Contemporary, Policy & Development Essays — UPSC CSE Mains Essay, Section B

Weightage: 125 marks, one essay from four topics. Section B is the section every General Studies candidate feels equipped for, and that feeling is the source of the problem — because the material transfers and the register does not.

1. The trap that Section B sets

A candidate who has written four General Studies papers on governance, economy, society and technology opens Section B and sees familiar ground: social media and mental health, growth and inequality, technology and employment, water, urbanisation, federal friction. The relief is immediate and it is a trap.

What follows is predictable. The candidate writes 1,100 words organised as a GS answer: the issue, its dimensions, the constitutional and statutory framework, the committee that examined it, the challenges in implementation, and a way forward. Every fact is accurate, the coverage is broad, the structure is orderly — and the score comes back at 105.

The reason is that the Essay paper is not asking for a longer GS answer, and it is scored by someone who knows exactly what one looks like. The Commission already tested policy knowledge four times. This paper is testing something else, and a script that supplies the first thing has not supplied the second.

The distinguishing quality has a name: the essay must have a reflective register. It must go under the policy question to the value question, ask what the issue is actually about, and reach a position that is the writer's own rather than a synthesis of committee findings. A GS answer explains what is; an essay argues about what matters and why.

This chapter is about that conversion. The material a serious candidate already has is the right material. What has to change is what they do with it.

2. The five moves that keep a policy essay an essay

Move one: open at human scale

Not with the scheme, not with the statistic, not with the definition. With a person, a place, a moment.

A woman who has migrated from a district in eastern Uttar Pradesh to work in a garment unit outside Bengaluru is counted in one state's population, pays into a provident fund she is unlikely to claim, and is eligible for subsidised grain in a village eleven hundred kilometres away.

That is one sentence, it contains the entire architecture of the portability problem, and it commits the essay to talking about people rather than about frameworks. The statistics can follow in paragraph three, where they will land harder for having been earned.

Move two: ask what the issue is actually about

Every policy question sits on top of a question about value, and naming the second is the move a GS answer never makes.

• A debate about data protection is on the surface about consent and compliance. Underneath, it is about whether a person retains any claim over information generated by their own conduct, and whether consent means anything when refusing it means not participating in ordinary life.
• A debate about agricultural pricing is on the surface about procurement and subsidy. Underneath, it is about whether farming is a business the state should stabilise or a way of life it should protect, which are different obligations with different implications.
• A debate about urban housing is on the surface about supply and regulation. Underneath, it is about whether a city is a place people have a claim to inhabit or a market they may enter if they can pay.

Naming the underlying question, in one sentence, usually in the introduction, is the single most efficient way to lift a Section B essay out of the GS register.

Move three: state the mechanism, not the conclusion

This is the discipline carried over from GS3 and it matters more here, not less. "Free electricity has led to groundwater depletion" is a conclusion. The essay wants the chain: power priced at zero for agricultural connections makes the marginal cost of extraction the cost of the pump alone; a farmer facing no marginal cost pumps to the point of satiation rather than efficiency; the crop mix shifts toward water-intensive varieties because they are now viable and are assured a price; the water table falls; deeper pumping requires more power, which the subsidy also covers; and the loop closes. That paragraph explains something. The conclusion merely reports it.

Move four: name who bears the cost

Every recommendation moves a cost onto somebody, and an essay that never says onto whom is not analysing but advocating. This is also the most reliable source of genuine counter-argument, because the person bearing the cost usually has a case.

Removing a subsidy imposes on current beneficiaries a loss that is certain and immediate, in exchange for a gain that is diffuse and deferred — which is a complete explanation of why such reforms fail, and it does not require attributing bad faith to anybody. An essay that includes this observation reads as somebody who has thought about implementation; one that recommends removal without it reads as somebody who has not.

Move five: refuse the way-forward list

The GS way forward — a list of five measures, each a sentence, none argued — is the clearest possible signal of register failure. Its essay equivalent is a conditional statement: what would have to be true for the problem to be soluble, and what is currently preventing it.

None of this requires new legislation. It requires that the reasons for refusing an application be recorded at the moment of refusal, which is the cheapest accountability mechanism ever devised and the most consistently evaded — because the cost of recording a reason falls entirely on the person who benefits from not having one.

That is a way forward, and it is an argument.

3. Data discipline: how to use figures that may have changed

Section B invites data, and data is where careless essays get caught. Three rules, and they are not merely defensive — following them produces better prose.

Attribute the figure to its source and its round. "The multidimensional poverty measure recorded a fall from roughly 29% in 2013-14 to about 11% in 2022-23" is safe, precise about what it is claiming, and demonstrates that you know figures come from somewhere. A bare "poverty has fallen to 11%" is unattributed, undated and invites the question of which measure.

Prefer structural ratios to point values. Point values date; structural relationships hold for years and often say more. Agriculture's share of employment remains far above its share of output is worth more in an essay than any single year's growth rate, because it states the problem rather than a symptom of it. Similarly: the great majority of judicial pendency sits in district and subordinate courts, not the higher judiciary reframes a debate that is usually conducted about the Supreme Court.

Calibrate specificity to confidence. If you are certain, be precise. If you are not, describe the relationship: "far larger", "roughly half", "an order of magnitude". An approximate statement that is true beats a precise one that is wrong, and the second is worse than no figure at all, because a sceptical examiner who catches one number will discount the rest.

The same discipline applies to live matters. Where litigation is pending, a policy is under revision, or a target is disputed, write the structure of the disagreement rather than predicting its outcome. An essay that explains why two principles conflict is more valuable, and considerably safer, than one that forecasts which will win.

4. Eight contemporary theme families

As with Section A, prepare families rather than topics. For each: the central tension, anchors worth carrying, and the counter-case that keeps the essay honest.

Technology, work and society

Tension: whether technological change displaces effort or displaces discretion, and who bears the transition.

Anchors: the distinction between automation that substitutes for tasks and automation that substitutes for judgment; the two-century record in which aggregate employment recovered while the specific displaced generation did not; the asymmetry that the gains are diffuse and the losses concentrated, which explains the politics precisely; digital public infrastructure as a case where the state built rails and private actors built on them, and the questions of accountability that arrangement leaves open.

Counter-case: the technological unemployment prediction has a long record of being wrong, and the burden is on the pessimist to say what is different this time. The serious answer concerns speed relative to the length of a working life, not scale.

Growth, inequality and welfare

Tension: growth raises the floor reliably and narrows the gap unreliably, and the two are frequently conflated.

Anchors: the multidimensional poverty decline and its magnitude; the divergence between consumption-based and wealth-based measures of inequality, which tell different stories about the same country; agriculture's employment-to-output ratio as the structural statement of the labour absorption problem; the shift from price subsidies to direct transfers, and what is gained and lost in it; states with similar incomes and very different social outcomes, which locate the explanation in public provision rather than growth rate.

Counter-case: concern about inequality in a country still lifting people out of deprivation can become a rich country's preoccupation imported prematurely. The answer is not to deny this but to name the two mechanisms that make concentration matter regardless of income level — demand constraint, and the conversion of economic concentration into influence over the rules.

Environment, climate and the development question

Tension: the sequencing claim — that development must precede environmental stringency — against the observation that the costs of deferral are not borne by those who defer.

Anchors: two distinctions almost always missed, that emissions intensity is not absolute emissions and that installed capacity is not generation; the equity architecture of common but differentiated responsibilities and why it is contested; adaptation as the neglected half of climate policy despite being where a warming country's exposure is greatest; local pollution as the more immediate and more politically tractable version of the same problem.

Counter-case: the stringency argument has a genuine distributional edge — a uniform standard is a mild constraint on a large firm and an exit decision for a small one — and that is an argument about instrument design rather than about whether to regulate.

Federalism, institutions and the state's capacity

Tension: whether the observed problems are of constitutional design or of administrative capacity, since the remedies differ entirely.

Anchors: the three channels through which centre-state relations actually run — legislative, administrative and fiscal — and the fact that most friction is fiscal while most commentary is legislative; the finding that the great bulk of judicial pendency is in the subordinate judiciary; the police reform directions and their partial adoption across states; devolution achieved in form through elections and withheld in substance by retaining funds and functionaries.

Counter-case: capacity arguments can excuse everything and explain nothing. The test is whether the same state performs well on some functions and badly on others, which points to incentives rather than to capacity.

Cities, migration and internal mobility

Tension: the benefits of mobility accrue nationally and at origin, while the servicing costs arise at destination — and no fiscal mechanism connects them.

Anchors: remittances flowing to origin while demand for water, housing, schooling and health arises at destination; entitlements attached to place rather than to person, and portability as the design response; municipal revenue bases growing more slowly than the populations they must service; census-based transfer formulas that lag movement by years.

Counter-case: migrants also generate the activity that funds destination cities. Correct in the aggregate, and it does not address incidence — the activity accrues to private employers and landlords while the servicing obligation falls on a municipal budget with no claim on it.

Information, media and the public sphere

Tension: the cost of making a claim has collapsed while the cost of checking one has not.

Anchors: attention as the optimised quantity in an advertising-funded model, which makes the harm a by-product of the revenue mechanism rather than a defect in it; the impossibility of content-based remedies that do not engage free expression; design-side and transparency-side alternatives — provenance, ranking disclosure, friction on virality; the observation that the past's information environment felt reliable because it was concentrated, not because it was checked.

Counter-case: the crisis narrative frequently overstates measured effects and understates how much the previous arrangement excluded. The strongest version is about the collapse of a shared factual baseline, not about individual credulity.

Gender, family and social change

Tension: legal equality has advanced considerably faster than the distribution of unpaid work and of safety, and the gap is where the subject actually lives.

Anchors: female labour force participation and the fact that it does not move monotonically with income or education, which defeats the simplest explanations; unpaid care work as the invisible constraint that determines participation; the distinction between formal legal capacity and effective exercise of it, particularly in property; the observation that changes achieved through law are quickest where they align with an economic interest and slowest where they do not.

Counter-case: accounts that treat all of this as attitudinal ignore the infrastructure — transport, safety, childcare, working hours — that determines whether a preference can be acted on at all.

Education, skills and the young

Tension: a system that measured enrolment improved enrolment, and what was not measured was not managed.

Anchors: near-universal enrolment alongside repeated assessment findings of large fractions of children several grade levels behind; the distinction between a credential and a capability, and the labour-market consequence when employers stop trusting the first; skilling programmes that certify supply without reference to demand; the demographic window as a claim about timing, since a young population is an asset only if it is employed.

Counter-case: learning is genuinely harder to measure than attendance, and crude measurement of it has done damage elsewhere. The answer is that the choice is not between crude measurement and none.

Worked example 4.1 (a Section B topic converted from GS register to essay register, with the two versions compared). "Social media is triggering fear of missing out amongst the youth, precipitating depression and loneliness."

The GS version, in outline. Introduction defining social media and FOMO. Section on extent, with usage statistics. Section on causes: comparison, algorithms, peer pressure. Section on impacts: psychological, academic, social. Section on the regulatory framework: intermediary rules, data protection, age provisions. Section on challenges: free speech, jurisdiction, enforcement capacity. Way forward: digital literacy, parental awareness, platform accountability, school counselling, self-regulation. Conclusion calling for a balanced approach.

Accurate, comprehensive, well organised, and it scores around 105 because it never argues anything.

The essay version.

Opening at human scale. A seventeen-year-old scrolling at one in the morning, seeing eleven people she knows at a gathering she was not invited to — an experience that before this decade would simply not have occurred, because absence was not documented and distributed.

The underlying question, named. This is not a topic about technology. It is a topic about what happens when comparison, which every society has always had, becomes continuous, involuntary and conducted against a composite of other people's best moments that no single person actually lives.

The mechanism, stated. Feeds present curated peaks; the viewer compares a full interior life against an edited exterior one; the comparison is not made once but continuously; and the deficit it produces is not against any real person but against an average of many people's highest points, which is a standard no one meets, including everyone who contributed to it.

Why this is not incidental. The business model optimises engagement, not satisfaction — and the emotions that hold attention most reliably are not the ones that leave a user better off. The harm is therefore a by-product of the revenue mechanism rather than a defect in the product, which is why appeals to platform responsibility have produced so little.

The objection, at full strength. For an adolescent who is isolated, disabled, geographically remote, or living somewhere their identity is unsafe, these platforms supply connection that would not otherwise exist and sometimes supply the only such connection. A purely negative account is not merely incomplete; it is false about the population it most affects, and the people arguing for restriction are rarely the ones who would lose most.

Absorbing it. The objection is right and it separates two things that are usually fused. The benefit comes from connection; the harm comes from optimisation for attention. These are separable in principle, which is what makes design-side remedies — chronological options, friction on virality, the ability to see without being ranked — different in kind from restriction. What the objection rules out is the blunt instrument, and it is worth noting that the blunt instrument is the one most often proposed, because it is the one that is administratively easy.

The value question, made explicit. We are asking a technology optimised for attention to carry intimacy, and then treating its failure to do so as a surprise. The prior question is what connection consists of and whether something that satisfies its outward form — responsiveness, frequency, visibility — can perform its function, which is to make a person feel known.

Conditional close. Returns to the seventeen-year-old. The problem is not that she saw the photograph. It is that she will see one every night for a decade, and that no previous generation had to develop a defence against continuous involuntary comparison because no previous generation was offered one. What is being asked of the young is an unprecedented act of self-regulation, and it is being asked by people who did not have to perform it.

What changed. The material is nearly identical. The structure argues instead of surveying, the objection is genuine and is answered by absorption, the value question is named, and the close is a conditional statement rather than a list. That difference is worth thirty marks.

5. Choosing between Section B topics

The four Section B topics in a given year will usually include one on technology, one on society or the young, one on economy or environment, and one that is a proposition about governance. Two selection cautions specific to this section.

Prefer the topic where you can state a genuine tension over the one where you know the most facts. A topic on which you hold a great deal of material is dangerous here for the reason the whole chapter is about: the material will pull you into the GS register, and the essay will be comprehensive and unremarkable.

Watch for the topic that is phrased as a proposition rather than a subject. "The cost of being wrong is less than the cost of doing nothing" appeared in Section B and is far closer to a Section A topic in what it demands. When a Section B topic is phrased as a claim, treat it as a claim — parse it, find the quantifier, decide whether you agree — rather than as a subject heading to be covered.

Common traps UPSC sets here

• Writing a General Studies answer with a larger word count — the defining failure of Section B, and it is invisible to the candidate because everything in it is correct.
• The way-forward list — five measures in five sentences, none argued, is the clearest register tell in the paper.
• Stacking committee names and scheme titles for credibility, which accumulates authority in place of reasoning.
• Asserting volatile point values — a wrong figure discredits the essay's other figures.
• Recommending without naming who bears the cost — advocacy dressed as analysis, and it forfeits the best available counter-argument.
• Predicting the outcome of live litigation or unsettled policy rather than explaining the conflict of principles.
• Never reaching the value question underneath the policy question.
• Treating a Section B topic phrased as a proposition as though it were a subject heading.

Memory aids

• "Make the concrete reflective" — the Section B correction, exactly inverse to Section A's.
• "What is this actually about?" — the question that finds the value layer under the policy layer.
• "Mechanism, not conclusion" — explain the chain rather than reporting its result.
• "Who bears the cost?" — the test that turns advocacy into analysis and locates the counter-argument.
• "Attribute the round, prefer the ratio" — data discipline in six words.
• "A conditional, not a list" — what replaces the way forward.

Exam protocol

• Open at human scale — a person, a place, a moment — and let the statistics arrive in paragraph three.
• Name the underlying value question in the introduction, in one sentence.
• Explain mechanisms rather than asserting relationships, and say where the chain weakens.
• Attach every recommendation to the party who bears its cost, and take your counter-argument from there.
• Attribute figures to source and round, prefer structural ratios, and describe relationships where you are unsure.
• On live disputes, write the conflict of principles rather than a forecast.
• Close with a conditional statement about what would have to be true, not with a list of measures.