Heat Pump or Gas Boiler: the SCOP Your Local Prices Demand

One formula, and it is not the one on the brochure

A gas boiler and a heat pump both sell you the same thing: a kilowatt-hour of heat inside the house. The boiler buys it by burning gas at some efficiency; the pump buys it by moving heat with electricity at some seasonal coefficient of performance. Put the two costs side by side and everything cancels except one comparison: the pump is cheaper to run when its SCOP is greater than the electricity price divided by the gas price, times the boiler's efficiency. That single inequality is the whole decision on running cost, and the two prices in it are published by Eurostat twice a year for every member state.

The prices used throughout this article are Eurostat's household bands for the second half of 2025, all taxes and levies included: electricity for a household consuming 2,500 to 4,999 kilowatt-hours a year, and gas for one consuming 20 to 199 gigajoules. Per kilowatt-hour that is 24.35 cents of electricity and 14.05 of gas in Portugal, 25.61 and 14.36 in France, 29.66 and 14.81 in Italy, 28.96 and 12.28 across the EU, 26.69 and 9.55 in Spain, and 38.69 and 12.23 in Germany. Spain has the cheapest gas in this group and Germany the most expensive electricity, and between them they produce the two hardest cases for a heat pump in western Europe.

The gas meter flatters the boiler by exactly ten per cent

A condensing boiler is often sold as "98 % efficient", and that figure is real but it is measured against the net calorific value of the gas — the energy released without recovering the latent heat of the water vapour in the flue. Your meter does not bill you in net calorific value. Gas is metered and priced on gross calorific value, which includes that latent heat, and Eurostat's own calorific-value table settles the ratio precisely: for natural gas in inland consumption in 2023, the net value is exactly 0.900 times the gross in both the French and the German rows. So a boiler rated 98 % on net delivers 88.2 % of every kilowatt-hour you are actually billed for.

Forgetting this is the commonest arithmetic error in the whole comparison, and it always runs the same way: it makes the boiler look about a tenth better than it is, and therefore raises the SCOP the heat pump has to reach by about a tenth. In Germany that is the difference between a break-even SCOP of 2.79 and one of 3.10 — the difference between a machine that might just work on the existing radiators and one that certainly will not. Every figure in this article uses 88.2 % for the boiler, which is a well-set condensing unit; an older non-condensing boiler in a real installation is closer to 70 %, and that shifts every comparison decisively toward the heat pump.

Same machine, same house, six different answers

Take a house that needs 12,000 kilowatt-hours of heat a year — a figure your installer computes from the building, not from a table — and put the same air-source heat pump in it six times. At a seasonal coefficient of 3.5, it saves about 1,077 euros a year in Portugal, 1,076 in France, 998 in Italy, 678 at the EU average, 384 in Spain and 337 in Germany. Nothing about the equipment differs. Nothing about the house differs. The whole spread, a factor of more than three between the best and worst case, is the local ratio of two published prices.

Now drop the seasonal coefficient to 2.5, which is roughly what an air-source pump manages when it has to feed radiators sized for 70 degrees rather than underfloor pipes at 35. Portugal still saves about 743 euros, France about 724 and Italy about 591. The EU average falls to about 281. Spain lands at about 18 euros a year, which after rounding is the same as no saving at all. And Germany goes negative: about 193 euros a year worse than the condensing boiler it replaced. This is the case the brochures never model, and it is exactly the case a retrofit in an old building is most likely to land in.

Why the grant comes second, not first

A subsidy reduces the capital you have to recover; it does not touch the annual saving out of which you recover it. Express the payback per 1,000 euros of extra capital over a like-for-like boiler replacement and the point becomes unavoidable. At a seasonal coefficient of 3.5 on that 12,000-kilowatt-hour house, each 1,000 euros takes 0.93 years to repay in France, 0.93 in Portugal, 1.00 in Italy, 1.48 at the EU average, 2.60 in Spain and 2.96 in Germany. A grant of 5,000 euros therefore buys back about five years in France and about fifteen in Germany — the same grant, the same machine, three times the effect.

And where the annual saving is negative, the payback is not long — it does not exist, and no grant of any size creates one. That is the German high-temperature-radiator case, and it explains why the serious question at a survey is not "what am I entitled to" but "what flow temperature can this building be heated at". Meanwhile the policy side has already moved: the recast energy performance of buildings directive, Directive (EU) 2024/1275 of 24 April 2024, tells member states not to give financial incentives for stand-alone fossil-fuel boilers from 2025, with narrow exceptions. The comparison is therefore increasingly between a heat pump and a boiler you buy unaided.

What would change this verdict

Three things, and only one of them is about the machine. First, the price ratio is a policy variable, not a physical constant: taxes and levies were 28.9 % of the EU household electricity price in the second half of 2025, up from 27.9 % six months earlier, and any government that moves a levy off electricity and onto gas moves every break-even SCOP in this article. Second, the SCOP on a datasheet is measured under the conditions of the ecodesign and labelling regulations for space heaters, Regulations (EU) 813/2013 and 811/2013, at a stated climate and flow temperature; the number your installation achieves depends on your climate and your emitters, and the gap between the two is where most disappointed owners live. Third, a heat pump replaced with an old non-condensing boiler rather than a good condensing one starts from a boiler efficiency near 70 % instead of 88 %, which pulls every break-even SCOP down by about a fifth and flips the German and Spanish cases outright.

Frequently asked questions

Where do I find my building's annual heat demand?

The quickest honest answer is on last year's gas bill. Take the kilowatt-hours billed, subtract what you use for cooking and hot water if that runs on the same meter, and multiply by the boiler's efficiency on the billed kilowatt-hour — 88 % for a well-set condensing unit, nearer 70 % for an older one. That gives you the heat the building actually consumed, which is a measurement rather than an estimate. A heat-loss calculation from the installer gives the peak load, which is what sizes the machine, and the two are different numbers used for different purposes: never let a peak load be presented to you as an annual consumption.

Is a SCOP of 2.5 really what old radiators give?

It is a realistic figure for an air-source pump asked to deliver water at 65 to 70 degrees through radiators sized for a boiler, in a climate with real winters. The physics is unforgiving: the efficiency of a heat pump falls as the difference between the source temperature and the delivery temperature grows, so every degree of flow temperature you can remove is efficiency you get back. That is why the useful survey question is what the emitters can be persuaded to work at, and why oversizing the radiators or lowering the demand often does more for the SCOP than a more expensive machine does. Treat 2.5 as the pessimistic case to test against, not as a prediction.

Does a lower electricity tariff at night change the answer?

It can, and it is one of the few levers a household controls. The prices in this article are Eurostat's averages across all household contracts in a band, so a time-of-use tariff that shifts a meaningful share of heating consumption into cheap hours lowers the effective electricity price and therefore the SCOP the pump has to reach. Two conditions decide whether it actually helps: the building must be able to store the heat it produces at night, which favours heavy construction and underfloor systems and disfavours a lightweight house with radiators; and the tariff spread must be wide enough to matter after standing charges. Run the arithmetic with your own effective average price per kilowatt-hour rather than the headline tariff.

Why compare against a condensing boiler rather than the old one I actually have?

Because the honest counterfactual is not keeping the old boiler forever — it is what you would buy if you were not buying a heat pump. If your boiler is at the end of its life, the alternative purchase is a new condensing unit, and the extra capital to be recovered is the difference between the two quotes, not the whole price of the heat pump. Comparing a new heat pump against a twenty-year-old boiler's running cost while charging the heat pump's full price against zero is the single most common way these calculations are made to come out favourably. Use the difference in quotes and the difference in running cost, and the answer stops flattering anybody.

Do these figures include hot water and cooling?

No, and both omissions matter in opposite directions. Domestic hot water needs a higher delivery temperature than space heating, so it drags the annual average coefficient down — a system that heats rooms at 40 degrees and water at 55 does not achieve its space-heating SCOP across the year. Cooling, on the other hand, is a service the gas boiler cannot provide at all, so where summers are hot a reversible heat pump is displacing an air conditioner as well as a boiler, and the running cost it should be charged with is the sum of both. The arithmetic in this article is space heating only, deliberately, because it is the one comparison where a like-for-like number exists.