Incentivizing alternatives to agricultural waste burning in Northern India: trust, awareness, and access as barriers to adoption

The burning of agricultural residue from previous season’s rice crop, primarily in the states of Punjab and Haryana, is a key contributor to poor air quality during the winter across North India. Air quality can deteriorate to catastrophic levels during the Agricultural Waste Burning (AWB) season in October–November, when fine particulate matter (PM2.5) concentrations can exceed WHO daily maxima over a sustained period by an order of magnitude or more, over a large swathe of the Indo-Gangetic plain. Over the past decade, attempts by Indian governments to change farmer behavior by incentivizing the use of novel technologies for managing rice residue without burning it have been met with little success. This paper uses farmer and expert interviews, as well as secondary data, to examine the barriers to adoption of these technologies in the state of Punjab. We analyze how operational factors (such as farm size, timing, technology availability, and choice) affect a farmers’ decision to choose (or not) a rice residue management practice. We develop a financial model for analyzing the costs of residue management technologies that are consistent with the decision-making process of both small and large farmers. We find that more sustainable residue management practices can be cost effective relative to residue burn, especially when existing subsidies are applied. However, difficulties in accessing technological alternatives to AWB and subsidies for their use and a lack of trust in the government’s ability to deliver the full benefits of subsidies, all contribute the low adoption of technological alternatives to AWB.


Introduction
Exposure to air pollution is the single largest contributing factor to global mortality, responsible for 15% of all premature deaths and 275 million Disability-Adjusted life years in 2017 (Global Alliance on Health and Pollution 2019). Agricultural Waste Burning (AWB) is a major contributor of air pollution in many countries, particularly in South Asia (Gadde et al. 2009). Exposure to fine particulate matter less than 2.5 μm (micrometer), also known as PM2.5, released by AWB can cause premature mortality due to respiratory and cardiovascular diseases (Brauer et al. 2012;Brook et al. 2010;Makkonen et al. 2010). AWB is thus a major public health concern in India and the second biggest cause of exposure to PM2.5, estimated to be responsible for over 35,000 deaths (Maji 2019). Indoor burning of solid fuels and kerosene lighting is a leading source of PM2.5 exposure (Maji 2019).
There is considerable spatial variation in AWB-related PM2.5 concentration across India. The roughly hundreds of millions living in the Indo-Gangetic Plain (IGP) are the worst affected. Rice is harvested at the onset of winter in Northwestern Indian states of Punjab and Haryana and agricultural waste, in the form of straw and stubble left over on the fields after harvest, is burnt to clear fields rapidly and inexpensively (Jain et al. 2014). Each year, during the November/December AWB season, air pollution reaches catastrophic levels throughout the IGP. Modeling shows that AWB alone results in increases in PM2.5 concentration ranging from 50 to 150 μg/m 3 ; this is two to four times the PM2.5 concentration in other parts of India (Brauer et al. 2012;Maji 2019;Venkataraman et al. 2018), and public health emergencies from AWB routinely take place. For example, during October 2016, northern India experienced a public health emergency when it recorded two-week average concentrations of PM2.5 of 440 ± 265 μg/m 3 , which are 20-37 times above WHO's guidelines (Jethva et al. 2019).
The Northwestern state of Punjab is widely considered as India's granary, yet its agricultural sector faces a plethora of environmental challenges including an alarming reduction of groundwater tables (Rodell et al. 2009) in addition to AWB (Awasthi et al. 2011;Gupta et al. 2004;Sarkar 2011;Singh 2012). These concerns are intertwined. In particular, policies aimed at containing the reduction in groundwater has aggravated AWB and its effect on the ambient air quality of the Indo-Gangetic Plain (Singh et al. 2019); AWB is thus part of the food-water-energy nexus in the region (Bhuvaneshwari et al. 2019).
This paper centers on why farmers in Punjab continue to burn post-harvest rice stubble, despite a ban on AWB in the region. We use interviews with farmers and experts to quantify the tangible and intangible costs of alternatives to AWB and propose some approaches to alleviating the present crisis. The remainder of the paper is as follows. In Sect. 2 we present a background on the drivers of AWB and highlight the policy context within which the AWB crisis has emerged in India. In particular, we clarify how it is linked to groundwater depletion, another major environmental externality of rice farming in the Punjab. We further present technological solutions to reduce AWB. In Sect. 3, we present the mixed methods approach used for data collection and analysis. In Sect. 4, we present our results and discuss their implications on decision-making and policies. First, we analyze how operational factors (such as farm size, timing, and technology availability and choice) affect a farmers' decision to choose a rice residue management practice and develop a decision flowchart to describe this process. Second, we use evidence from interview data, and data from published literature to develop a financial model for analyzing the costs of residue management consistent with the decision-making process of farmers operating at different scales (e.g., farm size), and with differential access to residue management options, and government policies. Finally, we go on to address, based on our interview data, why farmers view government support to reduce agricultural waste burning as inadequate and examine some plausible alternatives. In Sect. 5, we provide conclusions with some final observations.

Rice production in Punjab and environmental externalities
Agricultural Waste Burning (AWB) in Punjab is a result of a confluence of policies going back to the introduction of Green Revolution technologies in India (Frankel 2015). For decades, farmers in Punjab have been incentivized to grow rice as part of India's food security policies. Guaranteed procurement of rice in Punjab at a Minimum Support Price (MSP) by Food Corporation of India (FCI), budgeted by the national government, reduced the risks of rice cultivation, and supported the expansion of rice production in Punjab (Niti Aayog 2016). Punjab occupies only 1.5% of India's geographical land area, yet it contributed 24.2% in 2014-15 to the government procurement of rice used for subsidized distribution to the poor (Grover et al. 2016). Figure 1 shows increase in average area and yield of rice consumption in Punjab between 1960-61 and 2017-18. This growth in both acreage and yields has co-evolved with increased mechanization of agriculture, particularly diesel pumps and harvesting technology, resulting in several environmental externalities. Combine Harvesters used on the spring rice crop leave behind stubble and straw on the field (Sidhu et al. 2007). Farmers burn the rice residue while preparing the field for wheat sowing season that follows. Further, migrant labor from neighboring states of Bihar and Uttar Pradesh has decreased (Sirhindi 2019) and resulted in an increase in the cost of agricultural labor during the rice harvest season (Sharma 2018). Farmers find it both convenient and economical to burn residue to clear and prepare their fields for the next crop. In 2018, over 20 million metric tons of rice residue was generated in Punjab, of which almost 10 million metric tons was subject to open-air burning (Ministry of Agriculture and Farmers Welfare & Department of Agriculture, Cooperation & Farmers Welfare 2019).
Along with AWB, the rapid depletion of ground water in Punjab is also a looming agricultural crisis. Populist policies that provide free electricity for agricultural pumps, have over the past two decades, led to an increase in the number of groundwater "tube wells" (Singh 2012), with the result that groundwater is being drawn for flooded rice cultivation at a pace that is far greater than the recharge rate of local aquifers (Shah 2009); this has led to a dramatic decline in ground water levels in the region. In an attempt to reduce the rates of groundwater decline, the Punjab state government enacted The Punjab Preservation of Subsoil Water Act in 2009 restricting paddy transplantation that was typically done before the onset of monsoons in May, to later in the season, i.e., on or after June 20th (Singh et al. 2019; The Punjab Preservation of Subsoil Water Act 2009). This policy was intended to increase the overlap between the growing season and the annual monsoon rains. However, the subsequent delay in the harvest of rice crops in autumn leaves rice farmers in the region with a short turnaround of about two weeks to harvest the crop, manage the rice residue, and sow the winter wheat crop. The move to a later harvest has also meant that the peak of AWB has shifted from mid-October to early November (Jethva et al. 2019;Liu et al. 2020). A shift to crop burning later in the season is coterminous with changing meteorological conditions over the Indo-Gangetic plain where reduced wind speed and temperature inversions result in greater pollutant loading. Thus, a shift in growing season, along with an overall increase in paddy cultivation, coincides with 40% increase in total burning and ~ 50% increase in regional aerosol loading (Liu et al. 2020).

Technological solutions to AWB
The Indian government has attempted to change farmer behavior and to incentivize other means of managing rice residue. In 2014, the Government of India formulated The National Policy for Management of Crop Residue (NPMCR), which includes a list of directives to reduce AWB, such as the monitoring of agricultural fires, and facilitating the collection and transportation of rice residue. The National Green Tribunal (NGT), 2015 also provided several directives to curb air pollution due to AWB. Under the directives of National Green Tribunal (2015), open burning of residue is banned, and fines have been imposed (Singh and Zaffar 2017).
Technological alternatives can be classified into two major forms: (i) in situ and (ii) ex situ management of residue. In situ residue management consists of incorporating paddy straw into the post-harvest soil using different types of machinery. Ex situ residue management consists of collecting rice residue and transporting it away from the field to be utilized for applications, such as biomaterials, energy production, and industrial products. Details of technology options are provided in the SI.
In situ residue management Among the many devices that can be used in situ residue management of rice residue (See Table SI 1), the Happy Seeder (Somanathan and Gupta 2017) is the most widely studied. A combination of a seed drill and stubble mulcher, Happy Seeders can help farmers simultaneously sow wheat crop while mulching the stubble (Sidhu et al. 2007). Uptake of Happy Seeders has been limited; farmers have expressed problems of greater rodent attack, poor mulch quality resulting from insufficient decomposition of rice straw, and poor access as major concerns with the Happy Seeder (Singh et al. 2019). Other machinery such as the Rotavator, Disc Harrow, and Cultivator can be used for wet and dry mixing of rice residue into the soil after harvest (Singh et al. 2019). As discussed below, each of these appliances come with varying constraints of availability, access, and costs.
Ex situ residue management Residues of rice, wheat, maize and sugarcane are rich in lignocellulose biomass that contains cellulose, hemicellulose, and lignin that in principle can be used as raw material for the production of biofuels. Ash from rice husk that is produced as a by-product of gasification and combustion processes is rich in silica content and has applications in cement and ceramic manufacturing (Zain et al. 2011). Rice straw also has applications in paper mills as a substitute for pulp (Kumar et al. 2015). In practice, however, the use of rice straw for power generation and other industrial applications has been limited. First, rice straw is a geographically dispersed and low-density material, the cost of collection, transportation, and processing of the rice residue can surpass the benefits gained from using it as raw material (Bhuvaneshwari et al. 2019). Analyses also show that it is not always economically feasible to generate power from rice residue (Kumar 2017) given the costs of production vis-à-vis the price of power. Furthermore, rice straw has a high silica content, making it poor-quality fodder for cattle unlike residue from other staples, such as wheat (Lohan et al. 2018;Na et al. 2014). For all these reasons ex situ management of rice residue has been limited.
Studies have analyzed that the techno-economics of in situ (Shyamsundar et al. 2019) straw disposal (Kumar 2017;Bhuvaneshwari et al. 2019). Shyamsundar et al (2019) provides a comprehensive comparison between the profits from 10 different residue management practices, including Happy Seeder, Baler, Rotavator, and residue burn options. They also examine public costs related to residue burning, including government subsidies, health and economic costs related to greenhouse gas (GHG) emissions, and groundwater depletion. In this study, we shed light on the actual decision-making processes used by farmers for managing rice straw. Consequently, the costs and considerations for key farming decisions are derived from farmer interviews, instead of idealizations of 'rational' farmer behaviors (Shyamsundar et al. 2019). Here, we focus on the process of decision-making used by farmers, as we seek to answer the following questions: • What are farmers' perspectives on the available clean alternatives available and the challenges in adopting them? • What are the costs of rice residue management of these clean alternatives as expressed in farmer interviews and how do these costs influence practice? • How do farmers perceive targeted government incentives and are these incentives effective in reducing agricultural waste burning?

Methods
A mixed methods approach was used to quantify the tangible costs (capital and operational) and to characterize the intangible constraints faced by farmers that are associated with sustainable management of rice residue. Farmer and expert survey interviews were used in the quantification of cost of residue management and post-harvest field preparation, as well as to provide insight into farmer views on access to technology and the role of available government policies. Content analysis of interview codes was used to quantify the common residue management practices adopted by farmers and the conditions under which they are operated. These interviews (and survey questions) helped highlight the factors influencing farmers' decision-making. The interview protocol contained questions related to technological alternatives to AWB, their operating conditions, financing/costs of field preparation for the next season, perceptions of government policies, and their impacts on residue burning, in addition to broader questions about their farming practices. Questions were aimed at covering environmental impacts of residue burning, costs of residue management, farmer perceptions toward government policies, and other policies. Expert interview protocol consisted of questions focused on the scope of in situ and ex situ residue management practices and their perceptions of challenges faced by farmers in their adoption, responsibility, and role of central and state governments to support farmers in reducing residue burning.
Interviews were conducted in December 2019 with 40 farmers in the North Indian state of Punjab. 12 farmers were interviewed from Ludhiana district, 14 from Patiala district, 8 from Bhatinda district, and 6 from various other districts in Punjab interviewed at a farming expo. One participant was a wage-earning farm worker, 38 participants cultivate rice in the Kharif season on land they own or lease, and one farmer cultivated vegetables along with rice as part of contract farming. During the Rabi season, 38 farmers cultivate wheat as the primary crop after harvesting rice, while one farmer cultivates vegetables, potato, and maize on the entirety of his field during the Rabi season. 15 of the interviewed participants cultivate on farms of 10 acres or less. 17% of interviewed farmers were small-scale farmers (5 acres or less), 31% were medium-scale farmers (6-15 acres), and 53% were large-scale farmers (greater than 15 acres). Figure 2 shows landholdings from Haryana and Punjab. The overrepresentation of large-scale farmers in this study is due to access to and availability of farmers for the interview. Expert interviewees consisted of 12 experts, including scientists, agricultural science professors from universities in Punjab, and government officials from Punjab.
All interviews were translated from Hindi/Punjabi and transcribed to English. The qualitative data analysis software, NVivo, was used to deductively code the interviews on the following primary themes: crops grown, cost of cultivation, perceived impacts of residue burning, available residue management options, and perception toward government incentives. Inductive themes such as concern for groundwater level, impact of residue burning on air pollution, and trust in the system were also coded.

Rice residue management -farmers' decision-making
Variables such as access to technology, cost of operation, or pest presence that influence farmers' residue management practices emerged as key themes from farmer interviews related to farmer decision-making. Residue management practices, their conditions of operation, and variables influencing each decision were aggregated into a flowchart (Fig. 3) that demonstrates the process used by farmers in decision-making for rice residue management. This flowchart helps present a more realistic picture of actions taken on the ground and their associated rationale and demonstrates that four primary factors influence farmer decisionmaking-technology choice, farm size, time available for residue management, and availability of machinery.

Choice of technologies used
Active residue management practices, i.e., management practices that do not require residue burn, generally require a Super Straw Management System (SMS) attached to the Combine Harvester. In the absence of SMS, the Rotavator scenarios can be implemented. However, managing rice residue without a Super SMS drives up fuel cost. Active residue management scenarios also require a high horsepower (HP) tractor to operate the machines (either a Super SMS or a Rotavator). This dependence on high HP tractor is a key step in the farmer decision pathway. Each technology comes with its own set of choices and constraints. A Rotavator alone is insufficient for in situ incorporation for a rapid two-week turnaround between harvests. Farmers may need to purchase or rent a Reversible Plow or Mulcher along with Rotavator. These considerations add to the capital, operational, diesel, and labor cost. Similarly, the adoption of Happy Seeder depends on farmers' perception of pest and insect attack in the wheat crop. The use of Baler to collect residue is only feasible if the farmer has means of disposing the residue or giving it away as raw materials to factories or power plants.

Small vs. large farms
Machinery used for residue management including (i) Super SMS mounted on a Combine harvester, (ii) Happy Seeder, (iii) Super Seeder, and (iv) Baler require high-power tractors (> 45 HP) to operate. Thus, the adoption of sustainable residue management is predicated on the availability of and access to a high Horse Power (HP) tractor. It is feasible for farmers cultivating on medium to large farms (10 acres of more) to purchase a heavy tractor of 50 horsepower or more, whereas purchasing a high HP tractor may be beyond the economic capacity for farmers operating on land less than or equal to 5 acres. Since machinery such as Super SMS, Happy Seeder, and Baler require high HP tractors to operate, the use of these machinery is limited to medium-to large-scale farms. Thus, access to a high-power tractor is a limiting first step in the decision to burn agricultural residue. Those without such access are more likely to burn the residue.

Timing and performance constraints
Sowing short-duration rice variety enables farmers to harvest the rice crop early, giving them a longer time window to manage the residue before sowing the next crop, primarily wheat or potato. Given more than 10-15 days between harvesting rice and sowing the next crop and when provided with reliable electricity supply, farmers can fill the post-harvest fields with water and allow the rice residue to decompose and mulch. This makes it easy to incorporate the rice residue into the fields with a simple Rotavator or Reversible Plow, without having the need to burn the residue. However, the commonly cultivated variety by the interviewed sample size is Pusa 44, a high-yield and long-growing hybrid that leaves farmers only 10-15 days to manage the residue and prepare the field. Thus, a farmer growing long-duration rice variety is very likely to burn the residue on the field in order to clear the field quickly.

Availability of machinery
Marginal and small-medium-scale farmers often need to rely on the availability of agricultural machines on rent through farmers' co-operatives, Custom Hiring Centers (CHCs), or wealthier farmers in the village. This is particularly challenging given the short-time period available to manage the residue and clear fields to sow the wheat crop. Subsidies provided by the central government for the purchase of selected agricultural machinery (50% of capital cost) from licensed manufacturers makes it feasible for medium-large farmers to purchase these machineries. Eight or more farmers can also apply for group (at 80% of capital cost) subsidy. Formation of farmer societies has been encouraged by the government

Rice residue management: a cost analysis
Residue management practices that emerged from farmer interviews had varying costs and incentives associated to them. We present the results of a cost analysis for six different residue management practices that emerge as candidates from farmer interviews in Fig. 4. Details of the cost analysis can be seen in the Supplementary Information, with the conditions related to each practice shown in Table SI-1. Farm equipment can either be purchased or rented, with each having their own associated costs and constraints. Renting machinery obviates the need for upfront high capital cost, while adding possible time and availability constraints since a limited amount of local machinery are available for rent. In addition, some technologies including Super SMS, Happy Seeder, Super Seeder, Reversible Plow, Zero-Till Drill, and Mulcher are eligible for government subsidies on purchases. These include a 50% discount on capital cost for individual farmers and an 80% discount for farmer groups with shared ownership. There are other subsidies as well. During the harvest season of 2019, the government provided an incentive (a "bonus") of INR 100 per quintal (100 kg) of yield to small-scale farmers (defined as those who own 5 acres or less of land) and who have not participated in agricultural waste burning. We also assume that large-scale farmers (defined as those who own > 5 acres of land) have the capacity to purchase machinery. Renting of machinery is assumed to be done only by small-scale farmers. The government incentives thus include fixed bonus incentive (available for small-scale farmers only) for active residue management through renting machinery, 50% subsidy on purchase of agricultural machinery (except for Rotavator), and monetary benefits for giving away the residue (when using Baler). Table SI 7.2 summarizes notes from farmers about the pros and cons of managing rice residue using Happy Seeder, Baler, and Rotavator. Figure 4 shows annualized net present value (NPV) for 10 years of cost of field preparation using the six common residue management practices with government subsidies. The cost for each practice includes cost of machinery, cost of operation (diesel, labor, and repair), and cost of field preparation costs (fertilizer and insecticide) for the next sowing season. For the purpose of this analysis, we have assumed that all equipment purchased is new.
In the absence of government subsidies, burning residue is the cheapest residue management option, followed closely by buying and renting a Rotavator. Government incentives of subsidy on machinery and incentives to not burn (for small farmers) are typically required to make alternatives less expensive than residue burn. Annualized costs for renting and buying Rotavator are only 2% and 12% higher than residue burning without subsidy. Including a government subsidy for capital costs and fixed bonus for small-scale farmers makes Rotavator Fig. 3 Farmers' rice residue management decision-making process. The farmer decisions depending on access to technology are presented in blue. The final residue management processes in green are environmentally friendly outcomes desirable by farmers and policymakers. The status quo of residue burn is represented in red as it is an overall undesirable outcome.  Fig. 4 Annualized Costs of field preparation (over 15 years) with different rice residue management practices with and without government incentives. The government incentives include 50% subsidy by the government on purchase of agricultural machinery, except Rotavator. Net Present Value (NPV) is calculated with a discount rate of 10% over 10 years. Assumed inflation rate is 3.34%. It is assumed that farmers make 20% down payment of capital costs when purchasing the machinery and request a loan from banks at the 7% interest that is paid off in the first 5 years. It is further assumed that farmers receive subsidy on purchase of machinery at the end of first year, based on responses from farmers use even more attractive-20% cheaper for rent and only 2% more expensive for purchase. The annualized costs for renting and buying a Happy Seeder are 64% and 54% higher than residue burn. In the presence of government subsidy and fixed bonus, these costs are 42% and 44% higher than residue burn, respectively, for renting and buying a Happy Seeder. Finally, the cost of renting a baler is 22% greater than residue burning without government incentives. However, with the bonus and monetary benefits received for selling the residue, renting a baler is 31% cheaper than residue burn. Ex situ residue management using Balers are the cheapest option with government subsidies. However, the costs do not account for the infrastructure required for disposal of residue and so it is difficult make a generalized comparison with other in situ methods. Figure 5 represents a breakdown of the field preparation expenses for the different residue management practices considered. These expenses include costs of machinery (purchasing capital and rental, including combine harvester), costs of operation (labor, diesel, and repair), as well as field preparation costs (fertilizer and insecticide) during the following season. The capital costs for the machinery purchase option correspond to full machinery cost before government subsidy.
Equipment costs include cost of using a combine harvester with Super SMS for harvesting, in addition to the cost of other machinery used for residue management. These capital costs make up a large part of the expense borne by the farmer for each residue management practice, from 17% for purchase of Happy Seeder to 44% for renting a Baler. Fuel costs make up a substantial portion of the total. For example, the use of a Rotavator in combination with other machines such as a reversible plow or mulcher to manage rice residue in situ, the tractor needs to be operated on the field several times which increases the diesel consumption. Diesel costs are lowest for full or partial burn as this alternative requires the least field preparation. Annualized diesel costs range from between 39% (Happy Seeder) and 13% (residue burn).
Farmers reported that fertilizer costs are lowest for in situ incorporation, i.e., using Rotavator. This is consistent with the idea that in situ incorporation rice residue balances the soil nitrogen content and adds organic carbon and other nutrients back into the soil (Kumar et al. 2015). Both farmers and experts highlighted the increase in insect and pest attacks in the winter wheat crop after using Happy Seeder to sow wheat. This is reflected in the high insecticide cost for purchase and rent of Happy Seeder.
These findings differ from those of (Shyamsundar et al. 2019) especially with respect to Happy Seeder, which their analysis finds to be the most cost-effective alternative. These differences emerge from differing assumptions used in the two studies. The differences in cost of field preparation can be attributed primarily to Shyamsundar et al.'s inclusion of irrigation and harvest operation costs. Our study further differentiates accessible alternatives based on size of land holdings. Through the farmer interviews, we conclude that medium-to large-scale farmers have the financial capability to purchase agricultural machinery, whereas small-medium-scale farmers rent the equipment required for harvesting and field preparation. Our cost analysis accordingly contains cost for machinery purchase and rent attributed to differences in land holdings.
Overall, we find that residue management practices can be cost effective relative to residue burn, especially when existing subsidies are applied. The costs of field preparation and considerations for key farming decisions used in this paper are derived from farmer interviews and are thus a reflection of farmer behavior as narrated by farmers through semi-structured interviews, instead of idealizations of 'rational' farmer behaviors. However, as we discuss below, a cost analysis alone does not paint a complete picture, and there are various reasons for the insufficient adoption of active residue management practices. These include, inter alia, farmer's capacity to purchase or rent high HP tractor, access to agricultural machinery, timing of and access to government subsidy, perceptions of pest infestation, and soil conditions post-residue management all play an important role in the adoption of any residue management practice. We explore these further in the next section.

Farmers' perspectives and Government incentives
As we highlight in this section there several reasons for the difference between economic analyses presented earlier and what farmers say is possible. These reasons are characterized in Table 1 and broadly fall into three categories: Trust, Awareness, and Access. In what follows we describe how increasing trust in government measures and increasing  Table 1 shows that access constraints where farmers are required to purchase machinery at cost from a licensed manufacturer and the subsidy is then received in the form of reimbursement between 6 months and 2 years later. This can add to the carrying costs of a farm and to the financial burden. Small-scale farmers also noted that they cannot purchase expensive agricultural machines and thus do not receive the benefit of the government subsidy.
In addition, the cost of residue management with the help of a Rotavator is lower than the residue burn option. However, this residue management option depends on the availability of Rotavators and other machines (e.g., Reversible Plow and Mulcher) in the short harvest period, as well as farmers' access to large tractors to operate the machinery.

Payment/bonus to not burn
In 2019, the Supreme Court of India required governments to provide an INR 100 per quintal (100 kg) of yield as a bonus to small-scale farmers who own 5 acres or less of land and who have not participated in agricultural waste burning. Since the average yield in Punjab is 3 tons per acre, small farmers stood to receive a revenue of INR 3000 per acre. This is above the amount of Rs 2500 per acre suggested by (Jack et al. 2021) on the basis of a randomized control trial, but only half of what the farmers union is currently demanding (Tribune News 2021). In our interviews small farmers were encouraged by the idea of such an incentive, which could be effective in offsetting the extra cost of active residue management. However, the provision of an incentive may not solve the problem of AWB, as the onus of residue management falls on farmers who often do not have access to the means and infrastructure to use AWB alternatives. As noted in Table 1, a key assumption of the scheme, that farmers get the appropriate government subsidy on time, is often not met. In addition, uncertainly in availability of the fixed bonus government incentive in future also makes it difficult for farmers to plan for residue management in advance.
Ex situ rice residue management Ex situ residue management after collecting residue from the field with Balers, was considered a preferred residue management practices by farmers and experts alike. Table 1 shows that farmers see a potential value in selling rice residue to power plants and factories, both in terms of personal monetary benefit and conversion of residue from waste to raw material. Farmers however recognize that it is not in the government's capacity to facilitate the collection of residues from all the fields around Punjab. The short-time window available for rice harvest and field preparation makes the task further challenging. Interviewed farmers suggest local residue collection centers be set up by the government in each village during the time of harvest where farmers can deposit their collected residue in time. A major access constraint for ex situ residue management is the access to Baler machines in the shorttime window and a heavy tractor required to operate it.
Interviewed experts stressed the utilization of rice residue as a useful raw material highlighting several possible, but as yet unrealized, applications of rice residue, such power generation, and use as raw material for paper and building board. However, small-scale farmers may not have the tractors required to operate a baler, and the availability of balers is limited during the short harvest period of rice, and farmers do not necessarily receive monetary benefits for the residue that they give away. There is currently limited but growing capacity to absorb large amounts of residue generated in the region and the greater cost of ex situ management falls on the government and the private sector to establish infrastructure and industry.
Our expert interviews also uncovered other possible solutions related to the AWB problem which we briefly describe below; these include growing of alternative rice varieties and changes to incentives for crop diversification. Currently the most common rice variety in Punjab is Pusa 44, a longduration hybrid variety that takes 130 days to mature after a 30-day nursery transplantation period. Growing short-duration rice variety provides farmers with a longer time between rice harvesting and sowing of wheat crop for the next season, enabling the anerobic decomposition of rice residue and so make in situ incorporation more feasible. Further, agricultural waste burning of short-duration varieties rice takes place in pre-winter conditions and thus have reduced impact on ambient air pollution in the Indo-Gangetic Plain. The Ministry of Agriculture's direction to the Food Corporation of India (FCI) and other state procurement agencies not to procure this variety in Punjab (Nibber 2020), if implemented, may be an effort to reduce area under Pusa 44. Both farmers and experts were convinced that encouraging farmers to cultivate Basmati and other short-duration varieties of rice can also help reduce AWB.
Further, the Food Corporation of India (FCI) buys rice grains from farmers in Punjab at MSP to meet India's national food security demands. The provision of rice procurement at MSP and guaranteed monitory return at the end of the Kharif season is the main reason farmers continue to grow rice despite contrary recommendations from scientists and policy experts and depletion of groundwater resources. Interviewed experts stressed in importance of crop diversification in Punjab and recommended phasing out rice cultivation in parts of the state while simultaneously setting up markets for other less water-intensive crops. Almost 90% of the interviewed farmers expressed interest in switching away from cultivating rice, if provided with a reliable market for other crops. Crops such as maize, vegetables, American cotton, and pulses were some of the emergent alternatives to cultivating rice in the Kharif season from the interviews. Farmers are aware of the impact of rice cultivation on water consumption, which paired with subsidy on electricity for irrigation is causing unprecedented groundwater depletion. They acknowledge the importance of moving away from rice production in Punjab, a crop consumed primarily for export to other states, to save depleting groundwater levels. Thus, they recognize that agrarian solutions to the foodwater-energy nexus in Punjab and mitigation of AWB are closely intertwined.

Conclusion
The three important solutions to AWB include in situ residue incorporation, ex situ residue management, and crop diversification. Active management of rice residue by renting or purchasing a Rotavator in combination with other machines, such as Reversible Plow, Straw chopper, and Mulcher, or renting a Baler for ex situ managed can in principle be more economical than burning residue. However, the adoption of both in situ and ex situ management methods may be limited due to limitations of trust, awareness, and access on their adoption. For example, access to high horsepower tractors is required to operate the heavy machinery, and the availability of machinery in the short harvest period is limited and monetary incentives are available for only some of the alternatives. Direct subsidies or bonuses to encourage farmers to not burn residue could be an important tool for managing residue. While the exact amount is subject to some debate, the current level of Rs 100 per small farmer can cover 13-21% of the total cost of residue management. Small farmers expressed concerns with the timing of government subsidy on agricultural machinery and its consequence on machinery market price inflation. Additionally, the subsidy covers roughly a third of the area (Ricciardi et al. 2018) where rice is grown in the Punjab and opens up the question of whether it should be given to all farmers and not just those who own small farms.
Overall, managing residue in situ is a complicated task. There are many machines involved, adding to financial and coordination costs of residue management, especially in the limited time available between harvests. While in theory there are various government subsidies and cash incentives to reduce costs, in practice there are a lot of uncertainty in the timelines and processes by which farmers can avail of them. Farmers thus continue to find that the best way to manage residue is by simply burning it on the field. Consequently, ex situ residue management that shifts the responsibility of managing residue away from farmers to the government or its delegates might be easier to implement. Ex situ residue management can help provide a solution to the acute phase of the Indo-Gangetic Plain's seasonal air pollution problem and could convert a waste product into a useful raw material. However, this puts the onus on governments and other private initiatives to expand infrastructure required to integrate rice residue in productive market applications (such as pellets or fiberboards). Unfortunately, none of these residue management alternatives address the continued unsustainable depletion of groundwater in Northwestern India. Solutions to agrarian problems that underlie AWB will need to more directly address the conundrums at the heart of India's food-water-energy nexus. Enabling the cultivation of crop diversification alternatives to rice production that are less dependent on the continued and potentially catastrophic, depletion of groundwater will be essential to resolve these conundrums.