Europe has learnt that it cannot afford another 2022. It has not learnt how to spend the money it does have.
Consider two European households. The first is comfortable: two cars, a large house, a swimming pool that needs heating. The second is not: one small flat, no car, a boiler that is switched on later each autumn than the owner would like. Their government, alarmed by the price of energy and moved by the plight of the second household, cuts the duty on road fuel and caps the price of electricity.
Over the course of a year, according to the IMF’s own arithmetic for the current shock, the first household collects about €34 from the fuel measure and €33 from the electricity measure. The second collects €9 and €11.
This is not a failure of implementation. It is what the policy does. A subsidy delivered through the price of a good is, by construction, proportional to how much of the good you consume — and the rich consume more energy than the poor, in absolute terms, by a very wide margin. Cutting fuel duty to help the poor is means-testing in reverse: the more you have, the more you get. It is the fiscal equivalent of offering everyone in the restaurant a discount proportional to what they ordered and then congratulating oneself on having fed the hungry.
European governments know this. They have been told so, repeatedly, by the IMF, the OECD, the European Commission and the European Central Bank, in documents running to hundreds of pages, since the summer of 2022. I now want to explain why they keep doing it anyway, why the excuses are better than economists usually admit, and why — on the numbers published in the past fortnight — the excuses have finally run out.
What 2022 cost
The scale of the last episode is easy to forget. Between September 2021 and January 2023, European governments allocated roughly €651 billion to shielding consumers from energy prices, of which €540 billion in the EU alone (Bruegel’s fiscal tracker). The European Commission’s own accounting puts the net budgetary cost at 1.2% of EU GDP in 2022, and 2.1% across 2022 and 2023 together.
The composition is the part worth memorising. Of that support, the Commission reports, 64% went on price measures — caps, duty cuts, VAT reductions — and only a quarter of the total was targeted at vulnerable groups. Three-quarters was sprayed indiscriminately. The OECD, counting differently across a wider set of economies, arrives at the same place from another direction: price support accounted for 66% of the value of all measures, and 94% of that price support was non-targeted (OECD, June 2022).
What did Europe buy for the money? Rather less than it hoped. A large ECB microsimulation across the euro area found that the measures compensated households for about a third of their welfare loss and closed around 60% of the inequality gap opened by inflation — but that they were “not particularly well targeted at low-income households, resulting in a higher than necessary fiscal burden” (Amores et al., ECB Occasional Paper 330). “Higher than necessary” is doing a great deal of quiet work in that sentence. The IMF has since put a number on it: fully compensating the poorest 40% of European households in 2022 would have cost 0.9% of GDP, against the roughly 2.5% actually spent on household support. Europe spent nearly three times what full protection of the bottom four deciles would have required, and still did not protect them fully.
There is a second cost, less visible on any budget line. High prices are unpleasant, but they are also the most effective conservation policy ever devised, and they work for nothing. German gas consumption in the second half of 2022 came in 23% below its temperature-adjusted baseline (Ruhnau et al., Nature Energy, 2023). No ministry designed that, no agency administered it, and no taxpayer funded it. Every euro spent suppressing the price signal was a euro spent switching that mechanism off — which is why the Commission’s own formulation is unusually blunt: price measures “weaken the incentives to adjust energy demand”, and failing to curb demand for expensive imported energy “results in a greater transfer of national income abroad to energy exporters.” One is paying, in other words, to send more money to the people who raised the price.
2026: the lesson half-learnt
Which brings us to the present. Since the outbreak of the Middle East war, euro area energy prices have risen sharply again — energy inflation ran at 10.0% in the year to July 2026, against headline HICP of 2.9% (Eurostat flash estimate, 31 July 2026). And on 6 August the ECB published a box in its Economic Bulletin setting out what governments have done about it.
The headline finding is genuinely good news. Discretionary support since the outbreak of the war amounts to around 0.1% of GDP at the euro area level (Bouabdallah, Checherita-Westphal and Muggenthaler-Gerathewohl, ECB Economic Bulletin 5/2026). The Commission, measuring the EU with an earlier cut-off, puts it at €14.5 billion, or 0.07% of GDP. Against 1.2% of GDP in 2022, this is a contraction of well over ninety per cent. The ECB is even willing to credit it: the restraint “may partly reflect lessons from the past, including the high costs of the compensatory measures taken in 2022 and 2023.”
Now the other half. Of that much smaller sum, the ECB finds, “about 60% of the gross fiscal cost comprises (mainly untargeted) direct price measures”. The Commission, on its slightly earlier vintage, is harsher: more than two-thirds price measures, and three-quarters of the support spent in a non-targeted manner. The two institutions are measuring different aggregates on different cut-off dates, which accounts for the gap; neither of them is describing a well-targeted policy.
Set the two episodes side by side and the pattern is unmistakable. In 2022, three-quarters of a very large sum was untargeted. In 2026, three-quarters of a very small sum is untargeted. The level lesson has been learnt with impressive speed. The design lesson has not been learnt at all.
I want to be precise about what this does and does not show, because the obvious objection is a fair one and I shall come to it. But note first what it rules out. It rules out the story in which 2022 was an emergency, emergencies force crude instruments, and calmer conditions permit better ones. Conditions are calmer. The instruments are not better. Whatever produced the untargeted design in 2022, it was not the panic.
The case for the defence
It would be easy, and lazy, to stop there. The case for untargeted measures is stronger than economists’ conference-room consensus allows, and it comes in four parts, of which one is close to unanswerable.
The unanswerable part is administrative. Targeting requires knowing who is poor, quickly, and having a means of paying them. A number of member states in 2022 did not have that, and the fact that the IMF published an entire follow-up paper on “targeted, implementable and practical” second-best measures — bill rebates, block tariffs — is itself the concession. The OECD is explicit that “technical obstacles to implement a targeted approach and political economy constraints help explain the type of support countries provided” (Hemmerlé et al., 2023). A government that cannot find the poor cannot target the poor, and telling it to do so anyway is not advice.
The second defence is theoretically elegant and deserves more attention than it gets. Dworczak, Kominers, Akbarpour and Tokarski have argued that when a government is genuinely ignorant about who needs what, price controls can outperform transfers, because they screen: a threshold price cap that subsidises consumption up to a basic quota and charges the market rate above it separates necessity from luxury without the government needing to know anything about the household (GRAPE Working Paper 81). This is not a defence of cutting fuel duty. It is a defence of the German gas price brake specifically, and it is a good one.
Third, and awkwardly for my case, price caps may beat transfers on the particular objective of relieving energy poverty — because energy poverty is not confined to the poorest quintile, and income-targeted transfers therefore miss a great deal of it. That finding comes from a paper otherwise sympathetic to targeting, which makes it more credible rather than less (Amores et al., Energy Economics, 2025).
Fourth, voters prefer universal subsidies, partly because they misunderstand who benefits (Mus et al., Energy Research & Social Science, 2024). A minister who cuts fuel duty is not being stupid. She is being re-elected.
Why it still does not wash
And yet. Take the strongest version of the defence — that targeting is administratively hard — and ask what it predicts. It predicts that governments unable to target would spend rather more than the theoretical minimum, because crude instruments leak. What it does not predict is the actual arithmetic of 2026, which is this: the IMF estimates that fully compensating the poorest fifth of EU households for the entire energy shock would cost about 0.03% of GDP, rising to 0.15% under a severe scenario. The average member state is currently spending about 0.18% of GDP. Europe is spending roughly six times what full protection of the bottom quintile would require, and the bottom quintile is receiving €9.
That is not leakage. That is a different policy wearing the costume of this one.
Three further details, each of which the coverage has passed over. The first is that the inflation relief is borrowed rather than earned: the Eurosystem estimates the measures reduced year-on-year HICP by around 0.2 percentage points in the second quarter of 2026, “with a comparable increase in the second quarter of 2027” as the temporary measures expire. Nothing has been prevented. It has been moved.
The second is that the measures do not pay for themselves, and the ECB has now said so with data. The comforting theory that inflation swells indirect tax receipts turns out not to survive contact with the record: after large adverse oil supply shocks in the past, indirect tax ratios “developed very heterogeneously across euro area countries but there was no general upward trend at the euro area aggregate level”, and euro area VAT revenues from March to May 2026 “remained within their historical ranges”. Worse, excise duties are levied on quantities rather than values — so they do not rise with prices, and cutting them, as around seven in ten member states have done, removes revenue permanently while relieving prices temporarily.
The third is the fiscal position from which all this is being financed. The euro area’s projected structural deficit for 2026 is 3.3% of GDP, against the 2.6% projected before the 2022 shock, with the debt ratio close to 89% in both episodes and ten member states currently under an excessive deficit procedure. Europe entered the last energy crisis with a worse deficit than it thought and is entering this one with a worse deficit still — while committing to a defence and infrastructure build-out that was loosening the stance before the war began.
The sentence that has not changed
The ECB box closes by recommending that fiscal support be “temporary, targeted and tailored”. The IMF said the same thing in July 2022. The OECD said it in June 2022. The Commission says it in every forecast. That the advice has not changed in four years is sometimes read as a sign of analytical consensus. I read it as a sign of comprehensive failure: advice that is repeated verbatim across four years and two shocks is advice that nobody has taken.
The uncomfortable conclusion is that the restraint on display in 2026 is not a policy achievement. Governments are not spending 0.1% of GDP because they have finally understood incidence. They are spending 0.1% of GDP because, with structural deficits above 3% and ten of them in the corrective arm of the Stability and Growth Pact, they cannot spend 1.2%. The composition tells you what they would do if they could. Given a budget constraint, they economised on quantity; given a free hand on design, they changed nothing at all.
Which is, I think, the useful thing to know about European fiscal policy. It is not that ministers are unpersuaded by the economics. It is that the economics has never been the binding constraint. The budget was.
Sources
Amores, A.F. et al. (2023), “Inflation, fiscal policy and inequality”, ECB Occasional Paper No. 330. https://www.ecb.europa.eu/pub/pdf/scpops/ecb.op330~2e42ffb621.en.pdf
Amores, A.F., Christl, M., De Agostini, P., De Poli, S. and Maier, S. (2025), “Limiting prices or transferring money?”, Energy Economics 147, 108506. https://doi.org/10.1016/j.eneco.2025.108506
Arregui, N. et al. (2022), “Targeted, Implementable, and Practical Energy Relief Measures for Households in Europe”, IMF Working Paper 2022/262. https://doi.org/10.5089/9798400227400.001
Ari, A. et al. (2022), “Surging Energy Prices in Europe in the Aftermath of the War”, IMF Working Paper 2022/152. https://doi.org/10.5089/9798400214592.001
Bańkowski, K. et al. (2023), “Fiscal policy and high inflation”, ECB Economic Bulletin 2/2023. https://www.ecb.europa.eu/press/economic-bulletin/articles/2023/html/ecb.ebart202302_01~2bd46eff8f.en.html
Bobasu, A. and Dobrew, M. (2026), “Feeling the heat unevenly: energy prices and household consumption”, ECB Economic Bulletin 5/2026. https://www.ecb.europa.eu/press/economic-bulletin/focus/2026/html/ecb.ebbox202605_03~2626ace009.en.html
Bouabdallah, O., Checherita-Westphal, C. and Muggenthaler-Gerathewohl, P. (2026), “Assessing the scope for compensatory fiscal measures in response to the recent energy shock”, ECB Economic Bulletin 5/2026, 6 August 2026. https://www.ecb.europa.eu/press/economic-bulletin/focus/2026/html/ecb.ebbox202605_08~381440ea4e.en.html
Bruegel (2023), “National fiscal policy responses to the energy crisis”, dataset, version of 26 June 2023. https://doi.org/10.64153/LQHK8283
Celasun, O. (2026), “The 2026 Energy Shock: How to Deliver Targeted and Temporary Support while Encouraging Energy Conservation”, IMF European Department note, 4 May 2026. https://www.imf.org/-/media/files/countries/europe/imf-note-on-how-to-design-support-in-the-face-of-the-2026-energy-shock-europe.pdf
European Commission (2026), “Policy measures in EU Member States to address the 2026 energy price shock”, box in the Spring 2026 Economic Forecast, 21 May 2026. https://economy-finance.ec.europa.eu/economic-forecast-and-surveys/economic-forecasts/spring-2026-economic-forecast-slowdown-growth-energy-shock-drives-inflation/policy-measures-eu-member-states-address-2026-energy-price-shock_en
Eurostat (2026), “Euro area annual inflation up to 2.9%”, flash estimate for July 2026, 31 July 2026. https://ec.europa.eu/eurostat/web/products-euro-indicators/w/2-31072026-ap
Fetzer, T., Gazze, L. and Bishop, M. (2024), “Distributional and climate implications of policy responses to energy price shocks”, Economic Policy 39(120), 711–756. https://doi.org/10.1093/epolic/eiae038
Hemmerlé, Y. et al. (2023), “Aiming better: Government support for households and firms during the energy crisis”, OECD Economic Policy Papers No. 32. https://doi.org/10.1787/839e3ae1-en
Mus, M., de Rouilhan, S., Chevallier, C. and Mercier, H. (2024), “Energy subsidies versus cash transfers”, Energy Research & Social Science 118, 103836. https://doi.org/10.1016/j.erss.2024.103836
OECD (2022), “Why governments should target support amidst high energy prices”, 30 June 2022. https://doi.org/10.1787/40f44f78-en
Ruhnau, O., Stiewe, C., Muessel, J. and Hirth, L. (2023), “Natural gas savings in Germany during the 2022 energy crisis”, Nature Energy 8(6), 621–628. https://doi.org/10.1038/s41560-023-01260-5
Tokarski, F., Akbarpour, M., Kominers, S.D. and Dworczak, P. (2023), “A market-design response to the European energy crisis”, GRAPE Working Paper No. 81. https://ideas.repec.org/p/fme/wpaper/81.html
Weber, I.M., Beckmann, T. and Thie, J.-E. (2023), “The Tale of the German Gas Price Brake”, Intereconomics 58(1), 10–16. https://doi.org/10.2478/ie-2023-0004