ABSTRACT
This study investigated how access to irrigation contributed to crop income and livelihood improvement in the study area in Myanmar. The study was conducted in the Pyawt Ywar pump irrigation command area in Myinmu Township, and stratified random sampling was used to collect data from 285 households. These included 208 beneficiaries of irrigation and 77 non-beneficiaries. The Full Information Maximum Likelihood Estimation of Endogenous Switching Regression model was employed to analyze the data. The results showed that the likelihood of a farmer using canal irrigation was associated with distance to the market, the number of water sources available, and the type of cash crop grown. In turn, crop income was influenced by farm size, market distance, block position, livestock units carried, non-farm income, and type of cash crop grown. It was determined that access to canal irrigation water enhanced a farmer’s income in the study area.
Keywords: canal irrigation, crop income
INTRODUCTION
Of the various methods and technologies that enhance agricultural productivity, expanding irrigation is the most viable option to achieve increased agricultural productivity in Myanmar. Irrigation stimulates agricultural productivity and economic growth (Van Den Berg & Ruben, 2006). Increasing the agricultural yield, cultivated land area, and cropping intensity can increase agricultural productivity (Awulachew et al., 2007; Movik, Mehta, Mtisi & Nicol, 2005). According to the Myanmar Census of Agriculture, the average land holding size in Myanmar is 2.6 hectares, which is declining (Department of Settlement and Land Records [SLRD], 2010). Hence, expanding cultivated areas is a finite option in Myanmar, and individual farmers have reached a limit.
Where agriculture is predominantly rainfed, irrigation development has become a crucial resource in improving livelihoods by increasing agricultural production and farm income. In Myanmar, however, irrigated area is only about 16.5% of the net sown area. Of this area under irrigation, 25.5% is subjected to multiple cropping (Ministry of Agriculture, Livestock, and Irrigation [MOALI], 2018). Thus, the more significant proportion of Myanmar’s cultivated area is rain-fed and vulnerable to unpredictable rainfall patterns. Increasing the percentage of irrigated land provides a greater chance of improvement in livelihoods than for farmers depending on rainfall (Lipton, Litchfield & Faures, 2003). Following the importance of irrigation in agriculture, the Government of Myanmar has made considerable efforts to expand irrigation through infrastructure development. Facilities include 235 dams, 202 river pump stations, 107 weirs, 168 sluices, and 7l tanks. Distribution methods include gravity-fed canals and reservoir schemes, river pumping, and groundwater systems (MOALI, 2018).
Several studies have been conducted in several different areas to assess the impact on household welfare after adopting new technologies, including irrigation. Gebregziabher, Namara, and Holden (2012) found that, in terms of their technical efficiency, irrigation farmers in northern Ethiopia operated on a higher production frontier with significant inefficiencies, while rain-fed farms were on a lower production frontier, with high-efficiency levels. Hussain, Giordano, and Hanjra (2003) showed that irrigation enabled households to improve crop productivity in South and Southeast Asia. A comparison between irrigators and non-irrigators in China demonstrated that irrigation increased yields for almost all crops and higher income for farmers in all areas (Huang, Rozelle, Lohmar, Huang & Wang, 2006). In Ethiopia, access to irrigation has had a positive and significant impact on the two outcome variables: income increased by 8.8% and asset formation by 186% compared to non-users (Gebrehiwot, Makina & Woldu, 2017).
However, in Ghana, Acheampong, Balana, Nimoh, and Abaidoo (2018) showed that there was only about a 3% increase in the income of vegetable farmers participating in irrigated vegetable production using small reservoir irrigation compared to those not irrigating, and this difference was statisticslly insignificant. In contrast to this, it was found that using propensity score matching (PSM) and switching regression, that improved access to irrigation in the rural savannah region of Ghana significantly improved household welfare via increases in net farm income, and there was more room for enhanced impacts (Owusu, Namara & Kuwornu, 201l).
However, in Myanmar, studies on the effect of irrigation on crop income are minimal. It is against this background that this study was undertaken. It assessed the impact of irrigation on crop income for households in the Pyawt Ywar Pump Irrigation Project (PYPP) area.
METHODOLOGY
The study area was selected the area of Pyawt Ywar Pump Irrigation Project (PYPP) that located in Myinmu Township, Sagaing Region. In Myinmu Township, the production and cultivation of the major growing crops were rice, groundnut, sesame, sunflower, black gram, green gram, pigeon pea, cotton and corn in 2016-2017. In monsoon season, most cultivated crops were rice, groundnut, green gram and pigeon pea. Groundnut was highly grown in winter season and followed by sesame and sunflower. Summer rice was grown under irrigated area (GAD (Myinmu), 2017).
It is one of the largest irrigation command areas in Myinmu Township, supplying water for 2,024 ha of the sown farming area. The PYPIP irrigation scheme became operational in 2005 using diesel fuel pumps to lift water from the Mu River. Pumps were upgraded in 2010-2011 to electrically driven pumps: this resulted in an expansion of irrigated land from 892 hectare in 2010 to 1,457 ha in 2016 (IWUMD, 2017). Problems occur during extremely wet years or years with extremely high rainfall when the high water levels in the river carry more debris and sediment load in the monsoon season. This has an impact on pump operation due to clogging of the filters which causes pump failures (IWMI, 2018).
The PYPP Project consists of three pump stations and three main canals near the Mu River. Other water sources, such as lakes and tube wells, are available. There is a water management committee comprised of Irrigation and Water Utilization Management Department officials and farmer representatives from irrigation command areas. This committee plans the rotational irrigation schedule for the allocation of irrigation water. The farmers in the study area usually sell and buy farm outputs and inputs at the nearest market, which is, on average l3.l kilometers (km) away from farm units. Monsoon and winter crops are the primary crops for farmers. In winter, onion is usually grown as the main cash crop.
The target population for this study consisted of farm households under the PYPIP command area. The study area was stratified into three segments (Blocks 1, 2, and 3) by using a stratified random sampling method based on the location of the pump stations. The farmers in each block were further divided into beneficiaries (farmers who could access canal irrigation) and non-beneficiaries (farmers who could not access canal irrigation). This dividing was done using farmer lists from the Myinmu Township offices of the Irrigation and Water Utilization Management Department and Welthugerhilfe (WHH). Of the beneficiaries, there was further stratification into head-end, middle, and tail-end beneficiaries. Respondents were sampled with a 5% margin of error and at a 95% confidence level. The data set consisted of beneficiaries (208) and non-beneficiaries (77). The data were thus collected using a structured questionnaire from 285 farmers based on their production activities for the 2016-2017 seasons.
MODEL SPECIFICATION
In practice, evaluating the impact of canal irrigation on an outcome variable using linear regression analysis can lead to a biased estimate if the underlying process which governs selection into the irrigation scheme is not incorporated in the empirical framework. In this case, participation into the irrigation scheme is not only by self-selection of farmers but also non-random selection by the irrigation scheme. Hence, participation decision could be influenced by the observed (farm and household characteristics), and unobserved factors (motivation and management skills) of farmers. The reason for this is that the effect of the irrigation scheme may be over or underestimated if the beneficiaries are more (or less) able (due to certain unobservable characteristics) to derive benefits compared to eligible non-beneficiaries (Zaman, 2001).
In this study, some unobservable characteristics such as skill, innovation, and attitude in farm households may affect not only the use of agricultural irrigation but also other farming decisions, leading to endogeneity and self-selection problems in the model (Di Falco & Veronesi, 2013). Therefore, if the endogeneity that arises in farming and use of irrigation scheme is not taken into account, the true impact on farming cannot be estimated. Hence, the endogenous switching regression model that enables us to jointly consider the use of both the irrigation scheme and farming within a single framework was used. It further allows implementing counterfactual experiments for answering what the impact of irrigation is, if non-beneficiaries use irrigation or if beneficiaries do not use it.
An endogenous switching regression model was used to estimate the relationship between the access to canal water and crop income in the study area, controlling for self-selection bias. The endogenous switching regression can be used to predict expected crop income for beneficiaries if they switched to not accessing canal irrigation and vice versa for non-beneficiaries (Lokshin & Sajaia, 2004).
An endogenous switching regression model follows two steps. In the first step, it models the decision of whether or not farmers use canal irrigation. In the second step, it models the outcome of farming depending whether on farmers are beneficiaries or non-beneficiaries. Let the decision to use one of the canal irrigation be a dichotomous choice, where a farmer decides to use canal irrigation when there is a positive perceived difference between using canal irrigation and not using the canal irrigation.
RESULTS AND DISCUSSION
Demographic and socioeconomic variables used in FIML ESR model estimation
The descriptive statistics of the data for the variables used in the model estimation are presented in Table 1. The average years of schooling undertaken by the respondents of the two groups were similar. There was also little difference between the average farming experience of the two groups. The average family and farm sizes were not statistically different between the two groups. However, the two groups had differences in annual crop and non-farm income. The non-beneficiaries received higher crop incomes, but beneficiaries earned more non-farm income than non-beneficiaries. In fact, beneficiaries mainly grow rice in summer and monsoon, whereas non-beneficiaries produce high-return cash crops such as onions in winter besides summer and monsoon rice.
The average number of livestock held (bullocks, cows, calves, dairy cows, goats, and sheep) was standardized by converting this different livestock into a common unit, here designated as a Tropical Livestock Unit (TLU), which is equal to the product of the type of animal and the related animal coefficient. In this study animal coefficients (bullock/ox: 1.10, cow/dairy cow= 1.0, calf =0.20, goat/sheep= 0.10) were adopted from Storck, Emana, Adenew, Borowiecki and Wolde-Hawariat (1991). As the results indicated, the average number of livestock held was similar for the two groups. Concerning the distance from the nearest market, beneficiaries have a far greater distance to travel than non-beneficiaries in this study area.
There were several water sources for crop production in the study area other than canal water. These include water from tube wells, lakes, and the Mu River. In determining the number of water sources accessed, beneficiaries accessed more sources than non-beneficiaries. For the measurement of the female participation in family decision-making in farm management activities (crop selection, seed selection, fertilizer application, crop protection, and crop marketing), there was no statistical difference between the two groups. Onion was the predominant cash crop produced by non-beneficiaries in the winter season. The involvement of non-beneficiary households in village development groups was relatively low compared with the involvement of beneficiary households.
Canal irrigation impact on annual crop income
The results from the FIML ESR model are presented in Table 2. This model jointly estimated both the selection and the outcome equations. The Wald χ2 test statistics indicated that the selected covariates provided a reasonable estimate of determinants to be applied to the model and were jointly and statistically significant. The likelihood ratio test statistics for joint independence (χ2) showed that the equations were dependent. Also, the chi-square statistic indicated that over-identification in the regression specifications of annual crop income was highly significant when different from zero. The results showed that the selectivity terms (ρbeneficiary and ρρnon-beneficiary) were statistically significant, indicating self-selection in undertaking irrigation-based farming in the study area. This result also implies that participation in canal irrigation may impact the non-beneficiaries if they choose to participate in the future (Lokshin & Sajaia, 2004).
Irrigation selection
The second column of Table 2 provides the estimates for the determinants for the decision to use canal irrigation. The distance to the market and the number of water sources used displayed a significant and positive relationship to the probability of participation in canal irrigation. The income variable from cash crops was negatively correlated to the selection of canal irrigation, and this might result from the irrigation scheme providing water in the monsoon and summer crop seasons, while cash crop (onion) growers preferred using other irrigation facilities, such as tube wells, in winter.
Outcome equation
The last two columns of Table 2 displayed the annual crop income coefficients for the two groups, with the natural logarithm of annual crop income as the dependent variable. The most notable results were those for farm size and cash crop growth (onion) for both groups. Farmers with larger farms had a higher annual crop income than those with smaller farms. Similarly, growing cash crops affected annual crop income positively, implying that cash crop (onion) growers received higher annual crop income than those who did not grow cash crops. The unit showing the total holding of livestock was significantly and positively correlated to levels of crop income for beneficiaries. Non-farm income, market distance, and block position have been identified as having significant effects on the annual crop income. Due to high transactional costs, market distance negatively affected the crop income of beneficiaries. Poor water availability for crop production would result in lower crop income for farmers in downstream block positions (Block 3). The amount of education undertaken by heads of families and the family size has been identified as negatively correlated to crop income for the non-beneficiary group, implying that these factors did not contribute to welfare in terms of the crop income of non-beneficiaries.
Impact of canal irrigation
The impact of canal irrigation determined by comparing the conditional and counterfactual conditions is presented in Table 3. The results of average treatment effect on treated farmers (ATT) and average treatment effect on untreated farmers (ATU) from the endogenous switching regression estimation showed positive and statistically significant results, indicating that canal irrigation positively impacted the annual crop income received by farmers. As a result, with access to canal irrigation by beneficiaries, income increased by about 34% annually with the use of other water sources. On the other hand, if they had no access, their crop income decreased by the same percentage. Based on the ATU, if non-beneficiaries could access canal irrigation, their crop income would increase by about 32% when accompanied by their existing use of other water sources. According to the base heterogeneity effects of beneficiaries (BH1), the non-beneficiaries would achieve a more remarkable improvement in annual crop income through being able to access canal irrigation in comparison to the crop income of beneficiaries.
CONCLUSION AND RECOMMENDATIONS
This study has analyzed the impact of canal irrigation on crop income and the determinants of farmers’ decision-making regarding canal irrigation in the PYPIP area using an FML ESR model. According to FML ESR results, variables relating to market distance, cash crop (onion) growing, and water source availability significantly impacted the farmer’s decision to use canal irrigation.
Block position negatively influenced the crop income of beneficiaries. Farmers from locally high-field elevations within the command area had lower crop incomes. The unequal water distribution among the blocks resulted in insufficient water and low crop yield. Livestock holdings also influenced crop income positively, probably due to the effect of both animal labour and the benefits of organic matter for crops. Animals can also be sold when there are financial constraints, and the cash realized can then be used for farm improvements which may increase crop income. It was also found that developing the rural livestock sector has become increasingly important.
According to counterfactual analysis, the ATT and ATU were positive and significant for both beneficiaries and non-beneficiaries, noting that access to canal irrigation has resulted in a positive impact on the crop income of farmers with the existing use of other water sources, primarily due to crop intensification and crop diversification. Receiving canal irrigation water increases crop income significantly for all farmers in the irrigation command area, and the effects are more critical for non-beneficiaries because they would have benefited more by receiving irrigation scheme water. Thus, the irrigation scheme positively impacts farm households in the irrigation command area.
This study observed that the impact of canal irrigation on crop income and cash crop growing is significant in the study area. Development of the growing of cash crops and the development of the rural livestock sector should be encouraged in the study area. Upgrading and expanding irrigation facilities will increase cash crop planting and reduce average irrigation costs. Simultaneously, market distance constraints could be solved with improvements to road infrastructure, transportation facilities, and marketing systems. Moreover, improvement in water management performance through the provision of a reliable water supply should be established to reduce the difficulties in the study area.
Due to the importance of irrigation in agriculture, the Government of Myanmar has made considerable efforts to expand irrigation using gravity-fed canal and reservoir schemes, river pumping, and groundwater systems. However, without proper community participation in the water management, the expected outcome cannot be fulfilled particularly in an area of water scarcity. Greater participation by farmers, through collective water management, would provide sustainability of the water resources development and efficient and effective utilization of water for increased crops production. Therefore, the rural livelihoods of Myanmar require the improvement of irrigation systems with community participation. By understanding the constraints and determinants of farmer participation in collective water management activities, local water management systems can be strengthened in order to produce sustainable water resources development and efficient and effective utilization of water for the improvement of rural livelihoods of Myanmar.
REFERENCES
Acheampong D., B. B. Balana, F. Nimoh & R. C. Abaidoo. (2018). Assessing the effectiveness and impact of agricultural water management interventions: the case of small reservoirs in northern Ghana. Agricultural Water Management, 209, 163-170.
Awulachew S. B., A. D. Yilma, M. Loulseged, W. Loiskandl, M. Ayana & T. Alamirew. (2007). Water resources and irrigation development in Ethiopia, 123, International Water Management Institute.
Di Falco S. & M. Veronesi. (2013). How can African agriculture adapt to climate change? A counterfactual analysis from Ethiopia. Land Economics, 89(4), 743-766.
Gebregziabher G., R. E. Namara & S. Holden. (2012). Technical efficiency of irrigated and rain-fed smallholder agriculture in Tigray, Ethiopia: A comparative stochastic frontier production function analysis. Quarterly Journal of International Agriculture, 51(892-2016-65167), 203-226.
Gebrehiwot K. G., D. Makina & T.Woldu. (2017). The impact of micro-irrigation on households’ welfare in the northern part of Ethiopia: an endogenous switching regression approach. Studies in Agricultural Economics, 119(3), 160-167.
Huang Q., S. Rozelle, B. Lohmar, J. Huang & J. Wang. (2006). Irrigation, agricultural performance and poverty reduction in China. Food Policy, 31(1), 30-52.
Hussain I., M. Giordano & M. A. Hanjra. (2003). Agricultural water and poverty linkages: Case studies on large and small systems. Retrieved from https://ageconsearch.umn.edu/record/158358/files/ H032548.pdf
Lipton M., J. Litchfield & J.-M.Faurès. (2003). The effects of irrigation on poverty: a framework for analysis. Water Policy, 5(5-6), 413-427.
Lokshin M. & Z. Sajaia. (2004). Maximum likelihood estimation of endogenous switching regression models. The Stata Journal, 4(3), 282-289.
Maddala G. S. (1986). Limited-dependent and qualitative variables in econometrics: Cambridge university press.
Ministry of Agriculture, Livestock and Irrigation. (2018). Myanmar Agriculture at a Glance. Naypyitaw, Myanmar: Author.
Movik S., L. Mehta, S. Mtisi & A. Nicol. (2005). A “blue revolution” for African agriculture? Retrieved from https://opendocs.ids.ac.uk/opendocs/bitstream/handle/20.500.12413/8482 /IDSB_36_2_10.1111-j.1759-5436.2005.tb00194.x.pdf
Narayanamoorthy A. (2001). Irrigation and rural poverty nexus: a statewide analysis. Indian Journal of Agricultural Economics, 56(1), 40.
Owusu E. S., R. E.Namara & J. K. Kuwornu. (2011). The welfare-enhancing role of irrigation in farm households in northern Ghana. Journal of International Diversity, 1, 61-87.
Settlement and Land Records Department. (2010). Report on Myanmar Census of Agriculture 1993 and 2010. Ministry of Agriculture and Irrigation. Naypyitaw, Myanmar.
Shah T. & O. Singh. (2002). Irrigation Development and Rural Poverty in Gujarat: A Disaggregate Analysis. Paper presented at the Annual Review Meeting, November, International Water Management Institute, Colombo, Sri Lanka.
Storck H., B. Emana, B. Adenew, A. Borowiecki & S. Wolde-Hawariat. (1991). Farming Systems and Farm Management Practices of Smallholders in the Harerghe Highlands-a Baseline Survey Farming Systems and Resource Economics in the Tropics. Kiel, Germany: Wissenschaftsverlag Vauk.
Van Den Berg M., & R. Ruben. (2006). Small-scale irrigation and income distribution in Ethiopia. The Journal of Development Studies, 42(5), 868-880.
Verbeek M. (2012). A guide to modern econometrics. Fourth edition. John Wiley & Sons.
Impact of Canal Irrigation on Crop Income in the Pyawt Ywar Irrigation Project Area, Myinmu Township
ABSTRACT
This study investigated how access to irrigation contributed to crop income and livelihood improvement in the study area in Myanmar. The study was conducted in the Pyawt Ywar pump irrigation command area in Myinmu Township, and stratified random sampling was used to collect data from 285 households. These included 208 beneficiaries of irrigation and 77 non-beneficiaries. The Full Information Maximum Likelihood Estimation of Endogenous Switching Regression model was employed to analyze the data. The results showed that the likelihood of a farmer using canal irrigation was associated with distance to the market, the number of water sources available, and the type of cash crop grown. In turn, crop income was influenced by farm size, market distance, block position, livestock units carried, non-farm income, and type of cash crop grown. It was determined that access to canal irrigation water enhanced a farmer’s income in the study area.
Keywords: canal irrigation, crop income
INTRODUCTION
Of the various methods and technologies that enhance agricultural productivity, expanding irrigation is the most viable option to achieve increased agricultural productivity in Myanmar. Irrigation stimulates agricultural productivity and economic growth (Van Den Berg & Ruben, 2006). Increasing the agricultural yield, cultivated land area, and cropping intensity can increase agricultural productivity (Awulachew et al., 2007; Movik, Mehta, Mtisi & Nicol, 2005). According to the Myanmar Census of Agriculture, the average land holding size in Myanmar is 2.6 hectares, which is declining (Department of Settlement and Land Records [SLRD], 2010). Hence, expanding cultivated areas is a finite option in Myanmar, and individual farmers have reached a limit.
Where agriculture is predominantly rainfed, irrigation development has become a crucial resource in improving livelihoods by increasing agricultural production and farm income. In Myanmar, however, irrigated area is only about 16.5% of the net sown area. Of this area under irrigation, 25.5% is subjected to multiple cropping (Ministry of Agriculture, Livestock, and Irrigation [MOALI], 2018). Thus, the more significant proportion of Myanmar’s cultivated area is rain-fed and vulnerable to unpredictable rainfall patterns. Increasing the percentage of irrigated land provides a greater chance of improvement in livelihoods than for farmers depending on rainfall (Lipton, Litchfield & Faures, 2003). Following the importance of irrigation in agriculture, the Government of Myanmar has made considerable efforts to expand irrigation through infrastructure development. Facilities include 235 dams, 202 river pump stations, 107 weirs, 168 sluices, and 7l tanks. Distribution methods include gravity-fed canals and reservoir schemes, river pumping, and groundwater systems (MOALI, 2018).
Several studies have been conducted in several different areas to assess the impact on household welfare after adopting new technologies, including irrigation. Gebregziabher, Namara, and Holden (2012) found that, in terms of their technical efficiency, irrigation farmers in northern Ethiopia operated on a higher production frontier with significant inefficiencies, while rain-fed farms were on a lower production frontier, with high-efficiency levels. Hussain, Giordano, and Hanjra (2003) showed that irrigation enabled households to improve crop productivity in South and Southeast Asia. A comparison between irrigators and non-irrigators in China demonstrated that irrigation increased yields for almost all crops and higher income for farmers in all areas (Huang, Rozelle, Lohmar, Huang & Wang, 2006). In Ethiopia, access to irrigation has had a positive and significant impact on the two outcome variables: income increased by 8.8% and asset formation by 186% compared to non-users (Gebrehiwot, Makina & Woldu, 2017).
However, in Ghana, Acheampong, Balana, Nimoh, and Abaidoo (2018) showed that there was only about a 3% increase in the income of vegetable farmers participating in irrigated vegetable production using small reservoir irrigation compared to those not irrigating, and this difference was statisticslly insignificant. In contrast to this, it was found that using propensity score matching (PSM) and switching regression, that improved access to irrigation in the rural savannah region of Ghana significantly improved household welfare via increases in net farm income, and there was more room for enhanced impacts (Owusu, Namara & Kuwornu, 201l).
However, in Myanmar, studies on the effect of irrigation on crop income are minimal. It is against this background that this study was undertaken. It assessed the impact of irrigation on crop income for households in the Pyawt Ywar Pump Irrigation Project (PYPP) area.
METHODOLOGY
The study area was selected the area of Pyawt Ywar Pump Irrigation Project (PYPP) that located in Myinmu Township, Sagaing Region. In Myinmu Township, the production and cultivation of the major growing crops were rice, groundnut, sesame, sunflower, black gram, green gram, pigeon pea, cotton and corn in 2016-2017. In monsoon season, most cultivated crops were rice, groundnut, green gram and pigeon pea. Groundnut was highly grown in winter season and followed by sesame and sunflower. Summer rice was grown under irrigated area (GAD (Myinmu), 2017).
It is one of the largest irrigation command areas in Myinmu Township, supplying water for 2,024 ha of the sown farming area. The PYPIP irrigation scheme became operational in 2005 using diesel fuel pumps to lift water from the Mu River. Pumps were upgraded in 2010-2011 to electrically driven pumps: this resulted in an expansion of irrigated land from 892 hectare in 2010 to 1,457 ha in 2016 (IWUMD, 2017). Problems occur during extremely wet years or years with extremely high rainfall when the high water levels in the river carry more debris and sediment load in the monsoon season. This has an impact on pump operation due to clogging of the filters which causes pump failures (IWMI, 2018).
The PYPP Project consists of three pump stations and three main canals near the Mu River. Other water sources, such as lakes and tube wells, are available. There is a water management committee comprised of Irrigation and Water Utilization Management Department officials and farmer representatives from irrigation command areas. This committee plans the rotational irrigation schedule for the allocation of irrigation water. The farmers in the study area usually sell and buy farm outputs and inputs at the nearest market, which is, on average l3.l kilometers (km) away from farm units. Monsoon and winter crops are the primary crops for farmers. In winter, onion is usually grown as the main cash crop.
The target population for this study consisted of farm households under the PYPIP command area. The study area was stratified into three segments (Blocks 1, 2, and 3) by using a stratified random sampling method based on the location of the pump stations. The farmers in each block were further divided into beneficiaries (farmers who could access canal irrigation) and non-beneficiaries (farmers who could not access canal irrigation). This dividing was done using farmer lists from the Myinmu Township offices of the Irrigation and Water Utilization Management Department and Welthugerhilfe (WHH). Of the beneficiaries, there was further stratification into head-end, middle, and tail-end beneficiaries. Respondents were sampled with a 5% margin of error and at a 95% confidence level. The data set consisted of beneficiaries (208) and non-beneficiaries (77). The data were thus collected using a structured questionnaire from 285 farmers based on their production activities for the 2016-2017 seasons.
MODEL SPECIFICATION
In practice, evaluating the impact of canal irrigation on an outcome variable using linear regression analysis can lead to a biased estimate if the underlying process which governs selection into the irrigation scheme is not incorporated in the empirical framework. In this case, participation into the irrigation scheme is not only by self-selection of farmers but also non-random selection by the irrigation scheme. Hence, participation decision could be influenced by the observed (farm and household characteristics), and unobserved factors (motivation and management skills) of farmers. The reason for this is that the effect of the irrigation scheme may be over or underestimated if the beneficiaries are more (or less) able (due to certain unobservable characteristics) to derive benefits compared to eligible non-beneficiaries (Zaman, 2001).
In this study, some unobservable characteristics such as skill, innovation, and attitude in farm households may affect not only the use of agricultural irrigation but also other farming decisions, leading to endogeneity and self-selection problems in the model (Di Falco & Veronesi, 2013). Therefore, if the endogeneity that arises in farming and use of irrigation scheme is not taken into account, the true impact on farming cannot be estimated. Hence, the endogenous switching regression model that enables us to jointly consider the use of both the irrigation scheme and farming within a single framework was used. It further allows implementing counterfactual experiments for answering what the impact of irrigation is, if non-beneficiaries use irrigation or if beneficiaries do not use it.
An endogenous switching regression model was used to estimate the relationship between the access to canal water and crop income in the study area, controlling for self-selection bias. The endogenous switching regression can be used to predict expected crop income for beneficiaries if they switched to not accessing canal irrigation and vice versa for non-beneficiaries (Lokshin & Sajaia, 2004).
An endogenous switching regression model follows two steps. In the first step, it models the decision of whether or not farmers use canal irrigation. In the second step, it models the outcome of farming depending whether on farmers are beneficiaries or non-beneficiaries. Let the decision to use one of the canal irrigation be a dichotomous choice, where a farmer decides to use canal irrigation when there is a positive perceived difference between using canal irrigation and not using the canal irrigation.
RESULTS AND DISCUSSION
Demographic and socioeconomic variables used in FIML ESR model estimation
The descriptive statistics of the data for the variables used in the model estimation are presented in Table 1. The average years of schooling undertaken by the respondents of the two groups were similar. There was also little difference between the average farming experience of the two groups. The average family and farm sizes were not statistically different between the two groups. However, the two groups had differences in annual crop and non-farm income. The non-beneficiaries received higher crop incomes, but beneficiaries earned more non-farm income than non-beneficiaries. In fact, beneficiaries mainly grow rice in summer and monsoon, whereas non-beneficiaries produce high-return cash crops such as onions in winter besides summer and monsoon rice.
The average number of livestock held (bullocks, cows, calves, dairy cows, goats, and sheep) was standardized by converting this different livestock into a common unit, here designated as a Tropical Livestock Unit (TLU), which is equal to the product of the type of animal and the related animal coefficient. In this study animal coefficients (bullock/ox: 1.10, cow/dairy cow= 1.0, calf =0.20, goat/sheep= 0.10) were adopted from Storck, Emana, Adenew, Borowiecki and Wolde-Hawariat (1991). As the results indicated, the average number of livestock held was similar for the two groups. Concerning the distance from the nearest market, beneficiaries have a far greater distance to travel than non-beneficiaries in this study area.
There were several water sources for crop production in the study area other than canal water. These include water from tube wells, lakes, and the Mu River. In determining the number of water sources accessed, beneficiaries accessed more sources than non-beneficiaries. For the measurement of the female participation in family decision-making in farm management activities (crop selection, seed selection, fertilizer application, crop protection, and crop marketing), there was no statistical difference between the two groups. Onion was the predominant cash crop produced by non-beneficiaries in the winter season. The involvement of non-beneficiary households in village development groups was relatively low compared with the involvement of beneficiary households.
Canal irrigation impact on annual crop income
The results from the FIML ESR model are presented in Table 2. This model jointly estimated both the selection and the outcome equations. The Wald χ2 test statistics indicated that the selected covariates provided a reasonable estimate of determinants to be applied to the model and were jointly and statistically significant. The likelihood ratio test statistics for joint independence (χ2) showed that the equations were dependent. Also, the chi-square statistic indicated that over-identification in the regression specifications of annual crop income was highly significant when different from zero. The results showed that the selectivity terms (ρbeneficiary and ρρnon-beneficiary) were statistically significant, indicating self-selection in undertaking irrigation-based farming in the study area. This result also implies that participation in canal irrigation may impact the non-beneficiaries if they choose to participate in the future (Lokshin & Sajaia, 2004).
Irrigation selection
The second column of Table 2 provides the estimates for the determinants for the decision to use canal irrigation. The distance to the market and the number of water sources used displayed a significant and positive relationship to the probability of participation in canal irrigation. The income variable from cash crops was negatively correlated to the selection of canal irrigation, and this might result from the irrigation scheme providing water in the monsoon and summer crop seasons, while cash crop (onion) growers preferred using other irrigation facilities, such as tube wells, in winter.
Outcome equation
The last two columns of Table 2 displayed the annual crop income coefficients for the two groups, with the natural logarithm of annual crop income as the dependent variable. The most notable results were those for farm size and cash crop growth (onion) for both groups. Farmers with larger farms had a higher annual crop income than those with smaller farms. Similarly, growing cash crops affected annual crop income positively, implying that cash crop (onion) growers received higher annual crop income than those who did not grow cash crops. The unit showing the total holding of livestock was significantly and positively correlated to levels of crop income for beneficiaries. Non-farm income, market distance, and block position have been identified as having significant effects on the annual crop income. Due to high transactional costs, market distance negatively affected the crop income of beneficiaries. Poor water availability for crop production would result in lower crop income for farmers in downstream block positions (Block 3). The amount of education undertaken by heads of families and the family size has been identified as negatively correlated to crop income for the non-beneficiary group, implying that these factors did not contribute to welfare in terms of the crop income of non-beneficiaries.
Impact of canal irrigation
The impact of canal irrigation determined by comparing the conditional and counterfactual conditions is presented in Table 3. The results of average treatment effect on treated farmers (ATT) and average treatment effect on untreated farmers (ATU) from the endogenous switching regression estimation showed positive and statistically significant results, indicating that canal irrigation positively impacted the annual crop income received by farmers. As a result, with access to canal irrigation by beneficiaries, income increased by about 34% annually with the use of other water sources. On the other hand, if they had no access, their crop income decreased by the same percentage. Based on the ATU, if non-beneficiaries could access canal irrigation, their crop income would increase by about 32% when accompanied by their existing use of other water sources. According to the base heterogeneity effects of beneficiaries (BH1), the non-beneficiaries would achieve a more remarkable improvement in annual crop income through being able to access canal irrigation in comparison to the crop income of beneficiaries.
CONCLUSION AND RECOMMENDATIONS
This study has analyzed the impact of canal irrigation on crop income and the determinants of farmers’ decision-making regarding canal irrigation in the PYPIP area using an FML ESR model. According to FML ESR results, variables relating to market distance, cash crop (onion) growing, and water source availability significantly impacted the farmer’s decision to use canal irrigation.
Block position negatively influenced the crop income of beneficiaries. Farmers from locally high-field elevations within the command area had lower crop incomes. The unequal water distribution among the blocks resulted in insufficient water and low crop yield. Livestock holdings also influenced crop income positively, probably due to the effect of both animal labour and the benefits of organic matter for crops. Animals can also be sold when there are financial constraints, and the cash realized can then be used for farm improvements which may increase crop income. It was also found that developing the rural livestock sector has become increasingly important.
According to counterfactual analysis, the ATT and ATU were positive and significant for both beneficiaries and non-beneficiaries, noting that access to canal irrigation has resulted in a positive impact on the crop income of farmers with the existing use of other water sources, primarily due to crop intensification and crop diversification. Receiving canal irrigation water increases crop income significantly for all farmers in the irrigation command area, and the effects are more critical for non-beneficiaries because they would have benefited more by receiving irrigation scheme water. Thus, the irrigation scheme positively impacts farm households in the irrigation command area.
This study observed that the impact of canal irrigation on crop income and cash crop growing is significant in the study area. Development of the growing of cash crops and the development of the rural livestock sector should be encouraged in the study area. Upgrading and expanding irrigation facilities will increase cash crop planting and reduce average irrigation costs. Simultaneously, market distance constraints could be solved with improvements to road infrastructure, transportation facilities, and marketing systems. Moreover, improvement in water management performance through the provision of a reliable water supply should be established to reduce the difficulties in the study area.
Due to the importance of irrigation in agriculture, the Government of Myanmar has made considerable efforts to expand irrigation using gravity-fed canal and reservoir schemes, river pumping, and groundwater systems. However, without proper community participation in the water management, the expected outcome cannot be fulfilled particularly in an area of water scarcity. Greater participation by farmers, through collective water management, would provide sustainability of the water resources development and efficient and effective utilization of water for increased crops production. Therefore, the rural livelihoods of Myanmar require the improvement of irrigation systems with community participation. By understanding the constraints and determinants of farmer participation in collective water management activities, local water management systems can be strengthened in order to produce sustainable water resources development and efficient and effective utilization of water for the improvement of rural livelihoods of Myanmar.
REFERENCES
Acheampong D., B. B. Balana, F. Nimoh & R. C. Abaidoo. (2018). Assessing the effectiveness and impact of agricultural water management interventions: the case of small reservoirs in northern Ghana. Agricultural Water Management, 209, 163-170.
Awulachew S. B., A. D. Yilma, M. Loulseged, W. Loiskandl, M. Ayana & T. Alamirew. (2007). Water resources and irrigation development in Ethiopia, 123, International Water Management Institute.
Di Falco S. & M. Veronesi. (2013). How can African agriculture adapt to climate change? A counterfactual analysis from Ethiopia. Land Economics, 89(4), 743-766.
Gebregziabher G., R. E. Namara & S. Holden. (2012). Technical efficiency of irrigated and rain-fed smallholder agriculture in Tigray, Ethiopia: A comparative stochastic frontier production function analysis. Quarterly Journal of International Agriculture, 51(892-2016-65167), 203-226.
Gebrehiwot K. G., D. Makina & T.Woldu. (2017). The impact of micro-irrigation on households’ welfare in the northern part of Ethiopia: an endogenous switching regression approach. Studies in Agricultural Economics, 119(3), 160-167.
Huang Q., S. Rozelle, B. Lohmar, J. Huang & J. Wang. (2006). Irrigation, agricultural performance and poverty reduction in China. Food Policy, 31(1), 30-52.
Hussain I., M. Giordano & M. A. Hanjra. (2003). Agricultural water and poverty linkages: Case studies on large and small systems. Retrieved from https://ageconsearch.umn.edu/record/158358/files/ H032548.pdf
Lipton M., J. Litchfield & J.-M.Faurès. (2003). The effects of irrigation on poverty: a framework for analysis. Water Policy, 5(5-6), 413-427.
Lokshin M. & Z. Sajaia. (2004). Maximum likelihood estimation of endogenous switching regression models. The Stata Journal, 4(3), 282-289.
Maddala G. S. (1986). Limited-dependent and qualitative variables in econometrics: Cambridge university press.
Ministry of Agriculture, Livestock and Irrigation. (2018). Myanmar Agriculture at a Glance. Naypyitaw, Myanmar: Author.
Movik S., L. Mehta, S. Mtisi & A. Nicol. (2005). A “blue revolution” for African agriculture? Retrieved from https://opendocs.ids.ac.uk/opendocs/bitstream/handle/20.500.12413/8482 /IDSB_36_2_10.1111-j.1759-5436.2005.tb00194.x.pdf
Narayanamoorthy A. (2001). Irrigation and rural poverty nexus: a statewide analysis. Indian Journal of Agricultural Economics, 56(1), 40.
Owusu E. S., R. E.Namara & J. K. Kuwornu. (2011). The welfare-enhancing role of irrigation in farm households in northern Ghana. Journal of International Diversity, 1, 61-87.
Settlement and Land Records Department. (2010). Report on Myanmar Census of Agriculture 1993 and 2010. Ministry of Agriculture and Irrigation. Naypyitaw, Myanmar.
Shah T. & O. Singh. (2002). Irrigation Development and Rural Poverty in Gujarat: A Disaggregate Analysis. Paper presented at the Annual Review Meeting, November, International Water Management Institute, Colombo, Sri Lanka.
Storck H., B. Emana, B. Adenew, A. Borowiecki & S. Wolde-Hawariat. (1991). Farming Systems and Farm Management Practices of Smallholders in the Harerghe Highlands-a Baseline Survey Farming Systems and Resource Economics in the Tropics. Kiel, Germany: Wissenschaftsverlag Vauk.
Van Den Berg M., & R. Ruben. (2006). Small-scale irrigation and income distribution in Ethiopia. The Journal of Development Studies, 42(5), 868-880.
Verbeek M. (2012). A guide to modern econometrics. Fourth edition. John Wiley & Sons.