Energy Subsidies: Two Numbers, Two Definitions
Subsidy figures are quoted constantly and understood rarely. The gap between the two standard estimates is more than tenfold, which is large enough that citing one without saying which is being used makes a claim nearly meaningless.
What Each Number Counts
An explicit subsidy is money. A government caps the domestic petrol price below the world price and compensates the seller; it grants a producer a tax allowance unavailable to other industries; it funds infrastructure that only one fuel uses. These appear in budgets and can be added up, and the International Energy Agency's estimate on this basis has ranged from roughly 300 to over 1,000 billion dollars a year depending on world prices - it rises automatically when prices spike, because holding a domestic price steady costs more.
An implicit subsidy is a cost imposed and not charged for. Burning coal produces particulates that cause illness and premature death, and carbon dioxide whose damage falls on people who did not buy the coal. If the polluter does not pay these costs, someone else does, and economists treat the difference as a subsidy in the sense of an unpriced input.
The International Monetary Fund's estimate on this combined basis reached around 7 trillion dollars, close to seven percent of world economic output. The great majority is implicit, and within that, local air pollution - not climate - is the single largest component, because respiratory and cardiovascular deaths are numerous, near-term and comparatively well quantified.
Both framings are legitimate and they answer different questions. The narrow figure answers what governments are paying today, which is the right question for a finance ministry. The broad figure answers what fuel would cost if its users paid for its effects, which is the right question for setting a carbon price. Trouble arises only when one number is quoted with the other's rhetorical force.
Who Subsidises What
The largest explicit fossil subsidies are consumption subsidies in energy-producing and middle-income countries. Iran, Saudi Arabia, Venezuela, Egypt, Indonesia, India and Russia have all at various times held domestic fuel or electricity prices well below international levels, and in several cases these programmes have exceeded spending on health or education.
The stated justification is protecting poor households. In practice the benefit is roughly proportional to consumption, so the largest share goes to those who consume most, who are not poor. Studies across many countries have repeatedly found the wealthiest fifth capturing several times what the poorest fifth receives, which is why cash transfers are the standard recommended alternative.
Producer subsidies are more common in wealthy countries and less visible, arriving as depletion allowances, accelerated depreciation, favourable treatment of exploration costs, and publicly funded decommissioning liabilities. They are smaller in total and harder to count, which is part of why estimates vary.
Renewable support has fallen substantially in relative terms. Early feed-in tariffs paid far above market rates and were expensive per unit of energy - Germany's early solar tariffs are the standard example of a scheme that bought expensive electricity and, in doing so, funded the manufacturing scale-up that made solar cheap for everyone else. Current schemes are mostly contracts for difference awarded by auction, and several recent offshore wind auctions have cleared at or below wholesale price, meaning no net subsidy at all.
Why Reform Fails
Removing a fuel subsidy is one of the most reliable ways to cause unrest, and the pattern is consistent enough to be predictive. Nigeria, Ecuador, Indonesia, Iran, France and Kazakhstan have all seen significant protest following fuel price increases, several of which forced partial or complete reversal.
The political structure explains it. The benefit of a cheap fuel price is concentrated, immediate and visible at the pump. The cost is diffuse - a budget line most people never see - and the beneficiaries of removing it are future taxpayers and people harmed by pollution, neither of whom form a constituency. Concentrated benefits and diffuse costs produce durable policy regardless of whether the policy is good.
There is also a fairness point that deserves acknowledgement rather than dismissal. In a country where fuel is one of the few tangible benefits citizens receive from national resource wealth, removing it while other public services remain poor is experienced as taking something away and giving nothing back. Reforms have succeeded where that exchange was made explicit.
The reforms that worked share features. Indonesia and India both reduced fuel subsidies substantially during periods of low oil prices, when the immediate consumer impact was small, and both paired removal with direct transfers to identified low-income households. Timing to a price trough and compensating visibly are the two things that distinguish successful reform from reversed reform.
Why This Matters for Anything New
Subsidy structure determines what gets built, often more than technology cost does, and a new source enters a market whose prices already reflect decades of these decisions.
The practical effect is that incumbent fuels compete at prices that exclude some of their costs, while new entrants compete at prices that include all of theirs. That is the substantive content of the phrase level playing field, which is otherwise used loosely enough to mean nothing.
It also explains why support schemes have shifted in form. A feed-in tariff pays a fixed high rate and creates a permanent constituency for keeping it. A contract for difference sets a strike price, pays the difference when the market is below it and reclaims the difference when the market is above, which means support falls automatically as a technology matures and ends without anyone having to legislate its end.
The general principle worth extracting is one that recurs across this whole section. A policy built on a price disappears when the price moves; a policy built on a structure persists. Feed-in tariffs were unwound; the auctions and carbon prices that replaced them are harder to reverse because no single group depends on their continuation.
Frequently asked questions
Why are fossil fuel subsidy estimates so different?
Because they measure different things. Explicit subsidies are money actually paid - price caps, tax breaks - and total a few hundred billion dollars a year. Implicit subsidies add the unpriced costs of air pollution and climate damage, which brings the total to several trillion. Both are legitimate and answer different questions.
What is the largest component of the broad subsidy figure?
Local air pollution, not climate. Respiratory and cardiovascular deaths from particulates are numerous, near-term and comparatively well quantified, so they dominate the estimate. Climate damage is a large but smaller component, and explicit fiscal transfers are the smallest part of the total.
Do fuel subsidies help poor households?
Much less than intended. The benefit is roughly proportional to consumption, so most of it goes to those who consume most, who are not poor. Studies across many countries find the wealthiest fifth capturing several times what the poorest fifth receives, which is why direct cash transfers are the standard recommended alternative.
Why do subsidy reforms so often fail?
Because the benefit of cheap fuel is concentrated, immediate and visible at the pump, while the cost is a diffuse budget line. Concentrated benefits with diffuse costs produce durable policy. Reforms that succeeded were timed to periods of low oil prices and paired with visible direct transfers to low-income households.
How has renewable support changed?
From feed-in tariffs paying far above market rates to contracts for difference awarded by auction. A contract for difference pays the gap when the market is below the strike price and reclaims it when above, so support falls automatically as costs fall. Several recent offshore wind auctions cleared with no net subsidy.
