Averages and Hotspots: Pesticide Intensity, Residue Limits and Trade

The Lift Line

A national average of 0.3 kilograms a hectare and a farmworker in a sprayed cotton block are both accurate descriptions of Indian pesticide use. Only one of them is an answer to a safety question.

Why This Editorial Matters for Your Exam

Agricultural-trade answers tend to stop at “non-tariff barriers restrict Indian exports” without explaining the mechanism. This supplies it: the shift of residue limits from a toxicological to an analytical basis, and the WTO discipline that makes it contestable. It also carries an unusually clean example of how a true statistic, the national application average, can be used to answer a question it does not actually address.

GS Paper 3: Agriculture, issues of buffer stocks and food security, marketing and transport constraints, e-technology in aid of farmers; economics of animal rearing and food processing; environmental pollution.

GS Paper 2: Bilateral, regional and global groupings and agreements involving India; effect of policies of developed countries on India’s interests.

Background and Context

The author is Sarath Babu Balijepalli, writing in The Statesman on 5 August 2026.

Declare the interest before using the piece. The author’s own bio line identifies him as President of the Plant Protection Association of India. This is an agrochemical-sector advocacy position, and the article is an industry defence of pesticide use rather than neutral analysis. That does not make its claims false, and several of them are sound. It does mean each figure has to be checked independently rather than accepted, and, as set out below, a number of them do not survive checking.

The context is an agricultural export sector that has moved in a narrow band for five years: $50.2 billion in 2021-22, a peak of $53.2 billion in 2022-23, $48.8 billion in 2023-24, $51.9 billion in 2024-25 and about $52.6 billion in 2025-26. “Stagnant” is fair; “about $50 billion” understates the level slightly.

One causal caution belongs here. The dip in 2023-24 was driven principally by India’s own export restrictions on rice, wheat, sugar and onions, not by European residue enforcement. Attributing the plateau to non-tariff barriers is not supported.

Do not reproduce the left-hand column. The editorial’s figures are drawn from at least four different vintages, and one is not merely old but superseded. FAO has revised China’s series downward by roughly seven to eight times: current FAOSTAT puts China at 1.70 kg/ha, and the series never exceeds 2.59 kg/ha across 1990 to 2024. The 13.07 figure reflects a pre-revision reading of formulated product rather than active ingredient. The United States value of 3.57 matches no year at all in the current series, which runs between 2.10 and 2.87. And India’s 0.3 comes from Ministry of Agriculture consumption data, not from FAOSTAT, which carries a frozen and evidently artefactual constant for India.

The core claim survives; one of its headline comparisons does not. On current FAOSTAT for 2023, India at 0.24 kg/ha ranks 168th of 195 countries, roughly ten times below the world average and twelve times below the United States. India is genuinely among the world’s least pesticide-intensive agricultures, and that is the point worth carrying.

But the China comparison collapses. The real ratio is about seven to one, not the forty-three to one the editorial’s figures imply, and China is no longer an outlier at all on the current data. An answer that reproduces “China 13.07” is reproducing a withdrawn number.

There is an irony here that the editorial does not intend. Its whole thesis is that a true-looking number can mislead when its provenance is not stated. Its own comparative table is the clearest illustration in the piece.

The Core Argument / Issue

Why the threshold argument is the strong one

The most persuasive element of the case is not the comparative-intensity table but the nature of the limit being enforced.

A toxicological residue limit is derived from an acceptable daily intake, itself derived from animal studies with safety factors applied, and then translated into a permissible concentration given expected consumption. It answers the question: at what concentration does this substance begin to pose a risk?

An analytical default limit of 0.01 mg per kg answers a different question: what is the smallest quantity our instruments can reliably detect? Those are not the same standard, and only one of them is a health standard.

The consequence is peculiar. As detection technology improves, the effective stringency of a default limit increases automatically, without any new toxicological finding and without any regulatory decision being taken. A consignment acceptable in 2015 can be rejected in 2026 on the same residue level, because the instrument now sees it.

This is where the SPS Agreement bites. It permits members to adopt measures necessary to protect human health, but requires that such measures be based on scientific principles and not maintained without sufficient scientific evidence, and that where an international standard exists, notably a Codex MRL, a member adopting a stricter one must justify it by risk assessment. A default threshold set at the limit of detection, applied to a substance with an established Codex limit, is precisely the kind of measure that discipline exists to test.

Why the intensity argument needs care

Here the editorial’s own case has to be handled with more discipline than advocacy usually applies, and a strong answer says so.

A national average is a quotient, not a description of practice. India cultivates a very large number of distinct crops, put at 554 in the editorial though the figure could not be traced to an official source, across agro-climatic zones ranging from Kerala’s wet tropics to Rajasthan’s arid west. By most estimates only a quarter to a half of cultivated land receives structured plant protection at all, with published work putting pesticide application at around a quarter of the cultivated area. That means the denominator includes very large areas of low-input and rainfed cultivation that apply almost nothing, which pulls the national mean sharply down.

The corollary is that intensive blocks, in cotton, in vegetables under polyhouse cultivation, in table grapes grown for export, can and do apply at rates far above the mean. The relevant figure for a farmworker’s occupational exposure, or for a specific consignment’s residue level, is the local application rate, not the national average.

An argument that meets a hotspot concern with a national mean is answering a different question. The trade argument survives this; the general safety argument does not follow from it.

The comparative cancer figures, and why they prove less than they appear

The editorial cites age-standardised cancer incidence per 100,000 from IARC’s GLOBOCAN, and the set it uses mixes at least three editions. Its world figure of 196.9 and Australia at 462.5 are GLOBOCAN 2022. Its India figure of 104.1 is GLOBOCAN 2024. Its New Zealand, Ireland and United States values match neither edition, and Ireland at 375.5 is roughly thirty points above the 2022 reading.

On a single consistent edition the picture is: GLOBOCAN 2022, world 196.9, Australia 462.5, New Zealand 427.3, United States 367.0, India 98.5; or GLOBOCAN 2024, world 194.9, Australia 461.3, United States 358.5, India 104.1. Pick one and stay in it.

The figures are real. The inference drawn from them is weaker than it looks, for a reason worth understanding.

Recorded incidence is a function of detection. A country with universal screening programmes, high diagnostic access and a complete population-based cancer registry will find and record cancers that a country with limited screening, later presentation and partial registry coverage will not. India’s cancer registry coverage extends to a fraction of the population, and a substantial share of Indian cancers present at an advanced stage, which is itself evidence of under-detection at earlier stages.

Some of the gap between India and the high-incidence countries is therefore real difference in incidence, driven by age structure, since cancer is overwhelmingly a disease of older populations and India’s median age is far lower, and by risk-factor prevalence. Some of it is difference in finding. The data cannot separate the two, and a comparison used to argue that Indian agriculture is safe is loading more weight on it than it will bear.

The “cancer train” and what can honestly be said about it

The editorial addresses the well-known train running overnight across roughly 325 km from Bathinda in Punjab to Bikaner in Rajasthan, which acquired its popular name from the number of cancer patients travelling on it.

One correction on the identification. The editorial gives the train as No. 14888/339. 14887/14888 is the Rishikesh-Barmer Express, a 1,190 km service that does pass through both Bathinda and Bikaner but is not the train in question. The service actually so nicknamed is the Abohar-Jodhpur connection, historically numbered 339 and now operating as 14721/14722. The historical 339 in the editorial is right; the 14888 is not.

Its argument is that the traffic is explained by economics rather than environment: Indian Railways provides free tickets to cancer patients and steep concessions for an accompanying attendant, and the destination is the Acharya Tulsi Regional Cancer Hospital and Research Centre at Bikaner, which offers subsidised oncology. On this account, the train concentrates patients travelling toward affordable treatment rather than revealing a concentration of disease at the origin.

That mechanism is real and is frequently omitted from popular accounts. But it does not settle the question, and it should not be presented as though it does.

A subsidised-care explanation accounts for why patients travel on that route. It does not by itself establish the underlying incidence in the districts they come from, which is a separate empirical question requiring registry data rather than transport data. The editorial’s further claim, that Punjab ranks in the lower-to-moderate range on per-capita cancer incidence relative to other states, is the load-bearing assertion in this section and is contested in the literature. It should be treated as a claim requiring verification against population-based registry data, not as an established finding.

The registry data do not support the claim as stated, and there is a deeper problem with using registry data to make it at all.

On the numbers first. The Patiala Population-Based Cancer Registry records an age-adjusted incidence of 108.2 per 100,000 for men, which is close to the median of the 32 registry entries and makes “moderate” defensible for male incidence. For women it records 124.6, which places Patiala in the upper third nationally, above Mumbai and Bhopal. “Lower to moderate” is therefore not supportable for women. Both figures sit far below the highest-incidence registries, which are in the Northeast, where Aizawl records 269.4 for men and Papum Pare 219.8 for women.

The decisive objection is about coverage, not level. Patiala is Punjab’s only population-based cancer registry, and ICMR uses it alone to represent the whole state. Patiala is not in the Malwa cotton belt. Bathinda, Mansa, Muktsar and Faridkot, the districts where the pesticide-cancer claim is actually made, have no population-based registry at all.

So the registry data can neither confirm nor refute the Malwa claim, because it does not measure Malwa. A state average derived from a district outside the disputed area cannot be used to declare the concern unfounded. That is a null result from an instrument pointed elsewhere, presented as a positive finding, and the editorial’s word “confirms” is unwarranted.

Neither position establishes anything about causation, which is a separate question requiring exposure data rather than incidence data. Cancer is multifactorial: tobacco use, diet, obesity, infection, genetics and above all age. That is a reason for caution about attributing a cancer cluster to pesticides. It is equally a reason for caution about ruling the attribution out.

The problem the low average genuinely conceals

The strongest evidence that a low registered-use average is not a sufficient safety argument comes from the editorial’s own reporting.

The editorial cites market surveillance in Karnataka finding 250 samples entirely illicit, with up to twelve undeclared molecules. Three corrections are needed before this is used.

It is a 2020 dataset, arising from an RTI disclosure, not current surveillance, so presenting it as a present-day finding is six years out of date.

“100 per cent illicit” is misframed. The 250 samples were a targeted set of products marketed as bio-based or organic, tested precisely because they were suspected. Across general surveillance in 28 states over eight years, roughly 2.45 per cent of samples were off-specification. A targeted hit rate is not a population rate.

Pyridaben is registered in India. The editorial names it as unregistered; it appears at serial 272 of the CIB and RC registered list and is sold as a 20 per cent formulation for mite control on tea, cotton and chilli. It is not even named in the underlying Karnataka source. Nitenpyram does appear genuinely unregistered, and is named in the source.

What actually matters is simpler and survives the correction: the product was undeclared. A formulation sold without a truthful label claim is one whose residues no buyer can test for, no advisory can anticipate and no pre-harvest interval can govern.

Counterfeit and undeclared product does not appear in application statistics, because it is not sold through recorded channels and not counted. A country can therefore have a genuinely low registered application intensity and a real residue problem at the same time, and the two facts are not in tension.

How to Think About This (Analytical Frame)

When an aggregate statistic is offered as an answer to a distributional question, check whether the aggregate could be low while the thing being asked about is high. A national average pesticide intensity, a national per-capita income, a national sex ratio, a national forest cover figure: each is a true number that can coexist with severe local conditions, because an average is a quotient over a denominator that may be dominated by cases unlike the one in question. The diagnostic question is always what the denominator contains. This is not an argument against aggregates; it is an argument for matching the level of the statistic to the level of the question, and it applies to almost every “India ranks X” claim a candidate will encounter.

The Diagram in Words

Picture India’s cultivated land as a very large sheet of paper, most of it left almost bare, with a scattering of small dark patches where spraying is intensive. Now weigh all the chemical applied and divide it by the whole sheet. The number that comes out is genuinely small, and it is genuinely India’s average. But nobody lives on the average. The farmworker lives in one of the dark patches, and the export consignment that gets stopped at Rotterdam came from one too. Meanwhile, at the edge of the sheet, unrecorded product crosses a border and lands on patches that no ledger counts at all. The average is not a lie. It simply is not a map of where anybody is standing.

Way Forward

1. Contest analytically-derived residue limits through the SPS Agreement, pressing the requirement that measures rest on scientific principles and that departures from Codex be justified by risk assessment. Note the correct framing: India has raised Specific Trade Concerns in the SPS Committee and has not filed a WTO dispute on MRLs. Note too that Article 3.3 gives members an autonomous right to set stricter standards, so the stronger argument is not that the EU exceeds Codex but that a default limit set at the limit of detection is not the output of any substance-specific risk assessment, which is where Articles 5.1 and 5.6 bite.
2. Negotiate mutual recognition of accredited testing, so that a residue result certified in India is accepted at the port of entry, which addresses the cost of rejection without weakening the standard.
3. Build residue-testing capacity to export-market specification, since the ability to demonstrate what is in a consignment before it ships is worth more than any argument made after it is rejected.
4. Enforce pre-harvest intervals through farmer-producer organisations, which is the single highest-return on-farm intervention for residue compliance and requires advisory capacity rather than new regulation.
5. Attack the counterfeit and smuggled-molecule trade directly, through serialised traceability, and treat this as a food-safety measure rather than only as an industry-protection measure, since unregistered molecules are the residues nobody can predict or test for.
6. Legislate the pending reform. The Insecticides Act, 1968 remains the operative statute. The Pesticide Management Bill, 2020 was introduced in the Rajya Sabha in March 2020 and reported on by the Standing Committee on Agriculture in December 2021. It has never been passed, and because it was pending in the Rajya Sabha it did not lapse. The live instrument in 2026 is a fresh draft Pesticides Management Bill released for public consultation on 7 January 2026, which is a pre-legislative draft rather than an introduced Bill. Enact it, with periodic review of existing registrations, and pair it with support for the shift to green-chemistry molecules applied at grams rather than kilograms per hectare.
7. Publish disaggregated application data by crop and district, which would end the argument about averages by making the distribution visible.

PYQ Linkage and Practice

UPSC has tested agricultural exports, non-tariff barriers, the WTO agreements, food safety regulation and pesticide use across GS2 and GS3. The analytical-versus-toxicological distinction in residue limits is the technical detail that turns a general answer about protectionism into a specific one.

Practice question: “Maximum residue limits set at the limit of analytical detection are trade instruments wearing the clothes of food safety.” Examine this claim with reference to India’s agricultural exports, and assess what domestic capability India would need to contest such measures successfully. (250 words, 15 marks)

Interview angle: A national average of 0.3 kg per hectare and a farmworker in an intensively sprayed cotton block are both true descriptions of Indian pesticide use. If you were designing regulation, which number would you regulate against, and what would you lose by choosing the other?

Sources: The Statesman, APEDA, FAO, World Trade Organization

Source: Averages and Hotspots: Pesticide Intensity, Residue Limits and Trade — Ujiyari.com | Free UPSC & State PCS Editorial Analysis