Exit is the Hard Part

From cleaner skies in Beijing to lingering coal dependence on the Indo-Gangetic Plain, decarbonization is proceeding by accumulation rather than replacement. The result is a transition governed less by technology than by the management of loss. Flying south from Amritsar toward Delhi in winter, the Indo-Gangetic Plain reveals itself in layers. Thermal power stations rise at regular intervals, their stacks releasing pale plumes that flatten beneath the inversion. Solar arrays interrupt farmland; wind turbines crown low ridges; transmission corridors radiate outward from expanding cities. From above, the energy transition appears reassuring, as infrastructure accumulates and emissions seem to recede. On the ground, however, it feels less conclusive—more like a partial overlay, in which new systems are added atop old ones without dislodging the political, fiscal, and social arrangements of the status quo ante.

India is adding renewable capacity at remarkable speed, yet the older system hums on alongside it. Coal plants keep running even as new installations come online. Installed capacity rises; emissions fall more slowly. In 2025, coal-fired generation did fall—by roughly 3 percent, the first such decline outside the pandemic—but the drop reflected softer demand and forced curtailment rather than displacement. No capacity exited the system. At a moment when governments speak fluently of climate ambition amid energy insecurity and geopolitical strain, the aerial view reveals something more prosaic: systems are being built far faster than they are being allowed to end.

The difficulty lies less in intent than in sequence. Entry is comparatively easy. Generation can be added incrementally, project by project, megawatt by megawatt. Exit is another matter, as closure is discrete, final, and politically expensive. Systems that avoid it rarely complete the transition they announce. And where exits are postponed or misordered, transitions thicken rather than advance. Capacity accumulates, dependency lingers, and costs are pushed elsewhere.

Climate politics is often narrated in moral shorthand—leadership versus denial, ambition versus obstruction—categories that are emotionally satisfying and rhetorically useful, but analytically thin. They struggle to explain why some states install vast amounts of clean power with little corresponding decline in emissions, while others, accused of hesitation, manage more durable shifts. The difference is less sincerity than how loss is handled.

Decarbonization, in practice, is not a single undertaking but a sequence of closures. Each is uneven, irreversible, and politically charged. Assets are retired; functions must be replaced; labour and capital are stranded; liabilities surface somewhere. Costs arrive early and locally, while benefits arrive later and diffusely. Where these decisions are deferred, the atmosphere absorbs what balance sheets do not.

China illustrates this logic with particular clarity. Coal generation has begun to decline and the smog-choked winters of Beijing’s 2010s have largely receded. The improvement, however, has come less from eliminating cost than shifting it elsewhere. In the north, winter heating is not optional. The campaign to replace coal stoves with natural gas exposed the risks of exit pursued ahead of replacement. In several regions, coal systems were dismantled before secure gas supply was in place. Shortages followed, prices spiked and policy reversals ensued.

The crisis passed; what endured were the terms on which it was settled. Coal heating was highly polluting but it was cheap and locally controlled. Gas is cleaner, costlier, externally supplied, and politically priced. As transitional subsidies are withdrawn, the burden drifts downward, settling on the rural households that are least able to absorb it. Pollution has been reduced where it was most visible and politically salient, leaving costs to be displaced onto colder regions with weaker voice. The pattern is not confined to China. Coal-dependent Indian states—Jharkhand, Chhattisgarh, Odisha—find themselves in a similar bind, though within a different political economy. Coal plants there are not merely sources of power. They are employers, revenue anchors, and quiet guarantors of grid stability. Closing them is not the functional equivalent of adding capacity elsewhere. It is a wager with visible losers.

Distribution companies are financially fragile. Tariff reform is electorally radioactive. Coal employment is concentrated enough to matter, but not mobile enough to cushion its own displacement. India’s grid code now formalizes this stasis: coal plants are required to operate at a minimum load of roughly 55 percent and are compensated for availability even when renewables are abundant.Under these conditions, exit is deferred rather than confronted. Renewable capacity is added at the margins without forcing substitution. Emissions drift outward; dependency stays put. The consequences are now measurable. Between May and December 2025, India curtailed roughly 2.3 terawatt-hours of solar generation to preserve grid stability, compensating producers tens of millions of dollars for clean power not taken. Renewable energy accounts for roughly half of installed capacity but supplies barely a quarter of actual generation. The system pays twice to avoid deciding what closes.

China’s advantage lies not in virtue but in authority. Its power system is being reorganized even as coal capacity persists. Ultra-high-voltage transmission corridors reduce curtailment; dispatch increasingly favours renewable offtake; storage deployment has accelerated through mandate rather than price; coal plants are compensated more for availability than output. The system is not clean, but it is managed. Prices are contained, capital is directed, and losses are absorbed within state institutions rather than passed immediately to households.

India’s system is less insulated. Where China can sequence exit administratively and defer visible shock, India must negotiate it through fiscally weak utilities and politically sensitive tariffs. Coal is being repositioned as a balancing resource without the institutional capacity to absorb the wear, capital cost, and shortened plant lifespans that flexibility requires. The result is a familiar asymmetry: the transition advances in headlines while stalling in structure.

Other cases sharpen the contrast. Denmark reduced coal steadily by integrating generation, heating, and transmission early. Combined heat-and-power plants were converted; district heating expanded; wind displaced fossil generation within a stable institutional frame. Britain and Germany, by contrast, illustrate the costs of mis-sequencing. The UK’s rapid coal exit left gas to inherit coal’s stabilizing role without insulation from fuel-price volatility. Germany’s accelerated nuclear phase-out removed firm capacity before alternatives for system stability were secure, leaving coal plants to re-enter when pressure arrived.

Across these cases the pattern repeats. Transitions fail where exits are postponed, grids trail generation, and prices move faster than legitimacy. The green transition is less a march toward virtue than a process of managed loss. Entry is easy. What is harder is deciding what closes, when it closes, and who pays.

Over the Indo-Gangetic Plain, coal stacks keep rising because exit costs have never been faced. Cleaner air in China rests on a managed redistribution of loss; stalled substitution in India reveals what happens when that loss is left politically unassigned. Until closure is treated as a political act rather than a technical inconvenience, the transition will remain an overlay, and the aerial view will continue to reward self-deception.