Backup heat is where cold climate bills come from
A system that delivers 15% of its heat from resistance strips loses a quarter of its efficiency advantage.
By Ash Banerjee. Published 2026-08-31.
A heat pump moves heat rather than making it, which is why it can deliver three units of heat for one unit of electricity. Electric resistance heat makes heat directly and delivers exactly one unit for one unit. When a heat pump falls back on resistance strips, it stops being a heat pump for as long as they are running.
The question is how much that matters, and the answer is more than most people assume, because the averaging works against you.
The averaging is harmonic, not arithmetic
The instinct is to average efficiencies by weight. If 15% of your heat comes from strips at COP 1.0 and 85% from the compressor at COP 3.22, it is tempting to compute 0.15 times 1 plus 0.85 times 3.22 and get 2.89.
That is the wrong operation. You are not averaging efficiency, you are adding up electricity consumed to deliver a fixed amount of heat. The correct combination is:
where f is the share of heat delivered by resistance. With f at 0.15 and COP at 3.22, the effective COP is 2.35, not 2.89. The strips consume electricity out of proportion to the heat they deliver, so they pull the average down harder than their share suggests.
| Share of heat from backup | Effective COP | Efficiency lost |
|---|---|---|
| 0% | 3.22 | 0% |
| 5% | 2.83 | 12% |
| 10% | 2.53 | 21% |
| 15% | 2.35 | 27% |
| 25% | 1.97 | 39% |
| 40% | 1.60 | 50% |
Ten percent of heat from the strips costs you a fifth of your efficiency. Twenty five percent costs you close to two fifths.
What sets the share
The controlling variable is the balance point, which is the outdoor temperature at which the heat pump's output exactly matches the house's heat loss. Above it the compressor carries the whole load. Below it the shortfall has to come from somewhere, and in most installations that somewhere is a bank of resistance elements in the air handler.
Balance point is a property of the pairing between machine and house, not of the machine alone. The same heat pump in a well sealed house has a much lower balance point than in a leaky one. A unit sized for the cooling load, which is common practice in mixed climates, will have a higher balance point than one sized for heating.
Two other things push the share up. Many thermostats bring on strips during recovery from a setback, on the reasoning that the customer wants the temperature back quickly, so night setbacks in a heat pump house can cost more than they save. And defrost cycles run the strips to avoid blowing cold air into the room, which adds a share that is invisible on any equipment label.
Why HSPF2 does not save you here
HSPF2 is a seasonal figure and it does account for backup heat, which is a reasonable objection to everything above. The problem is which season and which house.
The rating is measured in one AHRI climate region against an assumed building load line. Your climate is not that climate and your house is not that house. A unit rated HSPF2 11.0 in the test region can deliver a materially lower seasonal COP in a colder one, and the gap is almost entirely backup heat.
Treating the label as though it described your installation is the same category of error as using a national average fuel price. It is a real number measured under conditions that are not yours.
What to do about it
Ask the installer for the balance point of the proposed system in your house, and for the expected share of annual heat from auxiliary. A contractor who has run a load calculation can answer both. One who cannot has not done the arithmetic, and the resulting bill is a coin toss.
If the share comes back above 15%, the options are a larger or better modulating unit, a tighter building envelope, or keeping the existing furnace as the backup instead of resistance strips. That last one is a dual fuel system, and in states with expensive electricity and cheap gas it is frequently the answer that the running cost arithmetic actually points to.
The calculator on this site exposes backup share as an input for exactly this reason. It is the assumption with the largest effect on the answer, and it is the one no state level dataset can tell you.
How this was checked
The harmonic combination is the part people get wrong, so I checked it at both limits before trusting it. Zero backup returns the heat pump COP unchanged. Full backup returns exactly 1.0. Both are what they should be.
For the real world share, I did not want to rely on my own intuition. Two Department of Energy funded field studies have instrumented heat pumps in occupied American homes and published measured auxiliary heat use. NREL found auxiliary energy exceeding 35% of compressor energy at 5 of 12 sites. PNNL, testing better equipment that had been sized by Manual J, found auxiliary heat to be a small fraction of delivered heat. I have written both up in full in the evidence note.
References
- Winkler, J. and Ramaraj, S. (2023). Field Validation of Air-Source Heat Pumps for Cold Climates. NREL/TP-5500-84745. National Renewable Energy Laboratory. May 2023.
- Mendon, V., Keene, K., Rosenberg, S., Rotondo, J. A., Nwe, K., Young, J., Wind, W., and Goetzler, B. (2025). Performance Results from DOE Cold Climate Heat Pump Challenge Field Validation. PNNL-37127. Pacific Northwest National Laboratory, prepared for the U.S. Department of Energy under Contract DE-AC05-76RL01830. January 2025.
- Air-Conditioning, Heating, and Refrigeration Institute. AHRI Standard 210/240-2023: Performance Rating of Unitary Air-Conditioning and Air-Source Heat Pump Equipment. This standard defines the M1 test procedure and the HSPF2 metric.
- Air Conditioning Contractors of America. Manual J: Residential Load Calculation, the standard method for sizing residential heating and cooling equipment.