A space heater is 100% efficient, and that is the problem

Search this and you get two answers, both stated confidently, by different pages. "Central heating is cheaper." "A space heater is cheaper if you only heat one room." Neither side shows the arithmetic, so there is no way to tell which one describes your house.

The arithmetic is not hard. It needs four published numbers and one fact from physics, and it produces a break-even you can check against your own situation.

The fact from physics, which settles half the question

An electric space heater is exactly 100% efficient, and so is every other one.

Every watt-hour of electricity that goes into a resistance heater leaves it as heat in the room. Not 95%, not 98% — all of it, because there is nowhere else for the energy to go. The fan, the indicator light and the losses in the cord all end up as heat in the same room.

This has two consequences that most of the pages on this question get wrong.

A 1500 W heater is a 1500 W heater. Ceramic, oil-filled, infrared, "energy-saving" — at the same setting they deliver the same heat and cost the same to run. A heater can differ in how quickly it warms you, where it points the heat, and whether its thermostat cycles sensibly. It cannot differ in efficiency, because there is no room above 100%.

And the ceiling is also the trap. 100% sounds like a perfect score. It is the worst score in home heating, because the alternatives are not limited to 100%: a heat pump moves heat rather than making it, and routinely delivers two or three units of heat per unit of electricity.

The four published numbers

Comparing heating costs means comparing the price of one kilowatt-hour of heat delivered into the house, not the price of one unit of fuel. Two conversions get us there.

For gas, the EIA publishes a residential price per thousand cubic feet, and separately states the heat content of the gas delivered to consumers: about 1,037 Btu per cubic foot, so a thousand cubic feet is 10.37 therms. A therm is 100,000 Btu, and a kilowatt-hour is 3,412.14 Btu, so a therm is 29.30 kWh.

For electricity, no conversion is needed, which is the whole point.

Both figures below are the US residential average for February 2026, published by the EIA. Why that month rather than the newest one is the next section, and it matters more than it sounds.

So per unit of heat, the space heater costs 3.4 times a modern gas furnace, 2.9 times an old one, and 2.2 to 3.0 times a heat pump.

One result there was not obvious to us before we did the arithmetic: at these prices a minimum-efficiency heat pump (8.02 ¢) and a high-efficiency gas furnace (5.18 ¢) are not close, but a heat pump at COP 3.0 (5.88 ¢) and an 80% furnace (6.15 ¢) very nearly are. The ranking depends on numbers that move every month, which is why we publish the month next to them.

The month is not a detail

Look at what the EIA's residential gas price did through the first half of 2026:

Gas did not become 73% more expensive between January and June. What changed is the divisor. That published price is total residential revenue divided by total residential volume, and it therefore carries the fixed monthly customer charge every household pays whether it burns anything or not. In January, that charge is spread over a winter's worth of gas. In June it is spread over a water heater and a stove.

This is our reading of why the series moves, not a statement the EIA makes on that page — but it is the reason a heating comparison must not use a June price. Do it and the space heater looks about 40% better than it is: against the June figure the ratio falls from 3.4× to 2.1×, and nothing about anybody's house has changed.

Almost every page comparing these two costs quotes whichever gas figure was newest. In summer, that is the wrong one.

So when does a space heater win?

It never wins per unit of heat. It wins by delivering less heat, and now the break-even is a number rather than an opinion.

At 3.4 times the cost per delivered kilowatt-hour, a space heater is cheaper only if it lets you deliver less than 29% of the heat the central system would have delivered over the same period. Against an old 80% furnace the threshold is 35%; against a good heat pump it is 33%.

That is a demanding target, and here is the part the "just heat one room" advice skips: heating one room out of five does not cut delivered heat to a fifth.

• You cannot let the rest of the house go cold. Pipes, and the fact that you will walk through it.
• Heat leaves the warm room into the cold ones, not just to the outdoors. The colder you let the rest of the house get, the faster your one warm room loses heat into it.
• A central system reaching a lower setpoint everywhere still runs, so its cost does not fall to zero — the saving is on the difference, not on the whole bill.

Which is why the honest answer has a shape rather than a value: a space heater wins in a small number of specific situations and loses in the general case. One person in one room, for a few hours, in a house whose thermostat genuinely drops for that whole period — plausible. A heater in the living room while the house stays at its usual temperature — that is added cost with no offset at all, and it is the most common way people actually use them.

What to check in your own case

The averages above are wrong for you specifically, in both directions and by a lot. State electricity ran from around 14 to over 52 cents in the same month. Do it with your own numbers:

1. Your electricity rate, from your bill, in cents per kWh. That is the cost of a delivered kWh from any electric resistance heater, full stop.
2. Your gas rate, from your bill. If it is per therm, divide by 29.30 for cents per kWh of gas energy; if it is per Mcf, divide by 10.37 first. Then divide by your furnace's AFUE.
3. Your heat pump's HSPF2, from the label, divided by 3.412 — that is its seasonal COP. Divide your electricity rate by it.
4. Divide. The bigger number over the smaller one is how much less heat the space heater has to deliver before it saves you anything.

If step 4 gives you 3, the space heater has to replace two-thirds of your heating, not supplement it. Most of the time it does not, and the bill says so in January.