Solar Payback Period Without the 30% Credit: Redo the Math

Berkeley Lab published its national update on distributed solar prices in August 2026. It covers roughly 5.3 million systems installed through the end of 2025 (U.S. Distributed Solar and Storage Data, Electricity Markets & Policy, read 27 September 2026). And 31 December 2025 is the last day the 30% federal residential credit existed.

So the most careful price dataset in the country contains exactly zero systems bought under the rules you are buying under. There is no published national median for a post-credit install yet, because nobody has finished collecting one.

Which means the arithmetic cannot be looked up. It has to be rebuilt from your own numbers, and the rebuild is not a matter of deleting one line. The credit sat inside the term that every other term in the formula gets measured against, and taking it out changes how much each remaining assumption is worth. This article is the input list, the formula, and one roof carried all the way to a year. What ended and on what legal terms is a separate question, answered in the article on what §25D actually said.

The term that vanished, and the one line left holding the offset

Before 2026, the net cost line in a residential payback calculation looked like this:

Net cost = Gross installed price x (1 - 0.30) - state and utility incentives

Now it looks like this:

Net cost = Gross installed price - state and utility incentives

One factor removed. The consequence is not that the calculation got one line shorter — it is that the entire offset now comes from whatever your state and your utility happen to run, and those programs are shaped completely differently from the credit they are replacing. The federal credit was a percentage with no dollar cap for solar electric property. State credits are typically a percentage with a cap. Utility and state capacity rebates are usually a fixed amount per watt that steps down as a program's blocks fill.

That difference in shape matters more than the difference in size. A percentage scales with the quote; a cap does not. Ask for a bigger system and a capped state credit covers a smaller share of it.

Nine inputs, and the document each one has to come out of

Every figure below is something you can find. None of them is a national average, because a national average has never described anyone's roof.

1. Gross installed price, before any incentive. This is the basis Berkeley Lab uses — the up-front price paid by the system owner, prior to any incentives, in dollars per watt DC. Take your quote's total, divide by DC kilowatts, and you have a figure comparable to a published distribution. Median prices across the top-100 residential installers in 2025 ran from $2.4 to $6.3 per watt once apparent outliers at either end are set aside, with roughly 60% below $4/W, almost 25% below $3/W, and about 10% above $5/W (2026 Data Update, Berkeley Lab, August 2026, read 27 September 2026). Within one installer, the 20th-to-80th-percentile band of project prices typically spanned $1 to $2/W, so being above a median is not by itself evidence of anything.

2. System size in kW DC. The median U.S. residential system was 7.7 kW in 2025, up 0.5 kW from 2024, with most systems between roughly 5 and 12 kW. Useful mainly as a sanity check on the divisor in input 1.

3. State income tax credit, after its own cap and its own rules. New York's is 25% of qualified solar energy system equipment expenditures, limited to $5,000, for a system installed and used at your principal residence in the state. It is not refundable, though unused credit carries over for up to five years (NYS Department of Taxation and Finance, Solar Energy System Equipment Credit, read 27 September 2026). Read that last part twice before you subtract anything. A nonrefundable credit is worth its face value only if you have the state tax liability to absorb it, this year or inside the carryforward window.

4. Capacity rebate in dollars per watt. NY-Sun assigns incentives to regions, splits each region into megawatt blocks, and lets the per-watt amount fall as blocks subscribe — incentives "remain available until all blocks within a region/sector are fully subscribed" (NYSERDA, Dashboards and Incentives, read 27 September 2026). The number is therefore not a constant and cannot honestly be quoted in an article. Read it off the live dashboard for your region and sector on the day you run your sheet, and see how to check a rebate still has funding behind it before it earns a row.

5. Sales tax and property tax treatment. Neither of these is a rebate. A sales tax exemption changes input 1. A property tax exemption or assessment freeze changes nothing in the payback calculation at all — but it does change what happens if the exemption sunsets before you sell.

6. First-year AC production in kWh. From your own modelling run, not the proposal's. The method is in checking a quote's production estimate against PVWatts. Two arrays of identical nameplate on the same street can differ by a quarter of their output.

7. Value of one produced kWh. This is two numbers pretending to be one: the marginal rate you avoid on a self-consumed kWh, and whatever your tariff pays for an exported one. Pulling the first off your own bill is covered in the four numbers on your electric bill. The second depends entirely on which compensation structure your state uses, and the two are rarely equal.

8. Fixed charges. List them, then set them aside permanently. A basic service charge, a minimum bill, and any demand component are not reduced by generation. They are also the reason an average revenue-per-kWh figure overstates what an array saves.

9. Ongoing cost. Monitoring subscription if the proposal carries one, cleaning if your site needs it, and one inverter replacement somewhere in the middle of the term. Whether this input changes your answer depends on which question you asked, which is worth a paragraph of its own further down.

One formula, and the slot people put the credit in wrongly

Net cost         = Gross price
                   - state income tax credit (after cap, and only if usable)
                   - capacity rebate ($/W x watts)
                   - any other cash incentive actually received

Year-1 saving    = Annual kWh x blended value per kWh

Growth factor g  = (1 + rate escalation) x (1 - annual degradation) - 1

Cumulative(N)    = Year-1 saving x ((1 + g)^N - 1) / g

Payback year N   = ln(1 + Net cost x g / Year-1 saving) / ln(1 + g)

Two things go wrong at the top of that stack more often than anywhere else in it.

The first is order of operations. A percentage credit and a fixed rebate do not commute. If a state credit is 25% of expenditures, it is 25% of what you spent — not 25% of what you spent after a utility handed you a cheque. Some programs define their base the other way round, and the program's own text is the only place that settles it. Run it both ways; if the two answers differ by less than a rounding error on your roof, stop worrying about it.

The second is timing. A tax credit arrives when you file, not when the panels go up, and a nonrefundable one may arrive in pieces across five years. Subtracting it from net cost on day one is a convenience, not a fact. If your answer lands near a threshold you care about, put the credit in the year you actually expect to receive it and see whether the answer survives the move.

A 7.7 kW roof in New York, carried to a year

Two of the figures below are stated assumptions rather than measurements of any real roof: the specific yield and the blended value per kWh. Both are marked, and both are chosen to be round enough to follow by hand. Every other figure traces to a source above. The yield in particular is not a modelling result for any address — it is a plausible mid-range Northeast number standing in until you run your own array's orientation, tilt and shade, which is the whole reason input 6 exists.

System size                7.7 kW DC            (LBNL median, 2025 installs)
Installed price            $3.60 / W DC         (inside the 60%-below-$4/W group)
Gross price                $27,720
State credit               25% x 27,720 = $6,930, capped at   -$5,000
NY-Sun block incentive     $0 in this run       (read your own dashboard)
Net cost                   $22,720

Specific yield             1,150 kWh per kW DC  <- assumption; use your PVWatts run
Annual production          8,855 kWh
Blended value per kWh      24.0 cents           <- assumption; use your own tariff
Year-1 saving              $2,125

Rate escalation            3.0% / year
Module degradation         0.5% / year
Growth factor g            (1.030 x 0.995) - 1 = 2.485% / year

On that blended value: EIA puts the average price of residential electricity in New York at 29.90 cents per kWh for July 2026, against 26.22 cents in July 2025 (Electric Power Monthly, Table 5.6.A, data for July 2026, released 24 September 2026, read 27 September 2026). The 24.0 cents used here sits below that deliberately, because the state average includes fixed charges that no avoided kWh removes. Do not read the 14% gap between those two July figures as an escalation rate either — it is one month against the same month a year earlier, and turning rate history into an escalation assumption is its own exercise, handled in the three dials that shorten payback on paper. The national residential figure moved 17.45 to 18.31 cents across the same two months, 4.93%.

Now the cumulative column, which is the only part that answers the question:

Cumulative savings pass the $22,720 net cost during year 10. The closed form puts it at 9.60 years.

Run the identical roof with the old 30% credit still in force and net cost is $27,720 − $8,316 − $5,000 = $14,404, cleared during year 7. 6.34 years.

Two sensitivities worth writing in beside them, since they cost nothing once the formula is in a cell. Every $0.10/W of capacity rebate is $770 on a 7.7 kW system, which pulls the answer in by about 0.29 years. And one more cent on the blended value takes off 0.34 years — while one cent less adds 0.37. The two are not equal, and that asymmetry is the whole reason to test a value you are unsure of downward rather than upward.

Why a 30% credit was worth 51% of the answer

Net cost went from $14,404 to $22,720, a rise of 57.7%. The payback year went from 6.34 to 9.60, a rise of 51.4%.

Neither of those is 30%, and the reason is the thing worth carrying away from this page. The credit was 30% of the gross price. But payback measures the time to clear net cost, and $8,316 was 57.7% of the $14,404 you were actually trying to earn back. A credit's headline percentage is quoted against the largest number in the calculation and then applied to the smallest one.

The capped state credit makes that sharper rather than softer. Strip New York's $5,000 out of the example and the federal credit would have been 42.9% of net cost instead of 57.7%. A fixed-dollar incentive shrinks the base, so whatever percentage credit sits beside it covers a larger share of what is left. The same mechanism is now running in reverse: with the percentage gone, that capped $5,000 is covering 18% of a $27,720 gross price, and it will cover a smaller fraction of every larger system you get quoted.

There is a second effect, and it is the one that should change how you read a proposal. Move the escalation assumption from 0% to 5% on the old, credit-bearing version of this roof and the payback year swings 0.83 years. Do the same on the 2026 version and it swings 2.03 years — the same dial, 2.4 times the leverage. The credit used to absorb sloppy assumptions. Now every soft input in the proposal moves the answer more than twice as far as it did last year, and the softest of them is usually the one on page four in small type.

Input 9 shows the opposite face of that sensitivity. An inverter replacement in year 12 does not touch a 9.60-year payback figure at all, because it falls outside the window — but a replacement quoted at $1,800 comes straight off a 25-year cumulative saving of $72,451. So one input is irrelevant or decisive depending on whether you asked "when am I even" or "what is this worth over its life." A proposal that answers only the first is not wrong so much as half a document.

Three ways the old answer gets quietly restored

The template. Look down the proposal for a row that subtracts roughly 30%, whatever it is labelled: federal credit, ITC, tax incentive, or sometimes an unexplained figure sitting in the net-cost stack. Then ask for the projected completion or permission-to-operate date in writing — not the contract date — and check it against the December 2025 cutoff.

The source you would have trusted. I pulled the IRS Residential Clean Energy Credit page on 27 September 2026 expecting the pre-expiry language to be gone from it. It is not, and the two halves are closer together than you would expect. Inside the single section headed "How it works", the first paragraph says the credit "is not available for any property placed in service after December 31, 2025" — and two paragraphs below it, under the same heading, the page says "You can claim the annual credit every year that you install eligible property until the credit begins to phase out in 2033" (IRS, read 27 September 2026). The statute has no operative 2033 language left in it. §25D(h), headed Termination, now reads in full that the credit "shall not apply with respect to any expenditures made after December 31, 2025"; the 2033 and 2034 step-downs were struck by Pub. L. 119–21 §70506 on 4 July 2025 and survive only in the amendment notes underneath (26 U.S.C. §25D, via Cornell Law School's Legal Information Institute, read 27 September 2026). A page can be accurate and stale two paragraphs apart, and every calculator, summary and sales deck downstream of it inherits whichever half it happened to read. When one number decides a five-figure purchase, the statute is the cheaper thing to open.

An incentive counted before it was confirmed. A per-watt block that has closed in your region, or a rebate whose appropriation is spent, is a row of zeros that looks like a row of dollars. Confirm the program is open, for your sector, in your region, on the day you run the sheet.

One boundary is worth stating flatly. All of the above is the arithmetic of a system you own. If the contract in front of you is a lease or a power purchase agreement then you are not the owner of the equipment, §25D was never the section describing your position, and the comparison works differently from top to bottom — that one is laid out in cash against loan, lease and PPA on total cost.

The sheet to build before you answer the quote

Open a spreadsheet with one column for the proposal's assumption and one for yours, and fill nine rows: gross price, $/W, kW DC, state credit after cap, capacity rebate, first-year kWh, blended value per kWh, escalation, degradation. Wherever the two columns disagree, the gap in years is the size of the disagreement, and you now have the closed form to price it.

Then request two dates in writing. The projected permission-to-operate date settles whether any federal line belongs on the page at all. The price expiry date tells you how long your own column has to be finished.

And run the sheet three times — at your escalation figure, at zero, and at the proposal's. If those three answers straddle a year that matters to you, the proposal has not given you the payback period. It has given you one of them.