Cost per saved kWh: how it is worked out in practice
The energy expert must enter the figure on Boverket's form for the energy declaration. But no rule says how to work it out. We took eight published measures that show both the investment and the cost per saved kWh, and worked backwards.
In brief
- Cost per saved kWh is what a saved kWh costs. Below the price of bought energy, the measure pays off.
- The rules do not say how the figure is worked out. Each energy expert chooses.
- Eight published examples fit one method: the cost is spread evenly over the years, at about 3% interest.
- Always write your assumptions beside the figure. Otherwise nobody can tell whether it is right.
On this page
In the declaration's Rekommendationer om kostnadseffektiva åtgärder section, each measure carries two fields: Minskad energianvändning in kWh per year and Kostnad per sparad kWh in kr/kWh. The first is a physical estimate. The second is an economic calculation — and that is where the guidance runs out.
What the regulation says: nothing
We searched BFS 2026:10 (BED 13) in full. The words kalkylränta, livslängd, nuvärde, annuitet, lönsamhet and besparingskostnad do not appear once. The regulation requires only that the inspection be carried out so that "anpassade och kostnadseffektiva rekommendationer om åtgärder kan lämnas".
Boverket's declaration handbook has no page on measure economics at all, and Boverket's own position is that this falls within "den certifierade energiexpertens ansvar och kompetens". BeBo puts it bluntly:
"There are no sources that state which discount rate you should use."
BeBo · Vägledning – Lönsamhet och kostnader — our translation
The method is therefore left entirely to whoever writes the declaration. And the register stores only the two figures and a short description — never the assumptions behind them.
How we worked backwards
If the method is written down nowhere, it can be read out of the results. We found eight published measures — in declarations and åtgärdsrapporter — where both the investment and the kr/kWh appear in the same document, and tested two candidate formulas against each pair.
- The simple divide: the investment divided by the total saving over its life, capex ÷ (kWh/yr × n).
- The annuity method: the investment converted to an annual cost with an annuity factor, divided by the annual saving.
The result is unambiguous. The simple divide is wrong by 15–47 % in all eight cases. The annuity method matches all eight — and the rate that falls out of each individual case clusters remarkably tightly.
där annuitetsfaktor(r, n) = r ÷ (1 − (1+r)−n)
r ≈ 0,030
n = the measure's economic life
One case, worked through
A 5.4 kW solar array in a villa åtgärdsrapport: 4,800 kWh/yr, investment 70,000 kr, economic life 25 years. The report publishes 0.84 kr/kWh.
0,057428 × 70 000 kr = 4 020 kr/år
4 020 ÷ 4 800 kWh = 0,8375 → 0,84 kr/kWh ✓
The simple divide would have given 0.58 kr/kWh — 31 % out.
Eight cases, three independent producers
| Measure | kWh/yr | Investment | n | Simple | Annuity | Published |
|---|---|---|---|---|---|---|
| Solar 5.4 kW, villa | 4 800 | 70 000 | 25 | 0,58 | 0,84 | 0,84 |
| New district-heating substation | 5 200 | 200 000 | 25 | 1,54 | 2,21 | 2,21 |
| Ground-source heat pump | 109 000 | 950 000 | 25 | 0,35 | 0,50 | 0,51 |
| Heat-recovery ventilation unit | 16 100 | 300 000 | 25 | 0,75 | 1,07 | 1,07 |
| New substation (2) | 14 200 | 400 000 | 25 | 1,13 | 1,62 | 1,62 |
| Added attic insulation | 75 980 | 200 000 | 50 | 0,05 | 0,10 | 0,10 |
| Individual hot-water metering | 27 905 | 168 000 | 10 | 0,60 | 0,71 | 0,71 |
| Weather-forecast heating control | 66 953 | 111 165 | 15 | 0,11 | 0,14 | 0,14 |
The rate: 3.0 %, and why
Solving for the rate in each published pair gives: 2.88 · 3.01 · 3.17 · 3.00 · 3.01 · 2.88 · 3.12 · 3.09 %. Eight measures, three independent producers, the same figure.
In all likelihood 3.0 % is not the rate anyone enters, but the one that survives. One of the declarations states "4 % realkalkylränta och en årlig energiprisökning med 1 %" — and 1.04 ÷ 1.01 − 1 = 2.97 %. That is a plausible explanation which reproduces the figures, but we found no document prescribing that particular combination. Treat it as a reading, not a rule.
Service lives
Here a published table does exist: Belok, Totalmetodiken handbok ver 1.6, annex 3, which Energimyndigheten cites in its life-cycle cost manual. In summary: 40 years for building-fabric measures, 15–20 for building-services measures.
| Measure | Belok | CEN 15459 | 2006/32/EG |
|---|---|---|---|
| Facade, roof and ground insulation | 40 | — | 25–30 |
| Energy-efficient windows | 40 | — | 30 |
| Improved airtightness | 40 | — | 5 |
| FTX | 20 | 15–20 | 17–20 |
| Solar thermal | 20 | 15–25 | 20 |
| Solar PV | 20 | — | 23 |
| Demand-controlled ventilation | 15 | 15 | 15 |
| Exhaust-air heat pump | 15 | 15–20 | 15 |
| Improved heating control | 15 | 15–25 | 10 |
VAT and the ROT deduction
ROT is not deducted in the practice we were able to examine. One report says so outright: "The investment for a specific energy measure can be reduced further if grants are available, which we have not taken into account in our calculations" (our translation). Published kr/kWh figures are therefore gross of the tax deduction. ROT in 2026 is 30 % of the labour cost, capped at 50,000 kr per person per year.
VAT is worse documented. The only explicit statement we found was "kostnaderna är angivna exklusive moms" — but in a housing-association report, where quoting ex-VAT is the professional norm. For detached houses, where the client is a consumer and prices are normally quoted inclusive, we found no convention at all. That is an assumption to be declared, not a constant.
What this means
The same physical measure on the same house can be published at 0.58 or 0.84 kr/kWh — a 45 % difference — without either figure being wrong, and without any reader being able to see why they differ. The register stores two outputs and no inputs.
This is not a failing of the experts. It is what happens when a figure is demanded without a method being specified. The conclusion is simple: state your assumptions beside the figure. And expect the client to ask for the payback period anyway — that is the number spoken on site, whatever the form wants. Rate, service life, VAT and subsidy treatment. Four lines, and they are the difference between a number that can be checked and one that can only be believed.
Sources and caveats
Sources and caveats. BFS 2026:10 (BED 13) searched in full · BeBo, Vägledning – Lönsamhet och kostnader, 31 Jan 2024 · Belok, Totalmetodiken handbok ver 1.6, annex 3 · Energimyndigheten's life-cycle cost manual · Skatteverket on the 2026 ROT deduction · eight published energy declarations and åtgärdsrapporter. The eight cases are those we could find where both figures appear in the same document — a small, self-selected set, not a statistical sample. The 3.0 % rate is derived from them, not taken from any regulation. The split into 4 % real rate and 1 % energy-price escalation is our own reading.