Grain procurement and seasonal air pollution

To provide farmers with stable remunerative prices for their produce, the government commits to buying all surplus grain at minimum support prices. This article demonstrates that when procurement prices are higher, farmers are encouraged to specialise and produce more rice, leading to an increase in crop-residue burning and air pollution. The mortality costs of pollution are estimated to be larger than gains to producers from higher prices.

Every winter season the capital city of India, New Delhi, is in the news, and not for good reasons. The entire northern India is covered by a thick layer of smoke. This smoke emanates from thousands of hectares of paddy fields set ablaze by farmers in the foodgrain producing states of Punjab and Haryana. These fires are so intense and at such a scale that satellites are able to capture them from outer space. The region is densely populated and is home to more than 400 million people who are directly exposed to toxic gases and high levels of particulate matter (PM) released from burnt crop residue. While the decision to burn leftover crop residue may be rational from a farmer's perspective, the health costs of resultant air pollution are significant.¹ Evidence shows that exposure to pollutants released from biomass burning has severe implications for human health, adding to the disease burden and stressing the limited resources and healthcare infrastructure available in the region (Shyamsundar et al. 2019, Singh et al. 2021, Pullabhotla and Souza 2022).

Fires and air pollution

Figure 1 shows the spatial distribution of fire events (panel (a)) and average annual PM 2.5² levels (panel (b)) for the 15-year period from 2002 to 2016. The highest levels of PM 2.5 are observed in the northern belt covering the rice- and wheat-cultivating states of Punjab, Haryana, and Uttar Pradesh. Fire activity seems more concentrated in the north-west and south-east parts of the country.

Figure 1. Spatial distribution of fire events and PM 2.5 levels

Note: Maps depict the averages for fire events and total PM 2.5 levels from 2002 to 2016.

Source: Author’s estimates based on ground-level data on PM 2.5 concentration from the Atmospheric Composition Analysis Group, and time and geocoded active biomass fire data from NASA's Fire Information for Resource Management System (FIRMS).

Figure 2 shows the seasonality in fire events, PM 2.5 levels, and wind speed. Fire activity seasonally peaks in the winter months of October and November (panel (a)). The increase in winter fire activity is matched by an increase in PM 2.5 levels during these months (panel (b)). The important thing to note is that the increase in fire activity and PM 2.5 levels observed during winter months is higher for districts with foodgrain procurement and overlaps with the rice harvesting season. Weather conditions also support the fire-air pollution link, as average wind speeds are the lowest in winter months (panel (c)). This, combined with low rainfall at this time, implies that pollutants from fires remain suspended in the atmosphere for longer durations. Based on the observed seasonal patterns and the rice harvest window, my research ((Negi 2024) focuses on total fire events and PM 2.5 levels for the winter months of October, November, and December.

Figure 2. Seasonal variation in fires, air pollution, and wind speed

Note: The graphs show monthly averages for 2002 to 2016 with 95% confidence intervals. A 95% confidence interval indicates that, if the experiment was repeated over and over with new samples, 95% of the time the calculated confidence interval would contain the true effect.

Government intervention in grain markets

A unique feature of the Indian agricultural marketing structure is that the federal government is the largest buyer of staple foodgrains like rice and wheat and commits to buying all the surplus at fixed price floors called the Minimum Support Price (MSP), which are announced at the beginning of the rice and wheat planting season. This policy is in place to provide farmers with stable remunerative prices for their produce but is not uniform in terms of its coverage. Figure 3 shows districts with government grain procurement operations.

On average, government agencies procure almost one-third of the total rice production in the country, with the purchase volumes varying across states. A few states, like Punjab, Haryana, and Andhra Pradesh, sell a large share of the total rice production to the government. For example, half of the rice and wheat produced in Haryana and two-thirds in Punjab is procured by the government. Is this policy leading to greater crop residue burning and higher air pollution in regions where the government interferes in local agricultural markets?

Figure 3. Spatial variation in foodgrain procurement

Note: Red shaded regions are districts with government procurement operations for the agricultural year 2012-2013.

Source: Procurement districts were identified based on the National Sample Survey Organisation's Situation Assessment Survey of agricultural households, conducted in 2013.

Grain procurement and residue burning

Figure 4(a) plots the trends in MSP for rice and wheat in real terms. The MSP for rice and wheat was on a slightly downward trend before 2006 but surged around 2007. This was because, as global rice and wheat prices rose around 2007-2008, domestic MSPs were also increased to maintain parity between global and domestic prices.

Figure 4. Assured prices, fires, and air pollution

Notes: (i) All prices are in real terms and deflated by the national income deflator. (ii) Panels (b) and (c) plot the difference in average outcome for treated (subject to government procurement) and control (not subject to intervention) districts over the years, with 95% confidence intervals. (iii) Estimates are from regressions with district fixed effects, year fixed effects, and a full set of control variables. Control variables include road length (in kilometres), number of bank branches, average nightlight intensity, wind speed, rainfall, temperature and pressure. (iv) The base year is 2006, indicated by the vertical red line. (v) On the horizontal axis, negative numbers indicate years preceding the base year and positive numbers indicate years after the base year. (vi) Total fire events and PM 2.5 levels are for October, November, and December.

Source: Data from the Ministry of Agriculture, Government of India.

Figure 4 panels (b) and (c) show the trends in fire incidents and PM 2.5 levels in districts with and without government procurement districts, using 2006 as the base year. Following the increase in MSP, fire incidents in procurement districts show a consistently upward trend starting from 2006. In contrast, estimates for control districts remain relatively stable, hovering near zero in most years. Similarly, PM 2.5 levels in procurement districts exhibit a significant increase post-2006. Although districts without procurement also experienced a slight upward trend in PM 2.5 levels after 2006, this increase is less pronounced than the sharp rise observed in procurement districts.

In terms of estimates, the study finds that higher MSPs post-2006 led to districts with government procurement of foodgrains experiencing a 43% increase in fire events and a 6% increase in PM 2.5 levels. This happened because higher support prices and procurement led to greater specialisation and higher rice production in districts with government intervention. Since farmers in northern India mainly depend on government agencies as the primary buyers of their output, they responded to higher prices by producing more rice. The result is more residue and more burning.

Health costs and redistribution

What are the health implications of additional fires and pollution exposure brought about due to higher MSPs and grain procurement policy? The study finds a 2 percentage point higher likelihood of illness in procurement districts after the price increase. This is primarily driven by respiratory diseases like asthma, tuberculosis, and heart diseases. The estimate for respiratory illnesses is 36% of the average incidence of respiratory illnesses in the country; for heart diseases, the estimate is larger. The analysis also suggests a 19% increase in the average per-person out-of-pocket medical expenditure. For procurement districts, the overall associated increase in out-of-pocket medical expenditure comes out to be US$29 million.

The study shows that government policy of support prices does lead to gains for foodgrain-producing regions, but losses are also higher because of the pollution externality and associated mortality costs. In districts with government procurement, estimates suggest an overall income gain of US$4.5 billion due to higher prices and a mortality cost of US$5.6 billion due to resultant pollution. On average, districts with government procurement experienced a net loss of around US$1 billion. In per capita terms, the net loss turns out to be US$2 per person per year.

Figure 5. Distribution of per capita net gains

Notes: (i) Panel (a) plots the distribution of per capita net gains for procurement districts. The vertical line indicates the mean (average). (ii) Panel (b) plots the non-parametrically estimated relationship between net gains per capita and the proportion of agricultural workers in the district. Agricultural workers include both land-owning cultivators and landless labourers.

These are aggregate estimates and mask the distributional aspect of these losses. Figure 5(a), which presents the distribution of net gains per capita, shows that some districts also experienced net losses. Positive net gains were primarily experienced in the surplus grain-producing states of Punjab and Haryana but at the cost of rest of the region. Finally, Figure 5(b) shows a negative association between the proportion of farm workers and net gains, implying that regions with greater employment in agriculture also experienced net losses.

Policy implications

The finding that the price-fire-air pollution link is driven by districts where the government is the largest buyer of foodgrains is telling of the agricultural policy-supported cultivation system practiced in India. Given the distorted nature of Indian agricultural markets, where the government procures surplus production at fixed price floors, upward movements in rice prices are matched with greater rice procurement. This procured foodgrain has found its way into the export markets in recent decades to the extent that India has now emerged as the largest exporter of rice globally. This is ironic given that farmers in northern India produce this rice with a slew of input subsidies and with immense environmental costs to the region. These findings highlight the need to factor in social costs of policies supporting agriculture.

The policy response to residue burning has mostly been in the form of bans and fines on such activities. Such policies, however, are poorly implemented and have been ineffective in curbing crop residue burning. An immediate short-run solution may lie in introducing new planting technologies that do not require residue burning or in creating markets for byproducts and crop residue. Evidence also suggests some success in financially incentivising farmers not to burn crop residue (Jack et al. 2025). Longer-run solutions would probably demand a rethinking of current agricultural policies.

Notes:

Farmers in northern India generally follow the rice-wheat cropping system. Rice is planted around June and July and harvested in October and November, when farmers are in a hurry to harvest rice in order to prepare the fields for the winter wheat crop. Due to rising labour costs, farmers resort to mechanised harvesting which leaves substantial residue in the fields, including rice stubble. Given the time and cost constraints, farmers practice residue burning as a cheap and fast method of preparing fields for the wheat crop.
Particulate matter of 2.5 microns diameter or less.