adaptation and Vulnerability’, Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, in M.L. Parry, O.F. Canziani, J.P. Palutikof, P.J. van der Linden and C. E. Hanson (eds.), Cambridge University Press, Cambridge, UK, 7-22.

IPCC (2014), ‘Climate Change 2014: Impacts, Adaptation, and Vulnerability, Part B:Regional Aspects’, Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, in V.R. Barros, C.B. Field, D.J. Dokken, M.D. Mastrandrea, K.J. Mach, T.E. Bilir, M. Chatterjee, K.L. Ebi, Y.O. Estrada, R.C. Genova, B. Girma, E.S. Kissel, A.N. Levy, S. MacCracken, P.R. Mastrandrea, and L.L. White (eds.), Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 5-10.

IPCC, (2014), ‘Summary for Policymakers, In: Climate Change 2014, Mitigation of Climate Change’, Contribution of Working Group III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, in O. Edenhofer, R. Pichs- Madruga, Y. Sokona, E. Farahani, S. Kadner, K. Seyboth, A. Adler, I. Baum, S. Brunner, P. Eickemeier, B.


Kriemann, J. Savolainen, S. Schlömer, C. von Stechow, T. Zwickel, and J.C. Minx (eds.), Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 4-31.

Jensen, R. (1982), ‘Adoption and diffusion of an innovation of uncertain profitability’, Journal of Economic Theory 27: 182–92.

Just, R.E. and D. Zilberman (1983), ‘Stochastic structure, farm size and technology adoption in developing agriculture’, Oxford Economic Papers 35: 307-328.

Kohler, T., M. Giger, H. Hurni, C. Ott, U. Wiesmann, S. W. von Dach, and D. Maselli (2010), ‘Mountains and climate change: a global concern’, Mountain Research and Development 30(1): 53–55.

Koundouri, P., C. Nauges, and V. Tzouvelekas (2006),‘Technology adoption under production uncertainty: theory and application to irrigation technology’, American Journal of Agricultural Econonomics 88 (3): 657-670.

Lohani, S.N. (2007), ‘Climate change in Nepal: shall we wait until bitter consequences?, Agriculture and Environment 8: 38–45.

Lynne, G. D., J. S. Shonkwiller and L. R. Rola (1988), ‘Attitudes and farmer conservation behavior’, American Journal of Agricultural Economics 70: 12-19.

Maddala, G. (1983), Limited Dependent and Qualitative Variables in Econometrics’, Cambridge University Press, Cambridge.

Malesu, M., B. Mati and A. Oduor (2006), ‘Promoting rainwater harvesting in eastern and southern Africa’, Draft ICRAF Working Report, International Council for Research in Agroforestry (ICRAF), Nairobi, Kenya.

Malla, G. (2008), ‘Climate change and its impact on Nepalese agriculture’, The Journal of Agriculture and Environment 9(7): 62-71.

MallaThakuri, M. B. (2011), ‘Revisiting the need for improved stoves: Estimating health, time and carbon benefits’, in A. K. Enamul Haque, M. N. Murty and P. Shayamsundar (eds.), Environmental valuation in South Asia, Cambridge University Press, pp. 348-374.

Mauceri, M., J. Alwang, G. Norton and V. Barrera (2005), 'Adoption of integrated pest management technologies: a case study of potato farmers in Carchi, Ecuador’, Proceedings of the Agricultural Economics Association Annual Meeting, Providence, Rhode Island.

Mendola, M. (2005), ‘Agricultural technology and poverty reduction: a micro-level analysis of causal effects’, Working Paper No. 2005-14, Universita degli Studi di Milano Bicocca, Milan, Italy.

Moges, G., H. Hengsdijk, and H.C. Jansen (2011), ‘Review and quantitative assessment of ex situ household rainwater harvesting system in Ethiopia’, Agricultural Water Management 98: 1215-1227.

MOI (2014), ‘Irrigation Policy, 2013’, Ministry of Irrigation, Government of Nepal, Official Site, [Available at] http://www.moir.gov.np/pdf_files/Irrigation-Policy-2070-final.pdf [Accessed on: 5 October, 2014].

Murgor, F.A., J.O. Owino, G.J. Cheserek and C.K. Saina (2013), ‘Factor influencing farmers’ decisions to adopt rain water harvesting technologies in Keiyo district, Kenya’, Journal of Emerging Trends in Economics and Management Science 4(2): 133-139.

Mutekwa, V. and S. Kusangaya (2006), ‘Contribution of rainwater harvesting technologies to rural livelihoods in Zimbabwe: the case of Ngundu ward in Chivi district’, Water SA 3(3): 437-444.

Namara, R.E., R.K. Nagar and B. Upadhyay (2007), ‘Economics, adoption determinants, and impacts of micro-irrigation technologies: empirical results from India’, Irrigation Science 25: 283-297.

Napier, T. L., M. Tucker and S. McCater (2000), ‘Adoption of conservation tillage production systems in three Midwest watersheds’, Journal of Soil and Water Conservation 53: 123-134.

NCPAH (2014), ‘National Committee on Plasticulture Application in Horticulture’, Ministry of Agriculture, Government of Inida, Official Site, [Available at] http://www.ncpahindia.com [Accessed on: 13 November, 2014].

NDRI (2009), ‘Project evaluation report on rainwater harvesting system in Makwanpur district for Plan Nepal’, Nepal Development and Research Institute (NDRI), Kathmandu, Nepal.

13


Ngigi, S.N. (2003), ‘Rainwater harvesting for improved food security: Promising technologies in the Greater Horn of Africa’, The Greater Horn of Africa Rainwater Partnership (GHARP), Kenya Rainwater Association (KRA), Nairobi, Kenya.

Ngigi, S.N., H.H.G. Savenije, J. Rockstrom and C.K. Gachene (2005), ‘Hydro-economic evaluation of rainwater harvesting and management technologies: farmers’ investment options and risks in semi-arid Laikipia district of Kenya’, Physics and Chemistry of the Earth, Parts A/B/C 30(11-16): 772-782.

Oweis T., P. Prinz and A. Hachum (2001), ‘Water harvesting: indigenous knowledge for the future of the drier environments’, International Centre for Agricultural Research in the Dry Areas (ICARDA), Aleppo, Syria.

Oweis, T. and A. Hachum (2006), ‘Water harvesting and supplemental irrigation for improved water productivity of dry farming systems in West Asia and North Africa’, Agricultural Water Management 80(1): 57-73.

Qaim, M., Subramanian, A., Naik, G., and D. Zilberman (2006), ‘Adoption of Bt cotton and impact variability: Insights from India’, Review of Agricultural Economics 28(1): 48-58.

Regmi, H.R. (2007), ‘Effect of unusual weather on cereal crop production and household food security’, Journal of Agriculture and Environment 8: 20–29.

Ribaudo, M.O. (1998), ‘Lessons learned about the performance of USDA Agricultural non-point source pollution programs’, Journal of Soil and Water Conservation 53: 4-10.

Sadashivappa, P. and M. Qaim (2009), ‘Bt cotton in India: Development of benefits and the role of government seed price interventions’, AgBioForum 12(2): 172-183. [Available at] http://www.agbioforum.org [Accessed on: 20 October, 2014].

Sadoulet, E. and A. deJanvry (1995), ‘A quantitative development policy analysis’, Baltimore: The Johns Hopkins University Press.

Sharma, B.R., K.V. Rao, K.P.R. Vittal, Y.S. Ramakrishna, and U. Amarasinghe (2010), ‘Estimating the potential of rainfed agriculture in India: Prospects for water productivity improvements’, Agricultural Water Management 97(1): 23-30.

Stata Corporation (2011), ‘Stata base reference manual release 12’, A Stata Press Publication, StataCorp LP, College Station, Texas.

UNFCCC (2007), ‘Climate change: impacts, vulnerabilities and adaptation in developing countries’, United Nations Framework Convention on Climate Change (UNFCCC), [Available at] http://unfccc.int/resource/docs/publications/impacts.pdf [Accessed on: 20 January, 2014].

Wani, S.P., T.K. Sreedevi, J. Rockström and Y. S. Ramakrishna (2009), ‘Rainfed agriculture: past trends and future prospects’, in S.P. Wani, J. Rockström and T. Oweis (eds.), Rainfed agriculture: unlocking the potential, CAB International 2009, London, UK, pp. 1-35.

Wiersum, K. F. (1994), ‘Farmers’ adoption of contour hedge rows intercropping: A case study from East Indonesia’, Agroforestry System 27:163–182.

World Bank (2006), ‘Managing climate risk: integrating adaptation into World Bank Group Operation’, World Bank Group, Global Environment Facility Program, World Bank, Washington DC.

World Bank (2012), ‘World Development Indicators’, [Available at] http://data.worldbank.org/data-catalog/world-development-indicators?cid=GPD_WDI [Accessed on: 28 April, 2012].

Yuan, T., L. Fengmin and L. Puhai (2003), ‘Economics analysis of rainwater harvesting and irrigation methods, with an example from China’, Agricultural Water Management 60(3): 217-226.

Zomer, R. J., A. Trabucco, M. J. Metzger, M. Wang, K. P. Oli and J. Xu (2014), ‘Projected climate change impacts on spatial distribution of bioclimatic zones and ecoregions within the Kailash Sacred Landscape of China, India, Nepal’, Climatic Change 125: 445-460.


Table 1: Description of main variables

Variables Description Expected Sign Variable Mean

RWH Technology RWH Technology adoption (1 if adoption, 0-otherwise) 0.50

Age Age of household head in years - 50.20

Gender Gender of the household head (1 if male, 0-otherwise) + 0.74

Education Education of household head (years of schooling) + 4.51

Economic active HHTotal household members aged 15 to 60 (economically

active)+ 4.10

ln Total land Total cultivated land in ropani (in natural log) + 1.76

Shareupland Upland in total cultivated land (percentage) + 80.27

TrainingTraining received on agriculture and livestock production (1

if yes, 0 otherwise)+ 0.46

ln Income off farm Annual off-farm income (NRs; in natural log) _ 3.70

ln Spade Number of spades (in natural log) + 1.03

Gulmi Survey district (if Gulmi-1, 0-otherwise) +/- 0.14

Syangja Survey district (if Syangja-1, 0-otherwise) +/- 0.15

Palpa Survey district (if Palpa-1, 0-otherwise) +/- 0.28

District fixed-effects +/-

ln Income farmAnnual household income from agriculture and livestock

(NRs; in natural log) 10.79

Poor Household is poor20 (if poor-1, 0-otherwise) - 0.39

Livestock standard unit Livestock standard units (cattle equivalent in number) + 3.10

Tables

dl

Table 2: Characteristics of rainwater harvesting (RWH) ponds (n=141)

Types of Pond Number Percent Water Holding Capacity (liter)

Cost of Construction (NRs.)

Total Irrigated Command Area(in ropani)

Plastic pond (Kachi)

115 81.6 44,792.31Range: 7200 to

75000

Labor : 9146.15Equipment: 16,134.62Other: 0.00Total: 25,372Subsidy: 30- 50 percent

6.13(0.5 to 30)

Cement constructed (Paki)pond

26 18.4 22875Range: 1000 to

72000

Labor: 5557.14Equipment: 19,000Other: 19,250Total: 35,333Subsidy: 40-60 percent

3.59(0.5 to 8.67)

Notes: Investment for maintenance of RWH pond per year = NRs. 913 (0 to 15000). Figures in parentheses indicate range.

20 A household is considered poor if its income is lower than USD 1.25/person/day.

15


Table 3: RWH pond location, water source, distribution, purpose and status (n=141)

Characteristics Frequency Percent

Where (Location):

Near homestead 133 94.3

In farm far from homestead but in same village 7 4.9

In farm far from homestead in another village 1 0.7

Water Sources:

Only from rainfall 64 45.4

Only from stream 6 4.3

Both rainwater and stream 56 39.7

Drinking water supply 15 10.6

Purpose for:

Production of vegetable and cash crops 137 97.2

Other (rice, wheat production, livestock) 3 2.8

Status:

Active 131 92.9

Non-active 10 7.1

Water Distribution Source:

Pipe 106 75.2

Pipe plus sprinkler 26 18.4

Using bucket 8 5.7

Other 1 0.7

Table 4: Socio-demographic and economic characteristics of RWH adopters and non-adopters

Particular Full Sample (N=282)

RWH Adopters (n=141)

Non-adopters(n=141)

Mean Difference

(t-test)

Socio-demographic Characteristics of RWH Adopters and Non-adopters:

Age of Household Head (in years) 50.21 48.91 51.50 -2.59

Gender of the Household Head (if male=1) 0.74 0.81 0.67 0.13***

Years of Schooling 4.51 5.31 3.69 1.62***

Caste (if higher caste =1)21 0.48 0.53 0.43 0.099*

Family Size 6.48 6.58 6.39 0.19

Economically Active Household Members (15 to 60 years old) 4.08 4.22 3.93 0.29

Upland Cultivated (in ropani) 6.47 7.03 5.91 1.12

Lowland Cultivated (in ropani) 1.94 1.89 1.99 0.10

Total Cultivated Land (in ropani) 8.41 8.93 7.91 1.29

Percent Shared by Upland in Total Land 79.5 80.2 78.7 0.41

Livestock Standard Unit (LSU)22 3.13 3.34 2.93 0.41

Number of Spades (type of physical asset) 4.7 5.2 4.3 2.88***

Extension Service (if yes =1) 0.43 0.46 0.40 0.06

Agriculture and Livestock Production Related Training

Received (if yes=1)

0.46 0.67 0.24 0.43***

Access to Credit (if yes=1) 0.77 0.81 0.74 0.08

Membership in Any Group, Cooperative and/or Organization

(if yes=1)

0.63 0.72 0.55 0.16***

21 Brahmin, Chettri and Takuri are the higher castes in Nepal.22 LSU is Livestock Standard Unit (based on cattle equivalent: 1 cow/cattle= 10 goats/lambs= 4 pigs and = 143 chicken/ducks)


Knowledge of Climate Change (if yes=1) 0.49 0.54 0.43 0.11*

Poor (if yes=1) 0.39 0.34 0.45 0.11*

Crop Production and Revenue for RWH Adopters and Non-adopters:

Production (in qt.)

Cauliflower/Cabbage 20.08(5.08)

22.05(6.41)

14.69 (7.43)

7.35

Tomato 50.39(26.82)

69.82(39.95)

11.04(2.21)

58.77

Other Vegetable 8.33(0.89)

9.76(1.32)

6.66(1.17)

3.09*

Rice 19.93(6.75)

9.78(1.94)

22.61(13.95)

-12.82

Wheat/Maize 17.36(9.04)

28.11(19.93)

8.47(1.22)

19.63

Revenue (in NRs.)

From Cauliflower/Cabbage 36284(7682)

42559(9932)

19252(8562)

23306

From Tomato 49179(5629)

59175(7557)

28945(6423)

30220***

From Other Vegetable 19711(1877)

22647(2833)

16297(2337)

6350*

From Rice 23712(2669)

26379(4546)

20723(2457)

5656

From Maize/Wheat 18373(2109)

19151(2533)

17729(3242)

1421

Annual Household Income from Different Sectors for RWH Adopters and Non-adopters:

Income from Cereal Crops 23671(2050)

23796(2697)

23546(3111)

250

Income from Vegetable and Fruit 44148(4663)

67446(8414)

20850(2960)

46595***

Income from Livestock 11603(1322)

13726(2247)

9480(1379)

4246*

Income from Employment /Services 40604(6532)

38082(9313)

43127(9191)

-5045

Income from Off-farm 34881(4311)

26134(5201)

43627(6815)

-17492**

Income from Foreign Employment 40730(6899)

47368(12313)

34092(6232)

13276

Income from Agriculture and Livestock (cereal + vegetable +

livestock)

79423(5739)

104969(9988)

53876(4812)

51092***

Total Annual Household Income 195640(11201)

216555(17713)

174724(13551)

41831*

Notes: Figures in parentheses are standard deviation. ***, **, and * indicates significance at the 1 percent level, 5 percent level and 10 percent level, respectively.

17


Table 5: Determinants of household income and RWH technology adoption (Maximum likelihood estimates): Treatment-effects models

Variable23 Model 1 Marginal Effect

(dy/dx)

Model 2 Marginal Effect

(dy/dx)

Model 3 Marginal Effect

(dy/dx)ln Income farm: Coefficient Coefficient Coefficient

Livestock standard unit 0.046*(0.025)

0.049**(0.024)

0.053**(0.024)

ln Spade 0.010(0.022)

0.009(0.023)

0.009(0.023)

ln Total land 0.592***(0.078)

0.583***(0.076)

0.543***(0.068)

ln Share upland -0.030(0.023)

-0.031(0.022)

-0.018(0.023)

RWH Technology 1.111***(0.190)

1.380***(0.199)

1.310***(0.212)

_cons 9.303***(0.190)

9.259***(0.171)

9.379***(0.169)

RWH Technology:

Age -0.014**(0.007)

-0.003** -0.016**(0.006)

-0.004** -0.017***(0.007)

-0.004**

Gender 0.506**(0.210)

0.117*** 0.705***(0.187)

0.155*** 0.705***(0.191)

0.154***

Education -0.005(0.042)

-0.006 -0.053(0.037)

-0.013 -0.059(0.041)

-0.015

Economic active HH 0.062(0.040)

0.015 0.053(0.043)

0.013 0.065(0.043)

0.016

Shareupland 0.004*(0.003)

0.001* 0.004*(0.003)

0.001* 0.004(0.003)

0.001

Training 1.103***(0.174)

0.326*** 0.950***(0.178)

0.274*** 0.987***(0.200)

0.285***

ln Income off farm -0.045***(0.015)

-0.011*** -0.048***(0.016)

-0.012***

Poor -0.895***(0.186)

-0.206 *** -0.808***(0.218)

-0.187***

constant -1.486***(0.4419)

-0.822*(0.432)

-0.808***(0.218)

DFE24 No No Yes

/athrho -0.497*** -0.854*** -0.819***-0.189***

-0.6740.827-0.558

/lnsigma -0.196*** -0.154**

rho -0.460 -0.693

sigma 0.821 0.856

lambda -0.378 -0.594

Treatment-effects Model-MLE

Wald test of indep. eqns. (rho = 0): chi2(1) =12.34***Prob> chi2 = 0.0004

Wald test of indep. eqns.(rho = 0): chi2(1) = 28.69***Prob> chi2 = 0.0000

Wald test of indep. eqns.(rho = 0): chi2(1) =22.08***Prob> chi2 = 0.0000

Notes: ***, ** and * indicate significance at 1 percent, 5 percent and 10 percent levels, respectively. dy/dx is probability

(margins) after mfx.

23 We use variables that describe the characteristics of the farmer/farm at the time the adoption decision was made. Some characteristics

have changed after adoption of RWH pond (like agricultural income, physical assets, and livestock units) and were not included in the model

to avoid any endogeneity problems.24 District fixed-effects.


Table 6: Cost benefit analysis of RWH plastic pond for upland crops production

RWH Plastic Pond (44792 liters water holding capacity) in 3 ropani Upland Vegetable Crops Production Scenario

Year Net Incremental Benefits/Income1(NRs.)

Total Cost2

(NRs.)Present Value of IncrementalIncome (at 12 percent) (NRs.)

Present Value of Total Cost (at 12 percent) (NRs.)

1 0 55372 0 55372

2 69457 913 55371 728

3 69457 913 49438 650

4 69457 913 44141 580

5 69457 913 39412 518

6 69457 913 35189 463

7 69457 913 31419 413

8 69457 913 28052 369

9 69457 913 25047 329

10 69457 913 22363 294

Total 330432 59716

B/C Ratio 6.1

NPV at 12 percent (in NRs.) 276649


IRR 33.9 %

Pay Back Period 2 years

If Total Cost Increases by 20 percent:3

B/C Ratio 5.1



IRR 33.2 %

If Total Benefits Decreases by 20 percent:3

B/C Ratio 4.9



IRR 33.1 %

If Total Cost Increases by 20 percent and Benefits Decrease by 20 percent:3

B/C Ratio 4.1



IRR 32.2 %

Notes

1. The incremental benefit from RWH adoption is 131 percent higher (i.e., NRs. 69,456 [or USD 700] per year) for adopters

relative to non-adopters based on Model 3, Table 5.

2. Annual costs include NRs. 25,372 is the construction cost and NRs. 30,000 is the opportunity cost of land for ten years.

We estimate the opportunity cost of land as the foregone revenue to the adopter from not growing crops in the pond land

(i.e., NRs. 3000 per year for 3 ropani irrigating capacity of the pond). The 10-year opportunity cost of land is NRs. 30,000.

Based on the farmer survey, annual maintenance cost is estimated to be labor costs of NRs. 931per household per year.

3. Incremental benefits from RWH technology adoption is 128 percent more relative to non-adopter in 95 percent confidence

intervals. If the estimated coefficient is 128 and the estimated standard error is 21, then the 95 percent confidence interval

is [131 - 2*21; 131 + 2*21] = [89; 173]. This allows us to calculate the cost-benefit ratio (based on an estimated benefit of

131) along with the corresponding 95 percent confidence interval.

19


Table 7: Rainwater harvesting benefits from district level training

a No. Trained per district (approx. 10% of population) 7000

b No. of Trainees who adopt RWH (29% of a) 2030

c Cost per trainee to government (NRs.) 3000

d Average annual Benefit to RWH adoptee (NRs.) from Table 6 69457

e Net annual Benefit from training to RWH adoptee (NRs.) (d-c) 66457

f Aggregate annual public benefits from training as a result of RWH adoption (NRs.) (e*b) 134,907,710

g Aggregate annual public benefits from training as a result of RWH adoption (USD) (f/100) 1,349,077

h Investment required for 7000 farmers (NRs.) from Table 6 387,604,000

i Investment required for 7000 farmers (USD) from Table 6 3,876,040


Figures

Figure 1: Location of the four districts

Figure 2: Number of vegetable and cereal crop producers among RWH adopters and non-adopters

Study Area (District)

Syangja

Makwanpur

Gulmi Palpa

21


Figure 3: Households farm income by training received and RWH technology adoption

Figure 4: Total annual average benefits and costs from RWH plastic pond


Appendices

Appendix 1: Average monthly rainfall in study area

Source: Metrology Department, GoN, 2012 and World Bank, Official Website

Notes: Nepal is characterized by heavy rainfall from July to September and low rainfall the rest of the year. Farmers in the

study area harvest rainfall during heavy rainfall months and they utilize harvested rainwater for vegetable crops production

when rainfall is low.

23


Appendix 2: Determinants of household income from agriculture and livestock sector (Using OLS and instrumental variable models)

Variable Model 1 (OLS) Model 2 (IV) Model 3 (IV)lnIncome farm: Coefficient Coefficient Coefficient

RWH Technology0.458*** (0.099)

1.219*** (0.264)

1.2384***(0.251)

Livestock standard unit0.032

(0.021)0.026

(0.023)0.027

(0.023)

ln Spade0.031

(0.024)0.011

(0.027)0.008

(0.027)

ln Total land0.645*** (0.064)

0.589*** (0.072)

0.531*** (0.071)

ln Shareup land-0.012 (0.024)

-0.025 (0.027)

-0.016 (0.027)

_cons9.2785***

(0.162)9.090*** (0.186)

9.227*** (0.184)

Instrument (training) No Yes YesDFE 25 No No YesSummary Statistics N = 279

F(5, 273) = 34.48*** Adj R-squared = 0.37

N = 279 F( 5, 273) = 29.02*** Centered R2 = 0.25

First stage regression:Test of excluded instruments:F(1, 273) = 54.97***

N = 279 F( 8, 270) = 21.86*** Centered R2 = 0.30

First stage regression:Test of excluded instruments:F(1, 270) = 56.70***

25 District fixed-effects (DFE): Gulmi, Syangja, Palpa (Makwanpur-omitted district)

SANDEEP.O. Box 8975, E.P.C 1056, Lalitpur, Nepal

Street address: c/o ICIMOD, Khumaltar, Lalitpur, Nepal

Tel: 977 1 5003222, Fax: 977 1 5003299, Email: [email protected], Web: www.sandeeonline.org

SANDEE Sponsors

rainwater harvesting and rural livelihoods in nepal · rainwater harvesting and rural livelihoods...

Documents