1. Introduction
Addressing food loss (FL) is vital for building up sustainable agriculture and agri-food systems. As outlined in Sustainable Development Goal 12, global leaders committed to halving per capita food waste at the retail and consumer levels and reducing food losses throughout production and the supply chain, including post-harvest losses, by 2030. The Food Loss Index includes post-harvest storage losses (PHLs) as part of the value chain, with approximately 14% of the world’s food loss occurring between harvest and the retail stage [1,2]. A projection highlights that by 2050, feeding a population of 9.1 billion will require over a 70% increase in food production [3,4,5]. That is why mitigating post-harvest losses is imperative.
PHLs encompass both quantitative and qualitative losses, diminishing economic value, and deteriorating taste and nutritional content [6,7]. Under unfavorable conditions, PHLs can reach up to 80% of the product [8]. Smallholding farmers have faced post-harvest losses in the field, during processing, and during storage [6]. Storage losses account for the maximum fraction of all post-harvest losses for cereals in developing countries and detrimentally affect farmer’s livelihoods [5].
Evidence suggests that improved storage (hermetic storage technology) through technology such as metal silos and hermetic storage bags (e.g., Grain Pro bags and Purdue Improved Crop Storage (PICS) bags) reduces post-harvest losses and alleviates food insecurity [9,10]. Also, it boosts household food availability and supply [11], further strengthening food security. Chegere [12] emphasized that introducing hermetic storage bags is simple and cost-effective for smallholding farmers. PICS is a hermetically sealed bag comprising three layers: one outer layer made of polypropylene and two inner layers made of high-density polyethylene (HDPE). It is a cost-effective and chemical-free storage technology for grain storage. Inside the storage bag, a depletion of oxygen and an accumulation of carbon dioxide or nitrogen takes place, which creates suffocation for living organisms [11,13,14].
Maize is a staple food crop for over 4.5 billion individuals in 94 developing nations, along with wheat and rice, consisting of at least 30 percent of the dietary calorie requirement [8]. It has become a major worldwide commodity and has a significant and incremental role in global agri-food systems, acting as a source for human food consumption and animal feeding [15]. The estimated storage losses of maize is 20–30 percent due to insect pests, mostly the maize weevil (Sitophilus zeamais), larger grain borer (Prostephanus truncatus), Angoumois grain moth (Sitotroga cerealella), and the lesser grain weevil (Sitophilus oryzae), negatively impacting food security and income generation [8]. The PHLs of maize are estimated to be 16.33% in Rwanda [16], 19 to 23% in Tanzania [13], and 15 to 25.5% in Ethiopia [17].
The global challenges occurring due to storage losses in maize, particularly caused by insect pests, underscore the importance of assessing the impact of storage technology on smallholding farmers in regions like Nepal. In Nepal, the maize weevil and the Angoumois grain moth are the main concerns of farmers in the post-harvest storage of maize [18]. A lack of proper post-harvest storage systems causes maize storage losses in the mid-hills region, where maize plays an important role in local livelihood, providing food and animal feed. This region, dominated by smallholding farmers, still heavily relies on traditional storage methods as it has a high infestation of insect pests. Hence, encouraging smallholding farmers to adopt hermetic storage technology could be an effective solution to reduce storage losses and enhance local food security.
Maize is the second largest staple cereal crop in Nepal after rice. The maize-growing summer season is from February/April to July/August, but sometimes, due to local geographical and climatic conditions, it concludes at the end of September. The cultivation area comprises 985,565 hectares with an annual production quantity of 3,106,397 metric tons and had a 3.15 Mt/ha yield in the 2021/22 period [19]. Maize covers 28.83% of production and 28.27% of the cultivation area of cereal crops. It contributed 7.6% of the Agriculture Gross Domestic Product (AGDP) in the 2022/23 period [20]. The agriculture sector contributed 24.09% of the National Gross Domestic Product (GDP) in the 2023/24 period [21]. Previous studies showed that there is a varying pattern of consumption of maize in Nepal. Timsina et al. [22] found that 60% of production is used for food consumption, 25% for animal feed, 3% for seed, and the remaining 12% is sold in the hilly districts.
The majority of farmers store their maize using traditional storage systems, which vary by region and locality. Bhandari et al. [18] emphasized that the traditional storage method is responsible for post-harvest storage losses. Maize is traditionally stored outdoors (e.g., on a Thangro/Suli—a timber or bamboo rack) and in indoor structures (e.g., in sacks or Kuniyu, an open semi-structure for storing maize in sheaths, and Kotha/Bhakari, a cylindrical storage made from bamboo splits) [23]. Piling up maize on a tree in front of the house and hanging it on the veranda are also traditional outdoor storage practices. Outdoor storage areas are exposed to moisture and rain and are vulnerable to insect pest infestation.
Traditional methods provide less physical protection from storage pests, intensifying post-harvest losses [18,23]. The maize weevil and Angoumois moth pose significant challenges to farmers regarding storage in Nepal [18]. The findings of numerous past surveys in Nepal vary due to agro-climatic conditions and survey methods. In the hilly districts of Nepal, as a whole, storage losses range from 10 to 20% of the total production of maize [18,24]. Additionally, Paneru et al. [25] reported a substantial post-harvest loss of maize, reaching up to 19.5% ± 12.5%, attributed to pests and disease cobs in the mid-western and far-western mid-hills of Nepal. Despite the recent emphasis on government policies, programs, and plans across federal, provincial, and local governments with a specific focus on agricultural value chain development, challenges in post-harvest management persist.
Recent government policies, programs, and plans have aimed to modernize the agriculture sector with greater emphasis on this sector. The Prime Minister Agriculture Modernization Project (PMAMP); a ten-year program, aims to encourage farmers to adopt improved and new technologies [26]. This project implemented development programs for Small Commercial Agricultural Production Centers (Pocket), Commercial Agricultural Production Centers (Block), Commercial Agricultural Production and Processing Centers (ZONE), and Comprehensive Commercial Agricultural Production and Industrial Centers (Super Zone) for major agricultural products including maize. In the 2022/23 period, PMAPM implemented supporting programs (e.g., subsidies of up to 85% in agriculture equipment, marketing support for agri-products, storage infrastructure, and establishing processing industries) [27].
Similarly, the Agriculture Development Strategy (ADS), 2015–2035, in Nepal focuses on value chain development at all levels and on enhancing food security nationwide [28]; it is a 20-year agriculture strategic plan that includes a 10-year action plan and roadmap based on an assessment of the present and past performances of the agriculture sector. Value chain development is one of four flagship programs. A lack of information, the nascent market of hermetic storage technology (HST), and economic factors are the main reasons why new technologies have not been adopted. However, hermetic storage bags (PICS bags and Super grain bags) were commercially introduced as viable alternatives [29,30]. Hence, to change storage behavior and reduce post-harvest storage losses, it is necessary to encourage smallholding farmers to adopt alternative storage techniques that are sustainable, straightforward to use, and cost-effective.
There is a growing body of literature dwelling on various aspects of post-harvest storage technology. Recent empirical studies are mostly associated with determinants of the adoption of hermetic storage technology. For example, Ndaghu et al. [31] found that age, awareness, extension contact, information sources, radio sources, cowpea income, and bag availability are associated with hermetic storage bag adoption among cowpea farmers in Northern Nigeria. Benimana et al. [16] concluded that household size and training on post-harvest handling and storage have a positive influence on the selection of hermetic storage technologies, such as silos and hermetic bags. In contrast, the genders of household heads and the distance to the input provider have a negative relationship in Rwanda. Alemu et al. [32] emphasized that the decision to adopt PICS bags was positively associated with cooperative membership, the education levels of household heads, livestock size, crop income, extension services, and the storage capacity of PICS bags. However, it was negatively associated with farmland size, maize production quantity, the non-food expenditure of households, the demand for agricultural inputs, and the cost of PICS bags in Northern Ethiopia. The authors of [33] suggested that learning from social networks (e.g., friends and relatives) and extension agents makes farmers more likely to adopt PICS bags more quickly. Cooperative membership is effective in post-harvest technology adoption in China [34].
Refs. [32,35] focus on the profitability of using PICS bags. Channa et al. [36] added that prior awareness of hermetic storage technology leads to a higher willingness to pay for and adopt it. Laboratory experiments emphasized the effectiveness of PICS bags in controlling insects, limiting damage, and preserving the quality of maize and shea nuts [37,38]. Odjo et al. [39] emphasized that hermetic storage technology is a satisfactory way to minimize the loss from insect damage to grains. Na’ang’a et al. [40] demonstrated that a PICS bag with three layers is efficient in maintaining the quality of maize by protecting it against aflatoxin and mold contamination. Shukla et al. [41] demonstrated that the benefits of adopting improved storage technology outweigh the traditional physical storage system in India.
There have been several impact evaluations of storage technology. However, there is no rigorous causal evidence on storage and post-harvest storage losses. For example, Aggarwal et al. [42] conducted a randomized controlled trial (RCT) in Kenya to assess the effect of a group storage intervention aimed to address storage constraints, specifically the collective storage of maize in hermetic bags. They found that farmers in the treatment group were 23% more likely to store maize compared to the control group. However, they did not find a statistically significant effect on the quantity of maize stored. Similarly, Chegere et al. [43] conducted an RCT in Tanzania and found a significant effect of a hermetic plastic bag intervention combined with training on storage losses. Farmers were randomly assigned a number of bags intended to store approximately 60% of the expected harvested maize. Since the decision to adopt new technology is not random [44,45], it might create some biased effects and be associated with awareness of the technology [36]. In addition, most previous studies estimated the intention-to-treat (ITT) effects, including the effect of being assigned a random treatment [10,11,13,46,47]. For instance, Brander et al. [10] studied the impact of randomly provided hermetic bags with training and found a significant effect on post-harvest losses by a percentage point of 11.1 in Tanzania. In contrast, Negede et al. [46] reported no effect of three randomly distributed hermetic bags on storage and storage losses in Ethiopia. Such studies were mostly focused on Africa and a few in Asia. Furthermore, there is no causal evidence of the effect of storage technology on storage and post-harvest storage losses in Nepal.
Thus, this study aims to assess the impact of PICS bags on storage quantity and self-reported post-harvest storage losses. To address the existing knowledge gap in the literature, we use a randomized encouragement design (RED) function as an instrumental variable (IV) approach to measure the impact of PICS bag adoption on the outcomes of interest. The IV approach can address the issue of endogeneity [48] and reconcile the ITT effect, which is the causal impact of random assignment for receiving treatment regardless of non-compliance. However, the IV approach considers non-compliance [48,49,50,51]. For this study, we randomly offered 100% and 50% subsidies on the price of up to three PICS bags to smallholding farmers.
The remaining part of this paper is organized as follows: Section 2 presents the materials and methods. Section 3 presents the results. Section 4 provides the discussion, followed by the conclusion in Section 5.
2. Materials and Methods
2.1. Study Site
We conducted this study at Mathagadhi rural municipality, which is located 25 km away from the headquarters of the Palpa district, Lumbini Province, Nepal. This area lies in the mid-hills of the mid-western part of Nepal (see Figure 1: Study site map). Mathagadhi rural municipality is a local government and has 8 wards. Wards are the smallest administrative level in Nepal. Geographically, it is the largest local government in size in the Palpa district out of 10, with an area of 215.49 square km. The settlements in the wards are scattered on the slopes of the folds of hills generally 1500 m above sea level, where the main mode of travel is on foot. For many, traveling to the municipality office for public services, including agricultural services and the market in the vicinity, could be a day’s work. The average annual temperature of Mathagadhi is approximately 23 °C. During the summer, temperatures can rise to as high as 33 °C, while in the winter, they can drop to as low as 3 °C. The main cereal crops produced are rice, maize, wheat, and millet.
We selected this study area due to four reasons. First, according to [52], about 84 percent of its active populations are engaging in the agriculture sector. Second, maize is the second largest crop in production, which plays a vital role in local food security and livelihood as a whole, being the source of food and animal feed, and it is used as an ingredient in religious functions and other cultural activities. The dried stalks and kernels of maize are used as fuel. The farmers in this area practice traditional storage methods, as mentioned in the previous section. They are facing insect pest problems with large post-harvest losses. Third, according to the government official of Mathagadhi rural municipality, it has a program aiming to subsidize hermetic storage bags in the near future. Thus, our findings may guide them in making the decision to use these bags. Finally, Mathagadhi represents a rural municipality in the mid-hills of Nepal, with similar climatic conditions, economic circumstances, and social aspects of local governments and other livelihood aspects across 28 other districts in Nepal.
2.2. Data and Sampling Strategy
We used primary cross-sectional data collected through a pre-structured questionnaire. We conducted a baseline survey from 25 August to 7 September 2023, which is the period in which farmers start storing the maize after harvest, jointly with local researchers using both a paper-and-pencil-based approach and a computer-assisted field entry (CAFÉ) approach. During the baseline survey, we collected socio-demographic data such as age, education years, gender, family income, maize production quantity, maize cultivated area, storage period, and quantity of storage losses, all of which might have an association with hermetic storage technology take-up (see Appendix A, Table A1).
In this study, we employed a two-step sampling method. First, we purposively selected Mathagadhi rural municipality from the Palpa district, one of the largest maize-producing mid-hill districts of Lumbini Province, Nepal. Second, smallholding farmers were randomly selected for the experiment, and interviews were conducted using a structured questionnaire. The list of maize farmers who cultivated maize in 2023 was obtained from the ward offices of the local government.
After collecting the list of farmers, we conducted a power calculation to determine the appropriate sample size for this experiment by following the method of [53]. For this, we selected a standard deviation of 0.56 for storage loss, a power of 80%, a significance level of 5%, and a minimum detectable effect (MDE) of 0.2. For sample sizes of different groups, we followed the method of [54], which suggests that if the experiment design has more than one treatment arm, it would be good to have a larger comparison group than the treatment arms. The proportion for encouragement-100 was p = 0.274, and the proportion for encouragement-50 was p = 0.366. Thus, the encouragement-100 group required a minimum sample size of 85, and the encouragement-50 group required a minimum sample size of 97.
Although PICS bags were introduced in the study area, farmers were not so familiar with them, which could have led to higher rates of non-compliance in adopting PICS bags during the experiment. Thus, we decided to increase the sample size to be higher than the required minimum to increase the sample in each treatment group, also considering the other constraints of budget, time, and the availability of research assistants. Considering all of these factors, 700 out of 2090 smallholding farmers in the list were selected randomly, and 50% were allocated to the control group, 30% to the encouragement-50 group, and 20% to the encouragement-100 group. Random selection was carried out through the STATA-17 software (STATACorp LLC, TX, US). Four farmers were not found during the baseline survey; hence, they were excluded, and 696 samples remained valid in the baseline survey.
We performed the end-line survey from 11 to 31 January 2024, the period when farmers start preparing for the next plantation of the crop and start to use the stored crops for various purposes, to collect outcome data. In the end-line survey, four farmers were excluded due to death, and other family members were not available due to absence when our enumerator visited the farmers. Finally, the sample size for data analysis was 692 smallholding farmers, comprising 347 farmers (50.14%) in the control group, 149 farmers in the encouragement-100 group (encouraged with 100% subsidy for up to 3 bags) (21.53%), and 196 farmers in the encouragement-50 group (encouraged with 50% subsidy for up to 3 bags) (28.32%).
2.3. Summary Statistics
Table 1 illustrates the descriptive statistics for our sample (post-attrition), including demographics, economic, agricultural, and post-harvest management information. As seen in panel A, the average age of household heads is 47.65 years with 6.08 years of education. Over three-fifths (62.4%) of household heads are male. The average size of the family (household) is 5.85. Moreover, the average ages of household heads in the control, encouragement-100 group, and encouragement-50 group are 47.35, 48.74, and 47.37 years, respectively.
Panel B shows that the average annual income of farmer households is NPR 279.66 thousand (USD 1 = NPR 133.78 as of December 2023) with an average cereal income of NPR 18.42 thousand at the farmgate price, and the average cash crop, forest, livestock, and off-farm income values are NPR 8.85 thousand, NPR 0.77 thousand, NPR 26.77 thousand, and NPR 217.69 thousand, respectively. The farmgate income of cereals crops is too low in comparison to off-farm income.
Panel C shows that the average agri-experience of household heads is 17.23 years. Their average agri-land size is 9.14 ropani (1 ha = 19.66 ropani). The average area of maize planting is 5.63 ropani, and the average production of maize is 520 kg. Also, the average maize production of households in the control, encouragement-100, and encouragement-50 groups are 518 kg, 538.2 kg, and 509.6 kg, respectively. Some 133 kg of maize, on average, is used for food consumption, 389 kg for animal feeding, and 27 kg for other purposes (brewing and making dry food in the winter season). Hence, the length of storage depends on the quantity produced, ranging from 2 months to 17 months. In addition, only 3.8% of the respondents used to sell maize grain.
Finally, panel D provides facts on post-harvest management. The average maize storage duration is 8.67 months. In the baseline survey, 89% and 69% of households reported that they have faced insect (maize weevil) and moth problems in the post-harvest storage of maize, respectively. A total of 44.32 kg of maize was lost in storage. Similarly, post-harvest storage losses among the control, encouragement-100, and encouragement-50 groups are 46.06 kg, 44.83 kg, and 40.84 kg. Only a small proportion (2% in the baseline survey) of farmers are adopting new technology (earlier HST is not regarded to be as effective as PICS bags by the farmers). In contrast, a significant proportion of households are still relying on traditional storage techniques. Thus, policy intervention is needed for the enhancement of post-harvest management to decrease post-harvest storage losses.
2.4. Balance Check
We conducted a balance test using the paired t-test to check the similarity between the groups with baseline characteristics. Table 2 shows the results from the balance test. We found that the pretreatment difference between the control group vs. the encouragement-100 group and the control group vs. the encouragement-50 group were balanced on average. Also, the variables were balanced between the encouragement-50 group vs. the encouragement-100 group.
2.5. Experimental Design
We used the randomized encouragement design, which functions as an instrumental variable (IV). We designed our experiment to assess the effect of PICS bag adoption with different levels of encouragement. This allowed us to estimate the effect in each pair of encouragement levels as suggested in [55]. To promote the use of hermetic storage bags, we randomly assigned 100 and 50 percent subsidy offers for the price of up to three PICS bags following the government practice of offering subsidies while promoting the usage of new machineries and technology. A total of 149 smallholding farmers received a 100% subsidy offer, 196 smallholding farmers received a 50% subsidy offer, and 347 farmers in the control group did not receive any subsidy offer. Due to randomization at the household level, there might be a chance of spillover of subsidy offer information among the groups. To minimize the possible spillover, our enumerator visited each household of the encouragement groups with 100% and 50% subsidies between 12 and 20 September 2023 without prior information and offered a price subsidy. First, we distributed the bags to the encouragement-50 group and later distributed them to the encouragement-100 group to avoid the disappointment effect and the leakage of information.
Local agriculture extension officials and enumerators played central roles in coordinating with smallholding farmers and distributing PICS bags. During home visits, enumerators offered subsidies to the farmers, who then decided whether they wanted to receive PICS bags or not. Subsequently, agriculture extension officers coordinated with the enumerators to distribute the bags from each ward office and nearby locations. Farmers were provided instructions on the proper use of PICS bags during the time of distribution. In the follow-up survey, enumerators ensured the correct handling of PICS bags. Figure 2 presents the detailed experimental design of our study. Based on the random assignment of subsidy offers, our study included the following three groups.
Control group: This group did not receive any offer of subsidy. They were provided 10–20 min of information on the harvesting of maize, on hermetic storage bags for seed and grain storage, and steps for the proper use of PICS bags during the baseline survey (see Supplementary Material Description S1).
Encouragement-100 group: This group received a 100% subsidy offer for the price of up to three bags and the same information as the control group.
Encouragement-50 group: This group received a 50% subsidy offer for the price of up to three PICS bags and the same information as the control group.
2.5.1. Treatment Variable
In this study, our treatment variable is the number of PICS bags used by smallholding farmers; each PICS bag can store 50 kg of shelled maize. The cost per bag is about USD 2.5, equivalent to about 6 kg of maize in the local market. Every smallholding maize farmer in the study site is free to use PICS bags. Hence, the treatment has an endogenous nature. Farmers can buy such kinds of PICS bags from local agro-vets (retailers of agricultural inputs), but they need to make an order in advance which, when available, often takes several days for delivery. The intervention period was three months, from the first week of October 2023 to the first week of January 2024, to estimate the treatment effect. We chose this period for intervention because, in the study area, farmers start to store newly harvested maize in September and start to prepare for plantation of the crop for the following year at the end of January. Starting from February, they start consuming the stored maize grains. Those with more grains stored for longer periods consume the crops as needed thereafter.
2.5.2. Instrumental Variables
In Nepal, federal, provincial, and local governments mostly subsidize farmers’ adoption of modern agriculture machinery or technology with full or 50% subsidies. So, it would be crucial to investigate which level of subsidy is a better option from policymakers’ perspectives to optimize the government budget and reach more farmers. Thus, we used two encouragements: 100% and 50% subsidy offers.
Our two random encouragements, 100% and 50% subsidy offers for PICS bags, are instrumental variables in the adoption of PICS bags. To measure the impact of the adoption of PICS bags, the instrumental variable could address the endogeneity issue if the assumptions of IVs are satisfied [48]. We evaluated the impact of the number of PICS bags adopted due to encouragement on outcomes, which is a causal effect of variable treatment. Both IVs, the 100% and 50% subsidy offers, satisfied all conditions of average causal response theory independence, including exclusion, first-stage, and monotonicity, as defined in [50]. We provide details about the validation of our instruments in the Results Section.
2.5.3. Outcomes
In this study, we examined the impact of the number of PICS bag treatments used on two outcome variables that responded to the end-line survey. The first outcome measures the quantity of maize stored using PICS bags in kg. The second outcome measures the total quantity of maize lost in storage, self-reported by farmers in kg. To measure storage losses, we first compared losses in storage with a PICS bag to losses in traditional storage methods and calculated the loss difference for each farmer within the group. Then, the differences in storage losses were compared among the groups. To ensure the losses were accurate, the enumerators checked the stored grains and confirmed the proportion of insect pest infestation.
2.5.4. Compliance Proportion
In our encouragement design, the compliance rate for the adoption of PICS bags by farmers in the encouragement-100 group is 0.66, while the compliance rate of farmers in the encouragement-50 group is 0.52 (Appendix A, Table A2). The same table shows that 92% of smallholding farmers use bags when they receive them for free, and 80% buy at least one PICS bag with 50% subsidies. Despite giving them brief information on hermetic storage bags for seed and grain storage, how they work, and the steps for their proper use during the baseline survey, the compliance rate was lower than expected.
2.6. Estimation Strategy
Our main objective was to measure the causal effect of the number of PICS bags used on outcomes (i.e., storage quantity of maize and post-harvest storage losses). For this, we used the two-stage least square (2SLS) method to estimate the local average treatment effects (LATEs) for the encouragement groups relative to the control group. Also, we estimated the LATE for the encouragement-100 group relative to the encouragement-50 group. Refs. [50,56,57] suggested that 2SLS can draw a weighted average per unit causal effect of continuous treatment. Continuous treatment was carried out to estimate the weighted average per unit causal effect, and we generated the weight for the treatment by following the methods used in [50,56,57]. To estimate the effects of treatment on the different outcomes of interest through two encouragement levels used as IVs, we set up the following first-stage and second-stage equations:
First-stage equation:
Ni = α1 + ϕ1Ei + γ1Xi + ε1i(1)
where Ni is the number of PICS bags; N = {0,1,2,…. }, representing the number of bags adopted by farmers; Ei is IV dummy 1 if farmers were assigned a subsidy offer (100%/50% subsidy offer, with only one subsidy being used as an instrument at one time); Xi is a set of control variables (age, education (year), family size, and gender); and ε1i is the error term.And second-stage estimates were obtained using the following equation after obtaining the first-stage fitted value, i.
Second-stage equation:
(2)
where Yi represent the outcomes (two outcomes: storage quantity and post-harvest storage losses), i is the fitted value of treatment, Xi is a set of the control variables (same as those in the first-stage equation), λ is the local average treatment effect (LATE), and ε2i is the error term.3. Results
3.1. The Impact of Subsidy Offers on PICS Bag Adoption in the First Stage
Encouraging smallholding farmers to use hermetic storage technology that reduces post-harvest storage losses, increases the storage period, and improves the grain quality on farm harvest grain seems noteworthy as it enhances the socio-economic benefits and food security of smallholder farmers [11,41]. Our experiment also basically supports this statement. Resolving socio-economic constraints in the adoption of new agricultural technology is always an independent concern of policymakers and academicians.
The results of the first stage show that encouragement with subsidy offers of 100% and 50% is highly predicated on the number of PICS bags used (Table 3). Columns (1) and (2) of the table indicate that the coefficient for encouragement with a 100% subsidy offer without and with covariates is 0.76 for both, which is significant at the 1% level (p-value < 0.01). Columns (3) and (4) of the same table illustrate that the coefficient for encouragement with a 50% subsidy offered without and with covariates is 0.41 for both, which is significant at the 1% level (p-value < 0.01). The coefficients are significantly far from zero, which indicates that encouragement with 100% and 50% subsidies are proper instrumental variables. Thus, the findings of the first stage show the impact of different subsidy levels on the number of PICS bags adopted.
3.2. The Impact of PICS Bag Adoption on Storage Quantity
Table 4 presents the local average treatment effects (LATEs) on storage quantity outcomes through two encouragements, 100%, and 50% subsidy offers. Columns (1) and (2) of the table show LATEs without and with covariates, respectively, for the farmers who were encouraged with a 100% subsidy. Likewise, columns (3) and (4) of the table show LATEs without and with covariates, respectively, for the farmers who were encouraged with a 50% subsidy. The results indicate that an additional unit of a PICS bag adopted increases maize storage by 42.84 and 42.86 kg without and with covariates, respectively, for those encouraged with a 100% subsidy. Similarly, it increases maize storage by 39.20 kg without and with covariates, respectively, for farmers encouraged with a 50% subsidy offer. The results show that there is a positive and significant effect (p-value <0.001) on storage quantity due to PICS bag adoption despite the controlling variables (Table 4).
Among the compliers in the 100% subsidy group, those who adopted one bag stored 36.59 kg, and those adopting two and three bags stored 82.5 kg and 122 kg, respectively. Similarly, among the compliers in the 50% subsidy group, those who adopted one bag stored 35.44 kg, and those adopting two and three bags stored 81 kg and 130 kg, respectively. The significant increases in storage quantity reflect a notable behavioral change towards the adoption of improved storage methods through the use of PICS bags. Thus, the results highlight the storage behavioral changes in smallholding farmers through the adoption of PICS bags.
3.3. Impact of PICS Bag Adoption on Post-Harvest Storage Losses
Table 5 provides LATE estimates on post-harvest storage loss outcomes through two encouragements: 100% and 50% subsidy offers. Columns (1) and (2) of the table illustrate LATEs without and with covariates, respectively, for farmers who were encouraged with a 100% subsidy. Similarly, columns (3) and (4) of the table illustrate LATEs without and with covariates, respectively, for farmers who were encouraged with a 50% subsidy. The results show that an additional unit of PICS bag adopted reduces post-harvest storage losses by 10.05 kg and 10.02 kg without and with covariates, respectively, for farmers who were encouraged with a 100% subsidy offer (Table 5). These values are consistent and significant at the 5% level (p-value < 0.05).
There was no significant effect on post-harvest storage losses for those encouraged with a 50% subsidy. This was due to the low adoption rate of the number of PICS bags. As described in the measurement of outcome storage losses, lower adoption leads to higher storage losses.
3.4. Impact on PICS Bag Adoption and Outcomes of Interest: 100% vs. 50% Subsidy Encouragement
Table 6 illustrates the first stage and IV estimates for farmers encouraged with a 100% subsidy relative to those encouraged with a 50% subsidy. Columns (1) and (2) of the table show the impact of a 100% subsidy on the number of PICS bags adopted without and with covariates, respectively, with coefficient values of 0.35 and 0.33 which are significant at the 1% level (p-value < 0.01), indicating that the 100% subsidy offer is a proper IV to estimate the impact of the number of PICS bags adopted on outcomes of interest. Similarly, columns (3) and (4) of the table represent the coefficient of the impact of PICS bag adoption on the maize storage quantity for farmers encouraged with a 100% subsidy relative to those encouraged with a 50% subsidy without and with covariates, respectively.
The coefficients suggest that one additional unit of PICS hermetic bag adopted results in maize storage of 47.15 and 47.68 kg without and with covariates, respectively, which are statistically significant at the 1% level (p-value < 0.01) and consistent regardless of the covariates.
Columns (5) and (6) of the table show the impacts on post-harvest storage losses with coefficient values of −8.24 and −10.94 without and with covariates, respectively (Table 6). However, the results show that the impact of PICS bag adoption on post-harvest storage losses is insignificant for farmers encouraged with a 50% subsidy in comparison to farmers who are encouraged with a 100% subsidy.
3.5. Instrument Validity
Our random encouragements, namely 100% and 50% subsidies offered to randomly selected farmers, are instrumental variables for those farmers. Our instruments satisfied the following conditions and are appropriate for estimating the impact of the number of PICS bags adopted on the storage quantity of maize and post-harvest storage losses.
3.5.1. Condition 1: Relevance
The first condition is that the instrumental variable must correlate with the endogenous regressor (treatment variable), Cov (Z, D) 0. Table 3 shows that encouragement with 100% and 50% subsidy offers are highly correlated with the use of the hermetic bag, and the coefficients of both encouragement levels are significant at a 1% level. Also, our IVs have higher first-stage F-statistics than the benchmark value, exceeding 10 (see Appendix A, Table A3) for strong IVs [58]. Hence, our IVs are acceptable to estimate the impact of PICS bags on the outcomes of interest.
3.5.2. Condition 2: Independence
The instrumental variable must not correlate with unobservable factors (Cov (Z, u) 0). In this study, we randomly offered 100% and 50% subsidies to smallholding farmers, so this may have distributed unobservable factors to all groups equally. The results from the balance check suggest the randomness of baseline variables between the encouragement-100 group relative to the control group, the encouragement-50 group relative to the control group, and the encouragement-100 group relative to the encouragement-50 group (see Table 2). This supports the validation of our instrumental variables.
3.5.3. Condition 3: Exclusion Restriction
Violation of the exclusion restriction condition leads to IV estimate bias and inconsistency [58]. Although we cannot conduct a direct empirical test for exclusion restriction, we can provide precise intuition [59]. Our random subsidy offers for PICS bags only provide smallholding farmers with access to the bags, and only after using the PICS bags can farmers store maize, which might reduce insect pest infestation while lowering storage losses. Hence, our encouragements (IVs) do not have a direct impact on outcomes if the PICS bags are not used. Due to our experimental setup, we ran a separate regression model for a 100% subsidy offer and a 50% subsidy. An equal number of exogenous instruments and endogenous variables could reduce the bias to near zero [60]. Therefore, our IVs satisfy the exclusion restriction condition.
3.5.4. Condition 4: Monotonicity
No farmers ‘defy’ the subsidy offers. Encouragement through subsidy offers motivates farmers to adopt hermetic storage bags, and the subsidy offers themselves do not demotivate them. When farmers do not receive subsidy offers, they are not motivated to adopt the bags, but they are also not demotivated by the absence of subsidy offers. In our design, farmers can decide not to adopt PICS bags if they do not want to.
3.6. Robustness Check
We estimated the LATEs for the outcomes using the limited information maximum likelihood (LIML) estimator and the generalized method of momentum (GMM) estimators to check the preciseness and consistency of our results (see Appendix A, Table A4 and Table A5).
4. Discussion
In this study, we delved into the causal impact of PICS bags on storage quantity and post-harvest storage losses among smallholding farmers in the rural mid-hills of Nepal, encouraging them to use HST by offering them subsidies for up to three PICS bags after providing HST information to all the farmers.
The findings from our study suggest that the adoption of PICS bags has significant effects on storage quantity when smallholding farmers in Nepal in both the encouragement-100 and encouragement-50 groups use PICS bags after being encouraged with subsidies. Hence, our findings on storage quantity contradicts the findings of [11,13,46]. They did not find a statistically significant impact on the storage quantity of maize in their studies, but they reported the ITT effect after conducting RCTs. A possible explanation for this discrepancy is the difference in treatment size relative to the average maize harvest. In their studies, the treatment size was too small compared to the average harvest, but in this study, the treatment could lead to up to approximately 30% of the average harvested maize being stored. Our findings on post-harvest storage loss resonate with the findings of [43]. They found that PICS bags with training reduced post-harvest storage losses by 6.1% points (76.8% reduction in storage losses). However, this finding might not be free from endogeneity bias since technology adoption is not random, and there is a chance of self-selection bias [44,45]. Also, our findings for post-harvest storage losses are in line with the findings of a 2.3% reduction in storage losses in Uganda [11], with hermetic bags causing a PHL reduction of 5.1% in Tanzania [10]. But our findings contradict the findings of [46], where the authors found the null effect (ITT) in post-harvest storage losses in Ethiopia.
Our insignificant result obtained for storage losses for the encouragement-50 group relative to the control group and that for the encouragement-100 group relative to the encouragement-50 group are mainly due to two reasons. First, most of the encouraged farmers adopted only a single PICS bag. Although we provided a subsidy for up to three PICS bags, which can store a maximum of 150 kg of shelled maize, most farmers did not take full advantage of this offer. The main reason for this might be the need for extra labor and time to shell the maize into grains before storing it in PICS bags. Traditionally, farmers store maize with husks in various ways, as mentioned in an earlier section. The out-migration of people for work has increased throughout Nepal, with 5.9 million labor permits provided in the 2023–24 period for working abroad [21]. The Lumbini Province is one of the provinces with high numbers of out-migration. This suggests a potential shortage of labor in rural regions of the province, including the studied rural municipality. This could be another reason for the low adoption of PICS bags in addition to the fact that a very high percentage of farmers took/bought PICS bags.
Second, due to the short period of intervention, farmers reported less insect pest infestation in the early storage period. Hence, they did not use PICS bags, although they took the bags if they had subsidy offers. They can use the PICS bags later, even in the next season and the following years. Generally, PICS bags can be used effectively for more than two years.
Our paper, indicating the behavioral change in farmers regarding the post-harvest management of grains, mainly has three contributions to the emerging literature on improved storage technology for the post-harvest storage of grains. First, we measured the causal impact of hermetic storage on storage quantity and post-harvest storage losses with a randomized encouragement design using an IV approach. This can address endogeneity bias by measuring the impact of the endogenous regressor [61]. Also, it provides a more rigorous and trustworthy causal impact than when measuring the effect of random treatment assignment on the outcome [48]. Since technology adoption is endogenous [44,45], previous RCTs on storage technology could not be free from endogeneity biases.
Second, most earlier studies measured the combined effect with other factors (e.g., training, demonstration, group storage, and community storage). They lack evidence for the singled-out effect of using the PICS bag. We estimated the effect of independent PICS bags disentangled from other factors. Thus, our findings show a more precise treatment effect of the PICS bag, reducing the effects of the other factors.
Third, to our knowledge, this is the first study which provides causal evidence on post-harvest storage management in rural mid-hills in Nepal. A previous descriptive study was conducted, titled “On-farm Grain Storage and Challenges in Bagamati Province in Nepal” [30].
Based on this study’s findings, we suggest two policy implications for improved storage and sustainable agri-food systems.
First, encouraging smallholding farmers with 100% subsidies leads to a broader adoption of HST. It increases the storage of maize and reduces post-harvest storage losses. There is a higher quantity of stored maize when using PICS bags, which reduces storage losses. It can enhance local food availability and accessibility. Similarly, encouraging famers with a 50% subsidy offer also increases the storage quantity through using PICS bags. The increase in storage quantity poses a behavioral change regarding HST. It improves the grains’ quality and controls insect infestation [37,38]. Thus, encouraging farmers by offering subsidies could be a viable policy option for post-harvest management, with preference to be given to the 100% subsidy considering the readily available PICS bags in such hilly rural municipalities.
Second, our study is the first study using a randomized encouragement design in agriculture economics, which used two values of encouragement. This allowed us to compare the impact of the PICS bag when smallholding farmers are encouraged with 100% and 50% subsidies. The results suggest that a 100% subsidy offer is more effective in storage quantity than a 50% subsidy, which has no significant effect on post-harvest storage losses. This shows that encouraging farmers with a 100% subsidy offer could be a better policy option to improve the storage system. On the other hand, a 100% subsidy imposes a significant burden on the government budget.
In contrast, a 50% subsidy offer is cost-effective for developing nations and will still provide long-term benefits. Encouraging farmers with a 50% subsidy offer could be traded off for a more significant number of farmers for HST adoption and can help control storage problems with a smaller budget. Hence, encouraging smallholding farmers with a 50% subsidy might be a feasible, sustainable, and expandable policy option for policymakers in developing countries.
A limitation of our study may be the intervention period, which was a bit short considering the storage period of maize of 8.61 months on average in the study area. However, some studies in the literature emphasize that a short-term subsidy for agriculture technology increases its demand and future adoption [36,62]. In future, it is recommended to conduct a study with a longer period of intervention to measure the impact of hermetic storage technology more robustly, including the cost-effectiveness of the subsidy levels.
5. Conclusions
The adoption of improved storage systems, such as PICS bags, can play a vital role in storage quantity, storage behavior change, and post-harvest storage loss reduction through encouragement with subsidy offers for PICS bags among smallholding farmers. This matters for enhancing food security as well as reducing hunger. This study aimed to measure the causal impact of PICS bags on storage quantity and post-harvest storage losses among smallholding maize farmers in Nepal with experimental data.
We conducted a randomized encouragement design using an IV among 692 smallholding maize farmers in the Mathagadhi rural municipality in the Lumbini Province in Nepal. We found a significant impact on storage quantity and post-harvest storage losses among smallholding farmers who were encouraged with 100% subsidy offers. Similarly, we revealed a significant impact on storage quantity, whereas no significant effect on storage losses was found in smallholding farmers who were encouraged with a 50% subsidy. In addition, compared to smallholding farmers encouraged with a 50% subsidy, those encouraged with 100% offers were likely to store more maize, although the impacts on post-harvest storage losses were insignificant.
For the sustainable adoption of PICS bags, policymakers need to consider the long-term feasibility of subsidy programs. While the findings highlight the effectiveness of 100% subsidies, implementing such a policy may cause government budget constraints. A phased or tiered subsidy approach could help balance budget constraints with program effectiveness. For instance, governments could begin with a higher subsidy level and gradually reduce it as adoption rates increase, leveraging the demonstration effect to encourage non-subsidized uptake. Moreover, training smallholding farmers to use PICS bags and market access can ensure their effective utilization and longevity. Providing technical support and creating awareness campaigns about the benefits of hermetic storage technology could further drive adoption.
Conceptualization, D.K.N. and K.L.M.; methodology, D.K.N. and K.L.M.; software, D.K.N.; validation, D.K.N. and K.L.M.; formal analysis, D.K.N. and K.L.M.; investigation, D.K.N.; resources, D.K.N. and K.L.M.; data curation, D.K.N.; writing—original draft preparation, D.K.N. and K.L.M.; writing—review and editing, D.K.N. and K.L.M.; visualization, D.K.N.; supervision, K.L.M.; project administration, D.K.N.; funding acquisition, D.K.N. and K.L.M. All authors have read and agreed to the published version of the manuscript.
This study was conducted in accordance with the Declaration of Helsinki and the survey questionnaire and experimental design were approved by the Institutional Review Board and research ethics review board of the Graduate School of Humanities and Social Sciences, Hiroshima University, Japan.
The corresponding author will make the data and codes available upon reasonable request.
The APC of this paper is partially funded by the Association for Regional Agricultural and Forestry Economics (ARAFE). This paper is based on the draft “The Impact of Hermetic Storage Technology Adoption on Storage Quantity and Post-harvest Storage Losses among Smallholders in Nepal: Evidence from Randomized Promotion” previously presented at the ARAFE 8th International Workshop held on 29 June 2024 at Aichi Gakuin University, Nagoya, Japan. The paper was revised by incorporating the feedback from the workshop. The authors would like to thank ARAFE for this. The authors and funding agency have no conflicts of interest. The authors are thankful for the Project for Human Resource Development Scholarship (JDS) organized by Japan Development Service Co., Ltd. (JDS), which facilitated the master’s study of D.K.N at Hiroshima University, Japan. The authors also acknowledge the cooperation of the farmers who participated in the field experiment.
The authors declare no conflicts of interest.
Footnotes
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
Enlarge this image.Figure 1. Study site map. Note: Numbers 1–8 represent the ward numbers of Mathagadhi Rural Municipality, and the triangle-marked wards indicate the areas where the study was conducted.
Descriptive statistics for the baseline variable.
| Variables | Full Sample | Control | Encouragement-100 Group | Encouragement-50 Group | ||||
|---|---|---|---|---|---|---|---|---|
| (n = 692) | (n = 347) | (100% Subsidy Offer) | (50% Subsidy Offer) | |||||
| (n = 149) | (n = 196) | |||||||
| Mean | SD | Mean | SD | Mean | SD | Mean | SD | |
| Panel A. Demographics | ||||||||
| Age | 47.65 | 13.92 | 47.35 | 13.86 | 48.74 | 13.96 | 47.37 | 14.05 |
| Education (year) | 6.08 | 3.71 | 6.16 | 3.57 | 6.09 | 3.94 | 5.93 | 3.78 |
| Gender | 0.62 | 0.48 | 0.64 | 0.48 | 0.63 | 0.48 | 0.59 | 0.49 |
| Family size | 5.85 | 2.48 | 5.88 | 2.43 | 6.04 | 2.72 | 5.66 | 2.39 |
| Panel B. Economical data (income in thousands of Nepalese rupees—NPR) | ||||||||
| Family income | 279.66 | 231.21 | 270.68 | 220.78 | 293.54 | 276.40 | 293.54 | 276.40 |
| Cereal income | 18.42 | 25.22 | 19.12 | 27.51 | 16.24 | 22.22 | 16.24 | 22.22 |
| Cash crop income | 8.85 | 28.60 | 10.86 | 37.57 | 6.82 | 14.54 | 6.82 | 14.54 |
| Forest income | 0.77 | 4.42 | 0.92 | 5.75 | 0.52 | 1.92 | 0.52 | 1.92 |
| Off-farm income | 217.69 | 225.68 | 205.08 | 214.82 | 234.65 | 264.38 | 234.65 | 264.38 |
| Livestock income | 26.77 | 32.01 | 27.66 | 37.13 | 27.73 | 32.19 | 27.73 | 32.19 |
| Panel C. Agricultural data | ||||||||
| Agri-experience (year) | 17.23 | 10.78 | 17.32 | 11.24 | 17.35 | 9.74 | 16.98 | 10.74 |
| Agri-land size (ropani) | 9.14 | 6.15 | 9.01 | 6.06 | 9.43 | 6.57 | 9.14 | 6 |
| Maize production quantity (Kg) | 520 | 364 | 518 | 359.69 | 538.2 | 372.6 | 509.6 | 366.2 |
| Maize plant area (ropani) | 5.63 | 3.86 | 5.59 | 3.84 | 5.81 | 3.91 | 5.57 | 3.89 |
| Panel D. Post-harvest management data | ||||||||
| Storage period (month) | 8.67 | 2.92 | 8.63 | 2.91 | 8.77 | 2.8 | 8.64 | 3.05 |
| Insects (dummy) | 0.89 | 0.31 | 0.89 | 0.32 | 0.91 | 0.29 | 0.89 | 0.32 |
| Moths (dummy) | 0.69 | 0.46 | 0.67 | 0.47 | 0.73 | 0.44 | 0.69 | 0.46 |
| Storage loss | 44.32 | 52.7 | 46.06 | 57.07 | 44.83 | 48.81 | 40.84 | 47.33 |
| New technology (dummy) | 0.02 | 0.14 | 0.02 | 0.15 | 0.03 | 0.16 | 0.01 | 0.07 |
Source: Field Survey, 2023 August. Note: SD = Standard Deviation, USD 1 = NPR 133.78 as of December 2023, and 1 ha = 19.66 ropani.
Baseline comparison among groups.
| Variables | Encouragement-100 vs. Control | Encouragement-50 vs. Control | Encouragement-50 vs. Encouragement-100 | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Diff | S. E. | t-Value | Diff | S. E. | t-Value | Diff | S. E. | t-Value | |
| Panel A. Demographics | |||||||||
| Age | −1.39 | 1.36 | −1.02 | −0.02 | 1.24 | −0.01 | −1.37 | 1.52 | −0.90 |
| Education (year) | 0.07 | 0.36 | 0.19 | 0.23 | 0.33 | 0.71 | −0.16 | 0.42 | −0.38 |
| Gender | 0.01 | 0.05 | 0.25 | 0.06 | 0.04 | 1.29 | −0.04 | 0.05 | −0.83 |
| Family size | −0.16 | 0.25 | −0.65 | 0.22 | 0.22 | 1 | −0.38 | 0.28 | −1.37 |
| Panel B. Economical (income in thousands of Nepali rupees (NPR)) | |||||||||
| Family income | −22.87 | 23.39 | −0.98 | −14.32 | 19.43 | −0.73 | −8.54 | 26.25 | −0.33 |
| Cereal income | 2.88 | 2.55 | 1.13 | 0.28 | 2.32 | 0.12 | 2.61 | 2.47 | 1.06 |
| Cash crop income | 4.04 | 3.17 | 1.27 | 4.01 | 2.8 | 1.43 | 0.03 | 1.60 | 0.02 |
| Forest income | 0.4 | 0.48 | 0.82 | 0.23 | 0.44 | 0.52 | 0.17 | 0.26 | 0.65 |
| Off-farm income | −29.57 | 22.60 | −1.31 | −22.03 | 19.10 | −1.15 | −7.54 | 25.65 | −0.29 |
| Livestock income | −0.07 | 3.50 | −0.02 | 3.17 | 2.86 | 1.11 | −3.24 | 2.81 | −1.15 |
| Panel C. Agricultural data | |||||||||
| Agri-experience (year) | −0.03 | 1.06 | −0.03 | 0.34 | 0.99 | 0.34 | −0.37 | 1.12 | −0.33 |
| Agri-land size (ropani) | −0.42 | 0.61 | −0.69 | −0.13 | 0.54 | −0.24 | −0.29 | 0.68 | −0.43 |
| Maize production quantity (Kg) | −20.18 | 35.61 | −0.57 | 8.38 | 32.35 | 0.26 | −28.56 | 40.11 | −0.71 |
| Maize plant area (ropani) | −0.22 | 0.38 | −0.57 | 0.02 | 0.34 | 0.06 | −0.23 | 0.42 | −0.55 |
| Panel D. Post-harvest management data | |||||||||
| Storage period (month) | −0.14 | 0.28 | −0.49 | −0.004 | 0.26 | −0.02 | −0.13 | 0.32 | −0.42 |
| Insects (dummy) | −0.02 | 0.03 | −0.7 | −0.003 | 0.03 | −0.11 | −0.02 | 0.03 | −0.55 |
| Moths (dummy) | −0.06 | 0.05 | −1.33 | −0.02 | 0.04 | −0.54 | −0.04 | 0.05 | −0.76 |
| Storage loss | 1.23 | 5.36 | 0.23 | 5.22 | 4.8 | 1.09 | −4.00 | 5.21 | −0.77 |
| New technology (earlier HST) (dummy) | 0 | 0.02 | −0.25 | 0.02 | 0.01 | 1.57 | −0.02 | 0.01 | −1.68 |
Source: Field Survey, 2023 August. Note: Diff = mean difference; SE = standard error.
The impact of subsidy offers on PICS bags adoption in the first stage.
| (1) | (2) | (3) | (4) | |
|---|---|---|---|---|
| Variables | No. of PICS Bags | No. of PICS Bags | No. of PICS Bags | No. of PICS Bags |
| 100% subsidy offer | 0.76 *** | 0.76 *** | ||
| 50% subsidy offer | 0.41 *** | 0.41 *** | ||
| Constant | 0.00 *** | −0.21 * | 0.00 | −0.10 * |
| Observations | 496 | 496 | 543 | 543 |
| R-squared | 0.29 | 0.30 | 0.19 | 0.20 |
| Covariates | No | Yes | No | Yes |
Notes: robust standard errors in parentheses; *** p < 0.01, and * p < 0.1; and covariates: age, education (year), family size, and gender.
Impact of PICS bag adoption on storage quantity.
| (1) | (2) | (3) | (4) | |
|---|---|---|---|---|
| Variables | Storage Quantity Using PICS Bag (kg) | Storage Quantity Using PICS Bag (kg) | Storage Quantity Using PICS Bag (kg) | Storage Quantity Using PICS Bag (kg) |
| No. of PICS bags | 42.84 *** | 42.86 *** | 39.20 *** | 39.20 *** |
| Constant | −0.00 *** | 2.61 | 0.00 *** | 0.97 |
| Observations | 496 | 496 | 543 | 543 |
| R-squared | 0.85 | 0.85 | 0.92 | 0.92 |
| Covariates | No | Yes | No | Yes |
Notes: robust standard errors in parentheses; *** p < 0.01; and covariates: age, education (year), family size, and gender.
Impact of PICS bag adoption on post-harvest storage losses.
| (1) | (2) | (3) | (4) | |
|---|---|---|---|---|
| Variables | Post-Harvest Storage Losses (kg) | Post-Harvest Storage Losses (kg) | Post-Harvest Storage Losses (kg) | Post-Harvest Storage Losses (kg) |
| No. of PICS hermetic bags | −10.05 ** | −10.02 ** | −11.93 | −9.37 |
| Constant | 30.53 *** | 9.89 | 30.67 *** | 5.61 |
| Observations | 496 | 496 | 543 | 543 |
| R-squared | 0.02 | 0.05 | ||
| Covariates | No | Yes | No | Yes |
Notes: robust standard errors in parentheses; *** p < 0.01, and ** p < 0.05; and covariates: age, education (year), family size, and gender.
Impact on PICS bag adoption and outcomes of interest: 100% vs. 50% subsidy encouragement.
| (1) | (2) | (3) | (4) | (5) | (6) | |
|---|---|---|---|---|---|---|
| First Stage | Second Stage | |||||
| Variables | No. of PICS Bag | No. of PICS Bag | Storage Quantity Using PICS Bag (kg) | Storage Quantity Using PICS Bag (kg) | Post-Harvest Storage Losses (kg) | Post-Harvest Storage Losses (kg) |
| 100% subsidy offer | 0.35 *** | 0.33 *** | ||||
| No. of PICS bags | 47.15 *** | 47.68 *** | −8.24 | −10.94 | ||
| Observations | 345 | 345 | 345 | 345 | 345 | 345 |
| R-squared | 0.04 | 0.06 | 0.82 | 0.82 | ||
| Covariates | No | Yes | No | Yes | No | Yes |
Notes: robust standard errors in parentheses; *** p < 0.01.
Supplementary Materials
The following supporting information can be downloaded at
References
1. FAO State of Food and Agriculture 2019. Moving Forward on Food Loss and Waste Reduction; Food & Agriculture Org: Rome, Italy, 2019; ISBN 9789251317891
2. Delgado, L.; Schuster, M.; Torero, M. Food Losses in Agrifood System: What We Know. Annu. Rev. Resour. Econ.; 2023; 15, pp. 41-62. [DOI: https://dx.doi.org/10.1146/annurev-resource-072722-025159]
3. FAO How to Feed the World in 2050. Available online: https://www.fao.org/fileadmin/templates/wsfs/docs/expert_paper/How_to_Feed_the_World_in_2050.pdf (accessed on 24 April 2024).
4. Godfray, H.C.J. The Challenge of Feeding 9–10 Billion People Equitably and Sustainably. J. Agric. Sci.; 2014; 152 (Suppl. S1), pp. 2-8. [DOI: https://dx.doi.org/10.1017/S0021859613000774]
5. Kumar, D.; Kalita, P. Reducing Postharvest Losses during Storage of Grain Crops to Strengthen Food Security in Developing Countries. Foods; 2017; 6, 8. [DOI: https://dx.doi.org/10.3390/foods6010008]
6. Abass, A.B.; Ndunguru, G.; Mamiro, P.; Alenkhe, B.; Mlingi, N.; Bekunda, M. Post-Harvest Food Losses in a Maize-Based Farming System of Semi-Arid Savannah Area of Tanzania. J. Stored Prod. Res.; 2014; 57, pp. 49-57. [DOI: https://dx.doi.org/10.1016/j.jspr.2013.12.004]
7. World Bank. Missing Food: The Case of Postharvest Grain Losses in Sub-Saharan Africa; Report No. 60371-AFR. World Bank: Washington, DC, USA, 2011.
8. Shiferaw, B.; Prasanna, B.M.; Hellin, J.; Bänziger, M. Crops That Feed the World 6. Past Successes and Future Challenges to the Role Played by Maize in Global Food Security. Food Secur.; 2011; 3, pp. 307-327. [DOI: https://dx.doi.org/10.1007/s12571-011-0140-5]
9. Tesfaye, W.; Tirivayi, N. The Impacts of Postharvest Storage Innovations on Food Security and Welfare in Ethiopia. Food Policy; 2018; 75, pp. 52-67. [DOI: https://dx.doi.org/10.1016/j.foodpol.2018.01.004]
10. Brander, M.; Bernauer, T.; Huss, M. Improved On-Farm Storage Reduces Seasonal Food Insecurity of Smallholder Farmer Households—Evidence from a Randomized Control Trial in Tanzania. Food Policy; 2021; 98, 101891. [DOI: https://dx.doi.org/10.1016/j.foodpol.2020.101891]
11. Omotilewa, O.J.; Ricker-Gilbert, J.; Ainembabazi, J.H.; Shively, G.E. Does Improved Storage Technology Promote Modern Input Use and Food Security? Evidence from a Randomized Trial in Uganda. J. Dev. Econ.; 2018; 135, pp. 176-198. [DOI: https://dx.doi.org/10.1016/j.jdeveco.2018.07.006]
12. Chegere, M.J. Post-Harvest Losses Reduction by Small-Scale Maize Farmers: The Role of Handling Practices. Food Policy; 2018; 77, pp. 103-115. [DOI: https://dx.doi.org/10.1016/j.foodpol.2018.05.001]
13. Channa, H.; Ricker-Gilbert, J.; Feleke, S.; Abdoulaye, T. Overcoming Smallholder Farmers’ Post-Harvest Constraints through Harvest Loans and Storage Technology: Insights from a Randomized Controlled Trial in Tanzania. J. Dev. Econ.; 2022; 157, 102851. [DOI: https://dx.doi.org/10.1016/j.jdeveco.2022.102851]
14. Murdock, L.L.; Margam, V.; Baoua, I.; Balfe, S.; Shade, R.E. Death by Desiccation: Effects of Hermetic Storage on Cowpea Bruchids. J. Stored Prod. Res.; 2012; 49, pp. 166-170. [DOI: https://dx.doi.org/10.1016/j.jspr.2012.01.002]
15. Erenstein, O.; Jaleta, M.; Sonder, K.; Mottaleb, K.; Prasanna, B.M. Global Maize Production, Consumption and Trade: Trends and R&D Implications. Food Secur.; 2022; 14, pp. 1295-1319.
16. Benimana, G.U.; Ritho, C.; Irungu, P. Assessment of Factors Affecting the Decision of Smallholder Farmers to Use Alternative Maize Storage Technologies in Gatsibo District-Rwanda. Heliyon; 2021; 7, e08235. [DOI: https://dx.doi.org/10.1016/j.heliyon.2021.e08235]
17. Debebe, S. Post-Harvest Losses of Crops and Its Determinants in Ethiopia: Tobit Model Analysis. Agric. Food Secur.; 2022; 11, 13. [DOI: https://dx.doi.org/10.1186/s40066-022-00357-6]
18. Bhandari, G.; Achhami, B.B.; Karki, T.B.; Bhandari, B.; Bhandari, G. Survey on Maize Post-Harvest Losses and Its Management Practices in the Western Hills of Nepal. J. Maize Res. Dev.; 2015; 1, pp. 98-105. [DOI: https://dx.doi.org/10.3126/jmrd.v1i1.14247]
19. MoALD. Statistical Information on Nepalese Agriculture 2078/79 (202122); Singha Durbar: Kathmandu, Nepal, 2023.
20. MoALD. Selected Indicators of Nepalese Agriculture; Singha Durbar: Kathmandu, Nepal, 2023.
21. MOF. Economic Survey Fiscal Year 2023-24; Singha Durbar: Kathmandu, Nepal, 2024.
22. Timsina, K.P.; Ghimire, Y.N.; Lamichhane, J. Maize Production in Mid Hills of Nepal: From Food to Feed Security. J. Maize Res. Dev.; 2016; 2, pp. 20-29. [DOI: https://dx.doi.org/10.3126/jmrd.v2i1.16212]
23. Upadhyay, I.P.; Shrestha, K.B.; Shivakoti, G.P. _Existing Farmer Practices of Maize Storage in the Eastern and Mid-Western Hills of Nepal; Regional Agriculture Research Station: Khajura, Nepal, 2002.
24. Upadhayay, I.P.; Shrestha, K.B.; Shivakoti, G.P. _A Literature Review On Post-Harvest Losses of Maize with Emphasis on Storage Losses.Pdf-1487049461; Regional Agriculture Research Station: Khajura, Nepal, 2002.
25. Paneru, R.B.; Paudel, G.P.; Thapa, R.B. Determinants of post-harvest maize losses by pests in mid hills of Nepal. Int. J. Agric. Environ. Biores.; 2018; 3, pp. 110-118.
26. PMAMP. Prime Minister Agriculture Modernization Project: Project Implementation Manual; Prime Minister Agriculture Modernization Project: Lalitpur, Nepal, 2020.
27. PMAMP. Annual Program and Progress Report (Fiscal Year 2022/23); Prime Minister Agriculture Modernization Project: Lalitpur, Nepal, 2024.
28. MoAD. Agriculture Development Strategy (ADS); Government of Nepal, Ministry of Agriculture Development: Kathmandu, Nepal, 2015; Part-1
29. Flock, J.H. Factors That Influence the Adoption of Hermetic Storage, Evidence from Nepal; Purdue University Press: West Lafayette, Indiana, 2015.
30. Kandel, P.; Kharel, K.; Njoroge, A.; Smith, B.W.; Díaz-Valderrama, J.R.; Timilsina, R.H.; Paudel, G.P.; Baributsa, D. On-Farm Grain Storage and Challenges in Bagmati Province, Nepal. Sustainability; 2021; 13, 7959. [DOI: https://dx.doi.org/10.3390/su13147959]
31. Ndaghu, N.N.; Abdoulaye, T.; Mustapha, A.; Choumbou Raoul Fani, D.; Tabetando, R.; Udeme Henrietta, U.; Lucy Kamsang, S. Gender Differentiation on the Determinants and Intensity of Adoption of Purdue Improved Cowpea Storage (PICS) Bags in Northern Nigeria. Heliyon; 2023; 9, e23026. [DOI: https://dx.doi.org/10.1016/j.heliyon.2023.e23026] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/38125509]
32. Alemu, G.T.; Nigussie, Z.; Haregeweyn, N.; Berhanie, Z.; Wondimagegnehu, B.A.; Ayalew, Z.; Molla, D.; Okoyo, E.N.; Baributsa, D. Cost-Benefit Analysis of on-Farm Grain Storage Hermetic Bags among Small-Scale Maize Growers in Northwestern Ethiopia. Crop Prot.; 2021; 143, 105478. [DOI: https://dx.doi.org/10.1016/j.cropro.2020.105478] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/33941995]
33. Manda, J.; Feleke, S.; Mutungi, C.; Tufa, A.H.; Mateete, B.; Abdoulaye, T.; Alene, A.D. Assessing the Speed of Improved Postharvest Technology Adoption in Tanzania: The Role of Social Learning and Agricultural Extension Services. Technol. Forecast. Soc. Chang.; 2024; 202, 123306. [DOI: https://dx.doi.org/10.1016/j.techfore.2024.123306]
34. Zhang, S.; Sun, Z.; Ma, W.; Valentinov, V. The Effect of Cooperative Membership on Agricultural Technology Adoption in Sichuan, China. China Econ. Rev.; 2020; 62, 101334. [DOI: https://dx.doi.org/10.1016/j.chieco.2019.101334]
35. Baributsa, D.; Njoroge, A.W. The Use and Profitability of Hermetic Technologies for Grain Storage among Smallholder Farmers in Eastern Kenya. J. Stored Prod. Res.; 2020; 87, 101618. [DOI: https://dx.doi.org/10.1016/j.jspr.2020.101618]
36. Channa, H.; Chen, A.Z.; Pina, P.; Ricker-Gilbert, J.; Stein, D. What Drives Smallholder Farmers’ Willingness to Pay for a New Farm Technology? Evidence from an Experimental Auction in Kenya. Food Policy; 2019; 85, pp. 64-71. [DOI: https://dx.doi.org/10.1016/j.foodpol.2019.03.005] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31217660]
37. Williams, S.B.; Murdock, L.L.; Baributsa, D. Storage of Maize in Purdue Improved Crop Storage (PICS) Bags. PLoS ONE; 2017; 12, e0168624. [DOI: https://dx.doi.org/10.1371/journal.pone.0168624] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/28072835]
38. Obeng-Akrofi, G.; Maier, D.E.; White, W.S.; Akowuah, J.O.; Bartosik, R.; Cardoso, L. Effectiveness of Hermetic Bag Storage Technology to Preserve Physical Quality Attributes of Shea Nuts. J. Stored Prod. Res.; 2023; 101, 102086. [DOI: https://dx.doi.org/10.1016/j.jspr.2023.102086]
39. Odjo, S.; Burgueño, J.; Rivers, A.; Verhulst, N. Hermetic Storage Technologies Reduce Maize Pest Damage in Smallholder Farming Systems in Mexico. J. Stored Prod. Res.; 2020; 88, 101664. [DOI: https://dx.doi.org/10.1016/j.jspr.2020.101664]
40. Ng’ang’a, J.; Mutungi, C.; Imathiu, S.; Affognon, H. Effect of Triple-Layer Hermetic Bagging on Mould Infection and Aflatoxin Contamination of Maize during Multi-Month on-Farm Storage in Kenya. J. Stored Prod. Res.; 2016; 69, pp. 119-128. [DOI: https://dx.doi.org/10.1016/j.jspr.2016.07.005]
41. Shukla, P.; Pullabhotla, H.K.; Baylis, K. The Economics of Reducing Food Losses: Experimental Evidence from Improved Storage Technology in India. Food Policy; 2023; 117, 102442. [DOI: https://dx.doi.org/10.1016/j.foodpol.2023.102442]
42. Aggarwal, S.; Francis, E.; Robinson, J. Grain Today, Gain Tomorrow: Evidence from a Storage Experiment with Savings Clubs in Kenya. J. Dev. Econ.; 2018; 134, pp. 1-15. [DOI: https://dx.doi.org/10.1016/j.jdeveco.2018.04.001]
43. Chegere, M.J.; Söderbom, M.; Eggert, H. The Effects of Storage Technology and Training on Post-Harvest Losses, Practices and Sales: Evidence from Small-Scale Farms in Tanzania. Econ. Dev. Cult. Chang.; 2022; 70, pp. 729-761. [DOI: https://dx.doi.org/10.1086/713932]
44. Mendola, M. Agricultural Technology Adoption and Poverty Reduction: A Propensity-Score Matching Analysis for Rural Bangladesh. Food Policy; 2007; 32, pp. 372-393. [DOI: https://dx.doi.org/10.1016/j.foodpol.2006.07.003]
45. Wu, F. Adoption and Income Effects of New Agricultural Technology on Family Farms in China. PLoS ONE; 2022; 17, 0267101. [DOI: https://dx.doi.org/10.1371/journal.pone.0267101]
46. Negede, B.M.; De Groote, H.; Minten, B.; Voors, M. Does Access to Improved Grain Storage Technology Increase Farmers’ Welfare? Experimental Evidence from Maize Farming in Ethiopia. J. Agric. Econ.; 2023; 75, pp. 137-152. [DOI: https://dx.doi.org/10.1111/1477-9552.12546]
47. Nindi, T.; Ricker-Gilbert, J.; Bauchet, J. Incentive Mechanisms to Exploit Intraseasonal Price Arbitrage Opportunities for Smallholder Farmers: Experimental Evidence from Malawi. Am. J. Agric. Econ.; 2023; 106, pp. 330-353. [DOI: https://dx.doi.org/10.1111/ajae.12376]
48. Angrist, J.D.; Imbens, G.W.; Rubin, D.B. Identification of Causal Effects Using Instrumental Variables. J. Am. Stat. Assoc.; 1996; 91, pp. 444-455. [DOI: https://dx.doi.org/10.1080/01621459.1996.10476902]
49. Angrist, J.D.; Pischke, J.-S. Mastering ’Metrics: The Path from Cause to Effect; 1st ed. Princeton University Press: Princeton, NJ, USA, 2014; ISBN 978-0-691-15284-4
50. Angrist, J.D.; Pischke, J.-S. Mostly Harmless Econometrics: An Empiricistís Companion; Princeton University Press: Princeton, NJ, USA, 2009.
51. Imbens, G.W.; Angrist, J.D. Identification and Estimation of Local Average Treatment Effects. Econometrica; 1994; 62, pp. 467-475. [DOI: https://dx.doi.org/10.2307/2951620]
52. National Statistical Office, Government of Nepal. National Population and Housing Census. 2021; Available online: https://censusnepal.cbs.gov.np/results/economic?province=5&district=52&municipality=9 (accessed on 7 December 2023).
53. Vu, H.T.; Tran, D.; Goto, D.; Kawata, K. Does Experience Sharing Affect Farmers’ pro-Environmental Behavior? A Randomized Controlled Trial in Vietnam. World Dev.; 2020; 136, 105062. [DOI: https://dx.doi.org/10.1016/j.worlddev.2020.105062]
54. Duflo, E.; Glennerster, R.; Kremer, M. Using Randomization in Development Economics Research: A Toolkit; Technical Working Paper Series; NBER: Cambridge, UK, 2006; 333.Available online: http://www.nber.org/papers/t0333 (accessed on 12 December 2023).
55. Mullally, C.; Boucher, S.; Carter, M. Encouraging Development: Randomized Encouragement Designs in Agriculture. Am. J. Agric. Econ.; 2013; 95, pp. 1352-1358. [DOI: https://dx.doi.org/10.1093/ajae/aat041]
56. Angrist, J.D.; Imbens, G.W. Two-Stage Least Squares Estimation of Average Causal Effects in Models With Variable Treatment Intensity. J. Am. Stat. Assoc.; 1995; 90, pp. 431-442. [DOI: https://dx.doi.org/10.1080/01621459.1995.10476535]
57. Baiocchi, M.; Cheng, J.; Small, D.S. Instrumental Variable Methods for Causal Inference. Stat. Med.; 2014; 33, pp. 2297-2340. [DOI: https://dx.doi.org/10.1002/sim.6128] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/24599889]
58. Bastardoz, N.; Matthews, M.J.; Sajons, G.B.; Ransom, T.; Kelemen, T.K.; Matthews, S.H. Instrumental Variables Estimation: Assumptions, Pitfalls, and Guidelines. Leadersh. Q.; 2023; 34, 101673. [DOI: https://dx.doi.org/10.1016/j.leaqua.2022.101673]
59. Aizer, A.; Doyle, J.J. Juvenile Incarceration, Human Capital, and Future Crime: Evidence from Randomly Assigned Judges. Q. J. Econ.; 2015; 130, pp. 759-804. [DOI: https://dx.doi.org/10.1093/qje/qjv003]
60. Angrist, J.D.; Krueger, A.B. Instrumental Variables and the Search for Identification: From Supply and Demand to Natural Experiments. J. Econ. Perspect.; 2001; 15, pp. 69-85. [DOI: https://dx.doi.org/10.1257/jep.15.4.69]
61. Sajons, G.B. Estimating the Causal Effect of Measured Endogenous Variables: A Tutorial on Experimentally Randomized Instrumental Variables. Leadersh. Q.; 2020; 31, 101348. [DOI: https://dx.doi.org/10.1016/j.leaqua.2019.101348]
62. Omotilewa, O.J.; Ricker-Gilbert, J.; Ainembabazi, J.H. Subsidies for Agricultural Technology Adoption: Evidence from a Randomized Experiment with Improved Grain Storage Bags in Uganda. Am. J. Agric. Econ.; 2019; 101, pp. 753-772. [DOI: https://dx.doi.org/10.1093/ajae/aay108] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/33281194]
© 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.
Details
; Keshav Lall Maharjan