Adoption of farm mechanization for clean air: Evidence from farm trials in Pakistan

Mishra, Ashok K.; Hazrana, Jaweriah; Yamano, Takashi; Sato, Noriko; Arif, Babur Wasim Working Paper Adoption of farm mechanization for clean air: Evidence from farm trials in Pakistan ADB Economics Working Paper Series, No. 758 Provided in Cooperation with: Asian Development Bank (ADB), Manila Suggested Citation: Mishra, Ashok K.; Hazrana, Jaweriah; Yamano, Takashi; Sato, Noriko; Arif, Babur Wasim (2024) : Adoption of farm mechanization for clean air: Evidence from farm trials in Pakistan, ADB Economics Working Paper Series, No. 758, Asian Development Bank (ADB), Manila, https://doi.org/10.22617/WPS240587-2 This Version is available at: https://hdl.handle.net/10419/310397 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/3.0/igo/ ASIAN DEVELOPMENT BANK ASIAN DEVELOPMENT BANK 6 ADB Avenue, Mandaluyong City 1550 Metro Manila, Philippines www.adb.org ADB ECONOMICS WORKING PAPER SERIES NO. 758 December 2024 Adoption of Farm Mechanization for Clean Air Evidence from Farm Trials in Pakistan Many rice farmers burn stubble and straw after harvest, which worsens air pollution. This paper examines the impact of training for farmers in Punjab, Pakistan, on the adoption of mechanized rice harvesters that leave short rice stubble, thereby reducing the need for crop burning. The results show that the training program increased farm performance among rice farmers who adopted the rice harvesters, highlighting the need for adoption of these harvesters to reduce open field burning and contribute to cleaner air. About the Asian Development Bank ADB is committed to achieving a prosperous, inclusive, resilient, and sustainable Asia and the Pacific, while sustaining its efforts to eradicate extreme poverty. Established in 1966, it is owned by 69 members —49 from the region. Its main instruments for helping its developing member countries are policy dialogue, loans, equity investments, guarantees, grants, and technical assistance. ADOPTION OF FARM MECHANIZATION FOR CLEAN AIR EVIDENCE FROM FARM TRIALS IN PAKISTAN Ashok K. Mishra, Jaweriah Hazrana, Takashi Yamano, Noriko Sato, and Babur Wasim Arif ASIAN DEVELOPMENT BANK The ADB Economics Working Paper Series presents research in progress to elicit comments and encourage debate on development issues in Asia and the Pacific. The views expressed are those of the authors and do not necessarily reflect the views and policies of ADB or its Board of Governors or the governments they represent. ADB Economics Working Paper Series Ashok K. Mishra, Jaweriah Hazrana, Takashi Yamano, Noriko Sato, and Babur Wasim Arif No. 758 | December 2024 Ashok K. Mishra ([email protected]) is Kemper and Ethel Marley Foundation Chair and Jaweriah Hazrana (jaw[email protected]om) is a post-doctoral fellow at Arizona State University. Takashi Yamano ([email protected]) is a principal economist at the Economic Research and Development Impact Department; Noriko Sato ([email protected]) is a senior natural resources specialist of the Sectors Group; and Babur Wasim Arif ([email protected]) is a consultant at Pakistan Resident Mission, Asian Development Bank. Adoption of Farm Mechanization for Clean Air: Evidence from Farm Trials in Pakistan Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) © 2024 Asian Development Bank 6 ADB Avenue, Mandaluyong City, 1550 Metro Manila, Philippines Tel +63 2 8632 4444; Fax +63 2 8636 2444 www.adb.org Some rights reserved. Published in 2024. ISSN 2313-6537 (print), 2313-6545 (PDF) Publication Stock No. WPS240587-2 DOI: http://dx.doi.org/10.22617/WPS240587-2 The views expressed in this publication are those of the authors and do not necessarily reflect the views and policies ofthe Asian Development Bank (ADB) or its Board of Governors or the governments they represent. ADB does not guarantee the accuracy of the data included in this publication and accepts no responsibility for any consequence of their use. The mention of specific companies or products of manufacturers does not imply that they are endorsed or recommended by ADB in preference to others of a similar nature that are not mentioned. By making any designation of or reference to a particular territory or geographic area inthis document, ADB does not intend to make any judgments as to the legal or other status of any territory or area. This publication is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) https://creativecommons.org/licenses/by/3.0/igo/. By using the content of this publication, you agree to be bound bytheterms of this license. For attribution, translations, adaptations, and permissions, please read the provisions andterms of use at https://www.adb.org/terms-use#openaccess. This CC license does not apply to non-ADB copyright materials in this publication. If the material is attributed toanother source, please contact the copyright owner or publisher of that source for permission to reproduce it. ADB cannot be held liable for any claims that arise as a result of your use of the material. Please contact [email protected] if you have questions or comments with respect to content, or if you wish toobtain copyright permission for your intended use that does not fall within these terms, or for permission to use theADB logo. Corrigenda to ADB publications may be found at http://www.adb.org/publications/corrigenda. Notes: In this publication, “$” refers to United States dollars and “PRs” refers to Pakistan rupees. ABSTRACT Many rice farmers burn stubble and straw after harvest, which worsens air pollution. This paper examines the impact of training on the adoption of advanced rice harvesting technologies that reduce the need for crop burning. It uses data from a cluster randomized controlled trial in Pakistan. The results reveal that the training program improved the farm performance of rice growers. The cost–benefit analysis shows that farmers who received the training and used improved mechanical rice harvesting generated higher profits, an average gain of PRs19,784 per acre. The results highlight the potential of targeted extension strategies to accelerate the adoption of productivity-enhancing and sustainable technologies among smallholder farmers. Keywords: rice farming randomized control trial, difference-in-difference method, revenues, harvest losses JEL codes: C31, O33, Q12, Q16 ______________________ The study was supported by ADB TA 9940-REG Mainstreaming Impact Evaluation, Methodologies, Approaches, and Capacities in Selected Developing Member Countries, Subproject 1. Corresponding author: Ashok K. Mishra ([email protected]). 1. Introduction Increasing agricultural productivity remains a key priority for policymakers and development organizations in many developing and emerging economies, where a significant proportion of the population lives in rural areas and relies on agricultural activities for their livelihood (Gollin and Udry 2021; Mcarthur and Mccord 2017). The situation is particularly stark in South Asia, where approximately 80% of the population lives in rural areas. The agriculture sector contributes a large share of the region’s gross domestic product (55%), provides 45% of employment opportunities, and employs more than two-thirds (65%) of the total labor force (World Bank 2017). Increasing productivity could potentially initiate a positive cycle of development, leading to improved livelihoods, increased food security, and equitable access to economic opportunities. Agricultural mechanization is broadly viewed as a crucial way to increase productivity, livelihood strategy, yields, and income. For this reason, governments and development partners in many developing and emerging economies allocate substantial resources to boost agricultural productivity. For instance, programs such as input subsidies, machinery financing schemes, extension services, and infrastructure development have been geared toward improving agricultural outcomes (Diao, Silver, and Takeshima 2016; Kienzle et al. 2013). Furthermore, mechanization is anticipated to catalyze structural transformation by diverting labor from agriculture and into more productive nonfarm sectors, ultimately contributing to overall economic growth (Gollin, Parente, and Rogerson 2002). However, inappropriate use of agricultural machinery can have unintended consequences. In Punjab, Pakistan, wheat is the main staple crop, and wheat combine harvesters have been promoted in the past. Because of their relatively high availability, 2 farmers use wheat combine harvesters to harvest rice, even though wheat harvesters cut the rice crop in the middle, leaving shredded rice straws and high stubbles in the fields. Unable to remove shredded rice straws and high stubbles, farmers tend to burn them. The practice of burning crop residue has significant environmental and health implications (Kumar, Kumar, and Joshi 2015; Kaushal and Prashar 2021) and contributes to air pollution and greenhouse gas (GHG) emissions (Bhattacharyya and Barman 2018).1 Proper rice harvesters, on the other hand, can cut the rice plants at the bottom, leaving only short stubble in the fields. In addition, long rice straws, properly cut at the bottom, can be sold at a higher price than shredded straws. The use of proper rice harvesters, therefore, is expected to reduce crop burning and help farmers increase their postharvest income. The objective of this study is to assess the impact of the ADB technical assistance program Enhancing Technology-Based Agriculture and Marketing (ETAM) in Rural Punjab on the farm performance indicators (outcome variables) of rice farmers. The ETAM program is a multifaceted intervention that aims to promote improved mechanization and enhance the livelihoods of smallholders. The ETAM program includes training on harvesting technologies and improved cultivation practices, postharvest processing methods, market linkages, and increased profitability along the rice value chain. An important component of the ETAM program is the promotion of a proper rice harvester, which can cut the rice crops at the lowest possible height. 1 In a more recent study, Bhattacharyya et al. (2021) estimated a net gain (both economic and environmental) of $664 per hectare from the production of bioethanol from straw, followed by the conversion of biochar ($183 per hectare) and conservation agriculture practices worth $131 per hectare. 3 To estimate the treatment effects of the ETAM program on farm performance indicators, we have conducted a standard difference-in-differences (D-I-D) analysis by comparing the change in farm performance indicators (outcome variables) between the treated and control samples before and after the implementation of the program. In addition, we have used a triple difference (DDD) strategy to examine the specific impact of adopting rice harvesters in the treated group. The DDD analysis has compared the change in farm performance indicators (outcome variables) for the adopters and nonadopters within the treated villages relative to the change in farm performance indicators profitability for the control sample. The study contributes to the literature on agricultural mechanization in several ways. First, by leveraging data from a randomized controlled trial, we can estimate the causal impacts of an intervention promoting mechanized rice harvesting and postharvest technologies. In contrast to many previous observational studies, our experimental design mitigates concerns about self-selection bias and endogeneity that often plague impact evaluations. Second, while most existing studies have primarily examined the effects of mechanization on productivity and/or profit, we discuss several potential pathways through which mechanization influences farm performance and profitability. The pathways include changes in crop yields, postharvest losses, output prices, and costs. Third, our study contributes to the limited evidence on the impact of agricultural mechanization2 in the context of Pakistan. This study provides valuable insights into the 2 Agricultural mechanization has the potential to boost productivity and increase profitability. The extent of the economic benefits often depends on contextual factors and the effective implementation of targeted interventions. 4 potential impacts, challenges, and scalability of policy initiatives in a low-income developing country setting like Pakistan. The remainder of the paper is organized as follows. Section 2 describes the background, while section 3 describes the conceptual framework of the study. Section 4 contains the survey and data as well as descriptive statistics. Section 5 outlines the empirical framework. Section 6 presents the results of the study. Section 7 reports on the benefit-cost analysis. Finally, section 8 presents a discussion and conclusions as well as policy implications. 2. Background Pakistan is home to 8.2 million farm families (Pakistan Bureau of Statistics 2010) who are responsible for meeting the basic food and nutrition needs of an estimated population of 212 million people. Approximately 65% of Pakistan’s total population lives in rural areas and is directly or indirectly dependent on agriculture for their livelihood. Agriculture is an important sector in Pakistan, employing over 37% of the labor force and contributing nearly 30% to gross domestic product (GDP) (Government of Pakistan 2023). Over the last few decades, the contribution of agriculture to GDP has declined in proportion to changes in the composition of the sector. However, the agricultural industry has a much greater potential to contribute to Pakistan’s GDP if it adopts innovative and modern technologies to replace traditional and outdated agronomic and farming techniques (USAID 2018). Mechanization of farming in Pakistan is still in its early stages, with limited adoption of modern machinery and technologies, especially among smallholders (Government of Pakistan 2023). The level of farm mechanization in Pakistan is estimated 11 drum. The tines effectively pick up and transport the paddy through the harvester’s system. The threshing drum, which is attached to the rear of the machine, separates the grains from the straw through a high-velocity threshing and cleaning process. The threshed grains are then collected in an integrated storage tank with a capacity of 0.8– 0.9 tons. Meanwhile, the intact straw is discharged from the harvester and laid back in the field in neat rows, making subsequent field operations easier. By demonstrating this advanced machinery, the program aimed to show smallholder farmers the efficiency, precision, and labor-saving potential of mechanical harvesting. The demonstrations highlighted the harvester’s ability to minimize crop losses, improve the timeliness of operations, and reduce the drudgery associated with traditional manual harvesting methods, ultimately contributing to higher productivity and better farm performance for adopters. 4.2. Sample Selection The surveys were conducted in three distinct phases in the agriculture sector of rural Punjab, covering the districts of Sheikhupura and Hafizabad. In the baseline survey, which was conducted in September–October 2020, 978 farmers were interviewed. The same farmers were interviewed in the midline and endline surveys in September– December 2021 and 2023. The study used a cluster randomized control trial (CRCT) design to prevent information and services from spilling over between farmers in the treatment and control groups, which is difficult to avoid when both groups live in the same village. To minimize such contamination, the study used the CRCT in which villages, 12 rather than individual farmers, were randomly assigned to treatment and control groups. A three-stage process was used to assign the villages. In the first stage, the tehsils were selected. One tehsil was deliberately selected in each target district, considering criteria such as homogeneity in terms of soil salinity, depth and quality of groundwater, and the availability of service providers, processors, and exporters. In the second stage, the qanungo halqas (QH) was selected. Each selected tehsil consists of four to seven QHs. In each selected tehsil, two QHs were selected based on the criterion of access to paved roads adequate for the movement of combine harvesters to the villages and transportation of produce to the location of traders, processors, and exporters. After listing all smallholder farmers through complete enumeration, one of the two selected QHs from each tehsil was randomly assigned to the treatment group. Finally, within each selected QH, a few villages in similar geographic proximity were chosen based on having access to a suitable road for transporting machinery and produce. Villages that were very large or had extreme socioeconomic diversity, conflicts, or influential people who could influence project implementation were excluded. A comprehensive enumeration exercise was carried out in each village to list all smallholder farmers producing the target crops. During the listing activity, information was collected on household land ownership under various tenancy contracts, land size, crops grown in different seasons, sources of livelihood of landless households and their involvement in farm labor, and contact details. In the villages within the treatment QH, all smallholder farmers were invited to participate in the demonstrations and use the services of the providers. Table 1 shows the tehsils, the QHs selected from each tehsil, and the 13 number of villages from each QH. A total of 41 villages were selected from the rice-wheat cropping zones. About 978 smallholder farmers randomly selected from the QHs participated in each survey. Table 2 provides information on the participants from the treatment and control QHs for the baseline survey. 4.3. Data Description During the surveys, comprehensive data were collected in eight thematic blocks to develop a detailed understanding of the respondents’ agricultural practices and livelihoods. This multifaceted approach was designed to provide a thorough analysis of the impact of the research interventions and to capture different aspects of the respondents’ agricultural activities and socioeconomic conditions. The survey collected general information about the respondents and key demographics such as gender, age, marital status, education level, and farming experience of the household head. In addition, this section asked about households’ primary sources of livelihood, housing conditions, and land ownership status, which provides information on respondents’ general livelihood strategies and standard of living. Information was also collected on land ownership, an important aspect of agricultural production. Respondents provided detailed information on the type of land tenure arrangements they had entered into, including ownership, rental, or other contractual agreements. Furthermore, the number of cultivated plots, the total cultivable area, soil characteristics, and irrigation sources were recorded, all of which influence agricultural productivity and farm decision-making. A specific part of the survey focused on cropped areas to gain a comprehensive understanding of respondents’ cropping patterns. Respondents provided information on 14 the specific crops grown in different seasons and the area allocated to each crop. These data were crucial for analyzing crop choices, diversification strategies, and the potential impact on income and food security. The survey also requested information on respondents’ agricultural practices, such as land preparation methods, use of inputs (e.g., seeds, fertilizers, pesticides), and postharvest handling techniques. This information was crucial for identifying potential areas for improvement, promoting sustainable practices, and assessing the impact of interventions on agricultural productivity and resource efficiency. Finally, a separate section looked at agricultural production and asked about the rice varieties grown, irrigation methods used, seed quantities utilized, seed procurement sources, production levels, planting techniques, agronomic practices used, and harvesting methods. The survey collected vital information from farmers on the marketable surplus and marketing channels for the rice produced. This section also included information on total production disaggregated by variety, quantities reserved for household consumption, quantities stored for future sale, and the various means farmers use to facilitate the sale of their rice produce. Furthermore, the survey included a section on access to agricultural extension services, sources of market information, credit providers, asset ownership, and poverty levels. 4.3.1. Outcome variables We use multiple outcome variables to rigorously evaluate the impact of the ETAM program on the farm performance of rice-producing households in Punjab province. The selection of performance indicators across the entire production cycle, from harvest to 15 postharvest processes, provides a comprehensive picture. Our first outcome is yield (maunds [Md] per acre),5 which is a direct measure of productivity. By making harvesting operations more efficient and reducing crop losses with manual methods, the use of rice harvesters can increase rice yield on cultivated plots. Second, we also quantify the losses and damage to the rice crop during harvest. Note that such measurements provide direct evidence of whether the technologies improve the efficiency of the harvesting process. Crop loss and damage rates during harvest capture the efficiency of the process itself. Manual harvesting often results in spillage, broken grains, and loss of output. Mechanized harvesting with specialized rice harvesters can minimize such losses, which benefits farmers. Third, we check if there is a price premium or incentive for paddy harvested mechanically. This measure sheds light on the potential revenue benefits for farmers resulting from the adoption of mechanical rice harvesters. A price premium for mechanically harvested paddy reflects the market demand for higher quality grain with less impurities (such as broken rice and rice husk). This premium represents a potential revenue benefit for farmers from adoption. Fourth, we also assess the value of extra straw by-products that are made available as fodder through mechanical harvesting. The by-products, such as baled rice straw, provide additional ancillary benefits, for example, livestock feed, animal bedding and fuel—rice pallets.6 The amount of straw by-products available for use as fodder or fuel also has monetary value. During mechanical harvesting, the pieces of straw are not damaged or shortened, which increases the usable straw output and thus offers an 5 1 maund = 37.32 kilograms. 6 Palletization of rice straw is a form of mass and energy densification (Ishii and Furuichi 2014). The pellets are easy to handle, transport, store, and utilize because of the increase in bulk density. 16 additional benefit. Finally, we estimate the cost savings from the adoption of the mechanical rice harvester in preparing the land for the next crop cycle. This indicator demonstrates a longer-term impact on time and budget management. Timely and efficient harvesting can facilitate the timely planting of the next crop and save land preparation costs. Indeed, such longer-term cost savings from the use of mechanical rice harvesters further contribute to farm profitability. In sum, the selected indicators include metrics that are directly linked to productivity and profitability throughout the production cycle— quantifying harvesting efficiency, cost savings, revenue gains, and ancillary benefits. The comprehensive coverage thus provides appropriate evidence for assessing the impact of the transition from traditional rice harvesting methods to mechanical rice harvesting technologies at the farm level. 4.4. Descriptive Statistics Table 3 presents the descriptive statistics for the pooled sample, the treated and the control groups. Of note is the first panel of Table 3, where the rice yield for all sampled farmers in the survey region was about 36 Md per acre. However, the yield of the treated group of farmers (37 Md per acre) is higher than that of the control group of farmers (35 Md per acre). The straw price received by farmers in the region also varied between the treatment and control groups. The price of straw for all sampled farmers in the survey region was about PRs7,589 per acre. However, the price of straw received is higher for the treated group of farmers (about PRs7,696 per acre) than for the control group of farmers (PRs7,384 per acre). The last row of the top panel in Table 3 shows the cost savings in preparing the land for the next sowing by using the mechanical rice harvester. In fact, these cost savings are much higher for the treated group of farmers, whose 17 average savings are PRs3,421 per acre compared to the control group of farmers (PRs2,855 per acre) and the farmers of the entire sample (PRs3,251 per acre). The second panel of the table shows the socioeconomic characteristics of the households in the sample. Farmers in the treated group are slightly younger (44 years old), have less farming experience (22 years), and derive a lower proportion of their income from farming (47%) than farmers in the control group. Figure 1 presents the kernel densities of the outcome variables for the control and treatment groups. The figure reveals that the distribution of rice yields for the treatment group is slightly shifted to the right compared to the control group, indicating higher average yields for the treated farmers. The distribution of harvest losses for the treatment group is skewed toward lower values compared to the control group. The pattern implies that the ETAM program was effective in reducing harvest losses for the treated farmers, possibly through improved harvesting techniques and technologies. Similarly, the distributions for the price of paddy and the additional price of straw are shifted to the right in the treatment group compared to the control group. The result suggests that the treated farmers were able to fetch higher prices for their paddy and earn additional income from straw sales. This is likely due to improved quality, better market linkages, or greater bargaining power promoted by the ETAM program. An important component of the ETAM program is the promotion and use of the Kubota rice harvester, a mechanized harvesting technology designed to address the labor-intensive and time-sensitive nature of rice harvesting operations. Figure 2 illustrates the proportion of farmers who used rice harvesters in both the control and treatment groups over the three waves of the survey. A notable upward trend in adoption can be 18 observed in both groups, with a particularly pronounced increase in the treatment group. In the baseline wave of the survey before the implementation of the ETAM program, the adoption rates of rice harvesters were relatively low at 7.95% in the control group and 19.11% in the treatment group. In the course of the subsequent survey waves, the proportion of farmers using rice harvesters increased in both groups. However, the difference in adoption rates between the control and treatment groups widened over time, from 11.16 percentage points (19.11–7.95) at the baseline survey to 18.25 (34.62–16.37) at the midline, and 22.65 percentage points (44.03–21.65) at the endline survey. The results indicate that the widening gap in the adoption of the rice harvesting machinery suggests that the ETAM program, with its targeted training and demonstrations, played a significant role in accelerating the adoption of rice harvesters by treated farmers compared to the control group. 5. Empirical Framework We begin by estimating a difference-in-differences (D-I-D) specification of the model. The D-I-D technique compares the outcomes in the treated and control groups before and after the technical assistance demonstrations. This allows us to measure the average effect of the demonstrations on outcomes. Formally, we estimate equation: FPit =α+βTP(Treatedit ∗Postt)+βXXit +δi+εit (1) where FPit represents the outcome variable for farm performance (yields, harvest losses, additional price of paddy, price of straw saved, and cost savings in land preparation7 for the next sowing) of household h and period t. Treatedit is a binary variable that takes the 7 All continuous variables in the regression are in logarithmic form. 19 value of 1 if a household is in the treatment group, 0 otherwise. Postt is a binary variable that takes the value of 0 if the year is before the ETAM program and the value of 1 after the implementation of the program. Xit is a vector of household characteristics, δi is the panel fixed effects, and 𝜀𝜀𝑖𝑖𝑖𝑖 is the error term. Note that the standard errors are clustered at the village level, which is consistent with the level of treatment assignment (Bertrand, Duflo, and Mullainathan 2004; Abadie et al. 2017). The coefficient of interest is βTP, which measures the D-I-D estimate of the average effect of ETAM on the outcome variables of farm performance; in other words, the change in outcomes between households that received the ETAM program intervention (i.e., treated) and those that did not (i.e., control) compared to the period before the intervention. However, the analysis assumes that without the treatment, both the treated group and the control group would have followed the same path with respect to their farm performance indicators (outcome variables). A positive coefficient 𝛽𝛽𝑇𝑇𝑇𝑇 indicates that the farm performance indicators of the farm households receiving treatment improved after the intervention compared to those not receiving the treatment. Although the D-I-D model is informative in obtaining the average treatment effects of the ETAM demonstrations, it does not allow us to measure the differential or additional impact of the Kubota rice harvester—an important component of the ETAM program. Therefore, to estimate the effect of the Kubota rice harvester on farm performance indicators, we modified our main specification and specified a triple difference design (DDD). The DDD approach allows us to compare the impact of the ETAM program on the treated group, which was differentiated by the adoption of the Kubota rice harvester (KRH) relative to the control group. 20 FPit =α+βTPK(Treatedit ∗Postt∗KRHt)+βTP(Treatedit ∗Postt)+βXXit +δi+εit (2) where KRHt is a dummy variable that indicates the adoption of the Kubota rice harvester in the posttreatment period; KRH takes the value of 1 if the household adopted the Kubota rice harvester in the posttreatment period; 0 otherwise. The remaining variables in equation (2) are defined as in the D-I-D model presented in equation 2. In the DDD specification, βTPK measures the effect of the ETAM program on the farm performance indicators (outcome variables) for households that adopted the Kubota rice harvester. The coefficient captures the variation in outcomes specific to adopter households in the treated group (relative to the control group) before and after the program rollout. In contrast, βTP measures the overall effect of the ETAM intervention on farm performance indicators (outcome variables) of households that were exposed to the treatment. Parallel trends assumption The D-I-D methodology relies on the assumption of “parallel trends,” according to which, without treatment, the treated group and the control group would have followed the same trend in the outcomes studied. This assumption is crucial because it helps to ensure that the observed effects after treatment are due to the treatment itself and not to preexisting trends. We test this parallel trend assumption by providing some suggestive evidence. First, we rely on baseline data collected in 2020 before the implementation of the ETAM program to show that household farm performance indicators (or outcome variables used in this study) followed similar trends in the treated and control groups. However, the present study has data limitations that prevent the establishment of parallel pretreatment trends. In particular, households were observed only once before treatment. 27 for paddy, income from straw sales, higher yields, and cost savings in land preparation. In this section, we present a comprehensive benefit–cost analysis of the ETAM program using average values from the survey data. Table 8 shows the benefit–cost analysis of the ETAM program across the three survey rounds for the treatment group that also adopted the mechanized rice harvester (KRH). While the initial cost of adopting a KRH is higher than that of a traditional combined harvester, this additional investment is offset by the numerous benefits derived from adopting the KRH and implementing the ETAM program’s training and practices. One of the significant advantages of the ETAM program is the reduction of harvest losses. Harvest losses are a major source of inefficiency in traditional rice farming, as a considerable portion of the crop is lost during harvesting and postharvest handling. The ETAM program helped farmers minimize these losses by training them in proper harvesting techniques and promoting the KRH technology. In all survey rounds, the reduction in harvest losses due to the adoption of KRH ranged from 3.5 to 3.8 Md per acre. This translates into additional revenue attributed to the reduction in harvest losses, which ranges from PRs6,243 to PRs14,062 per acre and contributes directly to farm income. In addition, the ETAM program enabled farmers to fetch a higher price for their paddy by improving the quality of the rice and facilitating better market linkages. Farmers who participated in the ETAM program with the KRH can earn additional revenue due to a higher price premium ranging from PRs75 to PRs138 per acre, depending on the survey round. Complementing the advantages mentioned above, participation in the ETAM program and the adoption of KRH generated additional income to farmers through the sale of straw. Traditionally, a significant amount of straw is lost or wasted during 28 harvesting and postharvest operations. However, by adopting the KRH and implementing the ETAM program’s practices, farmers were able to efficiently bundle the straw using the KRH and sell a larger portion of the straw. In this way, farming families were able to earn an additional income ranging from PRs5,273 to PRs11,441 per acre during the survey rounds. This additional income from straw sales can be an important source of revenue, especially in areas where straw is in high demand. Moreover, the adoption of advanced technologies and mechanization through the ETAM program has reduced the cost of land preparation for the next sowing cycle. By implementing efficient land preparation techniques, improved farm machinery, and better resource management practices promoted by the ETAM program and the adoption of KRH, farmers saved costs ranging from PRs2,196 to PRs5,690 per acre during the survey rounds. This cost-saving aspect further contributes to the overall economic viability of the ETAM program and the adoption of the KRH. The last row of Table 8 indicates that the net benefit per acre from participation in ETAM program practices and KRH adoption ranges from PRs10,584 in the baseline survey to PRs25,079 in the endline survey. The significant net benefits highlight the economic viability and potential for scaling up ETAM interventions to improve farm performance, reduce postharvest losses, and promote sustainable agricultural development in the region. 8. Conclusion and Implications Pakistan’s agriculture sector faces numerous challenges, such as low productivity, postharvest losses, and limited mechanization, especially among smallholder farmers. Despite the potential of agricultural mechanization to boost productivity, the adoption of 29 harvesting technologies among smallholder farmers in developing countries remains low. The Government of Pakistan and national research agencies have been working with the Asian Development Bank to educate and provide extension services and new technologies to increase rice productivity and farm performance in the rice value chain. This study examined the impact of the ETAM program on the farm performance of rice farmers in the rural Punjab province of Pakistan. The results revealed that the ETAM program had a positive impact on a variety of farm performance outcomes, including higher rice yields, reduced harvest losses, additional income from the sale of paddy and straw, and cost savings in land preparation. The increased adoption of a proper rice harvester is expected to reduce crop burning in open fields. The positive impact of the ETAM program aligns with the government’s efforts to increase agricultural productivity, achieve food self-sufficiency, and reduce air pollution. In addition, the reduction in harvest losses observed in this study is particularly noteworthy, as postharvest losses are one of the main causes of inefficiency in Pakistan’s agriculture sector. The ETAM program focused on promoting improved harvesting techniques and mechanized technologies such as the Kubota rice harvester. In addition to yield gains and reduced harvest losses, the ETAM program also enabled higher prices for paddy and additional income from the sale of straw. The long rice straws left by the Kubota harvester are more valuable than shredded rice straws left by wheat combined harvesters. Therefore, the findings of this study highlight the role of the ETAM program in improving market linkages, bargaining power, mechanization, and the overall value chain for rice cultivation in Pakistan. 30 Overall, the heterogeneity analyses underscore the importance of considering farm size and access to information when designing and implementing agricultural development programs. Tailored strategies, support mechanisms, and complementary policies may be required to address the varying constraints and needs of different subgroups and ensure equitable and inclusive benefits from such initiatives. It is noteworthy that the ETAM program had a positive and significant impact on farm performance indicators in the small and large holding farm categories, highlighting the potential for inclusive benefits from the interventions. Nonetheless, the results underscore the need for tailored support mechanisms and complementary policies to address the specific constraints and challenges faced by smallholder farmers in adopting and maximizing the benefits of mechanized technologies. Similarly, our findings highlight the critical role of access to information and extension services in facilitating the effective adoption and utilization of new technologies and practices. Households with better access to information are more likely to be aware of the potential benefits, optimal timing, proper implementation methods, and farm performance. The finding underscores the need to strengthen agricultural extension services and improve information dissemination mechanisms in Pakistan, especially for marginalized and resource-constrained farming communities. The findings of this study have several policy implications for promoting sustainable agricultural development and improving farm performance in Pakistan. First, policies such as the ETAM program, which promotes mechanization, improved cultivation practices, and market linkages, need to be scaled up and incorporated in the mainstream. These measures contribute significantly to achieve the country’s agricultural development 31 goals by boosting productivity, reducing postharvest losses, and increasing farmers’ incomes. Second, targeted support mechanisms such as subsidies, credit facilities, and capacity-building programs could be designed to address the specific constraints and needs of smallholder farmers. The heterogeneity analysis revealed that while the ETAM program had a positive impact on all farm sizes, the benefits were more pronounced for larger farms. The results of this study underscore the importance of ensuring that mechanization initiatives benefit all by providing tailored support to smallholder farmers who often face limitations in accessing credit, extension services, and economies of scale. Third, strengthening agricultural extension services and improving information dissemination mechanisms are crucial to facilitate the effective adoption and utilization of new technologies and practices. Targeted efforts to improve the reach and quality of extension services, coupled with effective communication strategies, can empower farming communities, particularly in resource-constrained areas, to maximize the potential benefits of such initiatives. Finally, complementary policies and investments in infrastructure, such as rural roads and storage facilities, are essential to support the adoption and scalability of mechanized technologies. Improved infrastructure can reduce postharvest losses, enhance market access for farmers, and facilitate the efficient transportation and handling of agricultural produce, thus contributing to the overall economic viability and sustainability of mechanization initiatives. 32 Figure 1: Kernel Density Estimates of Outcomes by Treatment Groups of Sampled Households, Pakistan Md = maund. 1 maund = 37.32 kilogram. Source: Authors’ calculation. 33 Figure 2: Percentage of Farmers Adopting the Kubota Rice-Harvester Before and After the Technical Assistance Demonstrations Source: ADB. Pakistan: Enhancing Technology-Based Agriculture and Marketing in Rural Punjab. https://www.adb.org/projects/52232-001/main. 34 Figure 3: Diagnostics for Parallel Trends Source: Authors’ calculations. 35 Table 1: Zone, Selected Tehsil, Selected Qanungo Halqas, and Number of Selected Villages for the Surveys Tehsil Qanungo Halqa No. of Selected Villages Muridke Ahdian 8 Muridke 11 Hafizabad Chak Chattha 9 Hafizabad 13 Source: ADB. Pakistan: Enhancing Technology-Based Agriculture and Marketing in Rural Punjab. https://www.adb.org/projects/52232-001/main. Table 2: Control and Treatment Samples in the Baseline, Midline, and Endline Surveys in Rural Punjab, Pakistan Group Qanungo Halqa Number of Participants Percentage of Participants Baseline Midline Endline Baseline Midline Endline Control Ahdian 169 151 137 90.01 95.45 90.31 Chak Chattha 176 165 153 91.01 96.85 91.64 Subtotal 345 316 290 90.49 96.12 90.95 Treatment Hafizabad 317 293 250 88.07 96.75 87.28 Muridke 316 308 273 90.05 98.54 90.67 Subtotal 633 601 523 89.13 97.70 89.08 Total 978 917 813 89.60 97.15 89.74 Source: ADB. Pakistan: Enhancing Technology-Based Agriculture and Marketing in Rural Punjab. https://www.adb.org/projects/52232-001/main. 36 Table 3: Summary Statistics of Farm, Household, and Plot Attributes of Sampled Households in Punjab, Pakistan Total Sample Treated Control Mean SD Mean SD Mean SD Yields (Md/acre)a 36.62 13.04 37.25 13.13 35.49 12.79 Harvest losses (Md/Acre) 3.55 1.29 3.539 1.25 3.58 1.41 Paddy price (PRs/Md) 97.88 76.59 97.363 78.50 99.05 72.18 Price of straw (PRs/acre) 7,589.04 4,150.24 7,695.85 4,264.06 7,384.00 3,916.36 Cost savings in land preparation for next season (PRs/acre) 3,251.31 2,202.74 3,421.02 2,253.99 2,855.08 2,025.62 Household Controls Household size (number) 8.361 4.26 8.44 4.26 8.32 4.26 Income from crops (%) 58.22 325.57 78.13 548.65 47.49 22.70 Farming experience (years) 22.44 13.29 23.34 13.31 21.953 13.26 Literacy of household head (%) 75.64 42.93 77.06 42.05 73.00 44.41 Age of household head (years) 44.74 14.56 45.18 14.33 44.51 14.68 Marital status (%) 88.06 32.43 89.20 31.05 87.44 33.14 Plot Controls Soil type (%) Clay loam 35.10 47.73 34.37 47.51 35.49 47.85 Clay 11.19 31.53 8.44 27.81 12.68 33.28 Loam 22.66 41.86 24.26 42.88 21.79 41.29 Others 0.35 5.88 0.15 3.89 0.45 6.70 Sandy 0.27 5.16 0.23 4.77 0.29 5.35 Sandy loam 5.52 22.83 4.94 21.69 5.83 23.43 Silty loam 24.92 43.26 27.60 44.72 23.47 42.39 Irrigation source (%) Canal 24.92 43.26 27.61 44.72 23.47 42.39 Tubewell 52.98 49.92 70.95 45.42 43.29 49.56 Canal + Tubewell 1.57 12.44 0.99 9.89 1.89 13.61 Observations 3,233 2,083 1,150 Md = maund, PRs = Pakistan rupees, SD = standard deviation. a1 maund = 37.32 kilograms. Source: ADB. Pakistan: Enhancing Technology-Based Agriculture and Marketing in Rural Punjab. https://www.adb.org/projects/52232-001/main. 43 Box: Rice Harvester Type by Sampled Households, Pakistan The Kubota rice harvester used by Centre for Agriculture and Biosciences International (CABI) as part of the technical assistance was responsive to all field conditions where the crop at maturity is either erect or lodged. The ER-112 was the latest model of paddy-specific half-feed harvester in Punjab having six straw walkers with a cutter width of 7 feet and 112-HP common rail engine. It essentially consists of four major parts: the header, the conveying chains, the threshing drum, and the grain storage tank. It cuts the paddy at the lowest possible height and the chains are installed on the header assembly with the tines mounted on the chains. The tines are used to grab the paddy and are guided through chains to the threshing drum. The grains are threshed and cleaned using high-velocity blowers and an integrated threshing drum mounted on the rear side of the machine. The threshed grains are stored in a storage tank with a capacity of 0.8– 0.9 tons. The intact/whole straw is moved out of the harvester and thrown back onto the field in rows. The Thinker XG750S is a full-feed paddy-specific harvester that performs three separate operations: reaping, threshing, and winnowing instead of separate machines. The wider header (7 feet) at the front of the harvester gathers the crop into the machine. The reel pushes the crop toward the cutter blade, which cuts the crop. The grains are then transported to the threshing drum via a conveyer belt. There, the grains are separated from the stalks and collected in a separate drum with a storage capacity of 0.3–0.4 tons, which is discharged through the machine’s side pipe auger. The harvester is capable of harvesting both standing and lodged crops. However, farmers only interested in harvesting the lodged crop with this harvester. Source: ADB. Pakistan: Enhancing Technology-Based Agriculture and Marketing in Rural Punjab. https://www.adb.org/projects/52232-001/main. 44 Figure A1: Histogram of Farm Size and Sampled Households, Pakistan Source: ADB. Pakistan: Enhancing Technology-Based Agriculture and Marketing in Rural Punjab. https://www.adb.org/projects/52232-001/main. 45 Figure A2: Percentage of Sampled Households Accessing Information, Pakistan Source: Source: Authors’ calculation. 46 REFERENCES Abadie, A. S. Athey, G. W. Imbens, and J. Wooldridge. 2017. When Should You Adjust Standard Errors for Clustering? National Bureau of Economic Research Working Paper Series. No. 24003. http://www.nber.org/papers/w24003.pdf. Adhvaryu, A. 2014. Learning, Misallocation, and Technology Adoption: Evidence from New Malaria Therapy in Tanzania. Review of Economic Studies. 81 (4). pp. 1331– 1365. https://doi.org/10.1093/restud/rdu020. Ahmed, A. U., K. Ahmad, V. Chou, R. Hernandez, P. Menon, F. Naeem, F. Naher, W. Quabili, E. Sraboni, and B. Yu. 2013. The Status of Food Security in the Feed the Future Zone and Other Regions of Bangladesh: Results from the 2011-2012 Bangladesh Integrated Household Survey. International Food Policy Research Institute. . https://ebrary.ifpri.org/digital/collection/p15738coll2/id/127518. Amir, S., Z. Saqib, M. I. Khan, M. A. Khan, S. A. Bokhari, M. Zaman-ul-Haq, and A. Majid. 2020. Farmers’ Perceptions and Adaptation Practices to Climate Change in Rain-Fed Area: A Case Study from District Chakwal, Pakistan. Pakistan Journal of Agricultural Sciences 57. pp 465–475. doi: 10.21162/PAKJAS/19.9030 Azhar, R., M. Zeeshan, and K. Fatima. 2019. Crop Residue Open Field Burning in Pakistan: Multi-Year High Spatial Resolution Emission Inventory for 2000–2014. Atmospheric Environment. 208. pp. 20–33. https://doi.org/10.1016/j.atmosenv.2019.03.031. Bandiera, O. and I. Rasul. 2006. Social Networks and Technology Adoption in Many in the Developing World. Economic Journal. 116 (1957). pp. 869–902. Beaman, L., A. BenYishay, J. Magruder, and A. M. Mobarak. 2021. Can Network Theory-Based Targeting Increase Technology Adoption? American Economic Review. 111 (6). pp. 1918–1943. https://doi.org/10.1257/AER.20200295. Bertrand, M., E. Duflo, and S. Mullainathan. 2004. How Much Should We Trust Differences-In-Differences Estimates? The Quarterly Journal of Economics. 119 (1). pp. 249–275. Bhattacharyya, P. et al. 2021. Turn the Wheel from Waste to Wealth: Economic and Environmental Gain of Sustainable Rice Straw Management Practices over Field Burning in Reference to India. Science of The Total Environment. 775. 145896. Bhattacharyya, P. and D. Barman. 2018. Chapter 21 – Crop Residue Management and Greenhouse Gases Emissions in Tropical Rice Lands. Soil Management and Climate Change. pp. 323–335. https://doi.org/10.1016/B978-0-12-812128-3.000215. Conley, T. G. and C. R. Udry. 2010. Learning About a New Technology: Pineapple in Ghana. The American Economic Review. 100 (1). pp. 35–69. 47 Das, B., P. V. Bhave, S. P. Puppala, K. Shakya, B. Maharjan, R. M. Byanju. 2020. A Model-Ready Emission Inventory for Crop Residue Open Burning in the Context of Nepal. Environmental Pollution. 266 (3). 115069. https://doi.org/10.1016/j.envpol.2020.115069. de Janvry, A., K. Emerick, E. Sadoulet, and M. Dar. 2015. The Agricultural Technology Adoption: What Can We Learn from Field Experiments? Revue d’Economie Du Developpement. 23 (4). pp. 129–153. https://doi.org/10.3917/edd.294.0129. Diao, X., J. Silver, and H. Takeshima. 2016. Agricultural Mechanization and Agricultural Transformation Xinshen. IFPRI Discussion Paper 01527. April. Gollin, D., S. Parente, and R. Rogerson. 2002. The Role of Agriculture in Development. American Economic Review. 92 (2). pp. 160–164. https://doi.org/10.1093/0195305191.003.0008. Gollin, D. and C. Udry. 2021. Heterogeneity, Measurement Error, and Misallocation: Evidence from African Agriculture. Journal of Political Economy. 129 (1). pp. 1–80. https://doi.org/10.1086/711369. Government of Pakistan. 2020. Mechanization of Agriculture Sector of Pakistan. Engineering Development Board. Government of Pakistan. 2023. Pakistan Economic Survey 2022–23. Finance and Economic Affairs Division, Ministry of Finance. Gunasekera, D., H. Parsons, and M. Smith. 2017. Post-harvest loss reduction in AsiaPacific developing economie. Journal of Agribusiness in Developing and Emerging Economies, 7(3). pp. 303–317. https://doi.org/10.1108/JADEE-12-2015-0058. Gustavsson J., C. Cederberg, U. Sonesson, R. van Otterdijk, and A. Meybeck. 2011. Global Food Losses and Food Waste. Food and Agriculture Organization of the United Nations. Ishii, K. and T. Furuichi. 2014. Influence of Moisture Content, Particle Size and Forming Temperature on Productivity and Quality of Rice Straw Pellets. Waste Management. 34 (12). pp. 2621–2626. https://doi.org/10.1016/j.wasman.2014.08.008. Kaushal, L. A. and A. Prashar. 2021. Agricultural Crop Residue Burning and Its Environmental Impacts and Potential Causes – Case of Northwest India. Journal of Environmental Planning and Management. 64 (3). pp. 464–484. https://doi.org/10.1080/09640568.2020.1767044. Kienzle, J., Ashburner, J.E., Sims, B.G. 2013. Mechanization for Rural Development: a Review of Patterns and Progress from around the World. Food and Agriculture Organization, Rome 48 Kumar, D. and P. Kalita. 2017. Reducing Post-Harvest Losses During Storage of Grain Crops to Strengthen Food Security in Developing Countries. Foods. 6 (1). pp. 1–22. https://doi.org/10.3390/foods6010008. Kumar, P., S. Kumar, and L. Joshi. 2015. Socioeconomic and Environmental Implications of Agricultural Residue Burning: A Case Study of Punjab, India. Springer. Lin, M. and T. Begho. 2022. Crop Residue Burning in South Asia: A Review of the Scale, Effect, and Solutions with a Focus on Reducing Reactive Nitrogen Losses. Journal of Environmental Management. 314. 115104. https://doi.org/10.1016/j.jenvman.2022.115104. Mcarthur, J. W. and G. C. Mccord. 2017. Fertilizing Growth: Agricultural Inputs and Their Effects in Economic Development. Journal of Development Economics. 127 (February). pp. 133–152. https://doi.org/10.1016/j.jdeveco.2017.02.007. Mottaleb, K.A., Krupnik, T.J. and Erenstein, O. 2016. Factors Associated with SmallScale Agricultural Machinery Adoption in Bangladesh: Census Findings. Journal of Rural Studies, 46, 155-168. https://doi.org/10.1016/j.jrurstud.2016.06.012 Mumin, Y., A. Abdulai, and R. Goetz. 2023. The Role of Social Networks in the Adoption of Competing New Technologies in Ghana. Journal of Agricultural Economics. 74 (2). pp. 510–533. https://doi.org/10.1111/1477-9552.12517. Nath, B. C., Y.-S. Nam, Md. D. Huda, Md. M. Rahman, P. Ali, and S. Paul. 2017. Status and Constrain for Mechanization of Rice Harvesting System in Bangladesh. Agricultural Sciences. 8. pp. 492–506. https://doi.org/10.4236/as.2017.86037. Pakistan Bureau of Statistics. 2010. Agriculture Census, 2010. Government of Pakistan Statistics Division Agricultural Census Organization. Pakistan Bureau of Statistics. 2022. Pakistan Statistcal Year Book (2022). Pakistan Bureau of Statistics, Government of Pakistan. State Bank of Pakistan. 2020. Annual Report 2020-2021 (State of the Economy). https://www.sbp.org.pk/reports/annual/arFY21/Anul-index-eng-21.htm. Takeshima, H., H. O. Edeh, A. O. Lawal, M. A. Isiaka. 2015. Characteristics of PrivateSector Tractor Service Provisions: Insights from Nigeria. Developing Economies. 53 (3). pp. 188–217. United States Agency of International Development (USAID). 2018. Pakistan Agricultural Technology Transfer Activity. Project Stakeholders’ Analysis Report (Primary Research Findings and Field Activity Status) USAID-391-C-17-00004. World Bank Group. 2017. World Bank Annual Report 2017. http://documents.worldbank.org/curated/en/143021506909711004/World-BankAnnual-Report-2017. ASIAN DEVELOPMENT BANK ASIAN DEVELOPMENT BANK 6 ADB Avenue, Mandaluyong City 1550 Metro Manila, Philippines www.adb.org ADB ECONOMICS WORKING PAPER SERIES NO. 758 December 2024 Adoption of Farm Mechanization for Clean Air Evidence from Farm Trials in Pakistan Many rice farmers burn stubble and straw after harvest, which worsens air pollution. This paper examines the impact of training for farmers in Punjab, Pakistan, on the adoption of mechanized rice harvesters that leave short rice stubble, thereby reducing the need for crop burning. The results show that the training program increased farm performance among rice farmers who adopted the rice harvesters, highlighting the need for adoption of these harvesters to reduce open field burning and contribute to cleaner air. About the Asian Development Bank ADB is committed to achieving a prosperous, inclusive, resilient, and sustainable Asia and the Pacific, while sustaining its efforts to eradicate extreme poverty. Established in 1966, it is owned by 69 members —49 from the region. Its main instruments for helping its developing member countries are policy dialogue, loans, equity investments, guarantees, grants, and technical assistance. ADOPTION OF FARM MECHANIZATION FOR CLEAN AIR EVIDENCE FROM FARM TRIALS IN PAKISTAN Ashok K. Mishra, Jaweriah Hazrana, Takashi Yamano, Noriko Sato, and Babur Wasim Arif