How Fast Do Solar Panels Degrade Over Time?

Solar quotes tend to carry a single degradation number, usually half a percent a year, presented as if it were a property of physics. It is closer to a summary statistic from a large and uneven body of field research, and the distribution behind it matters as much as the headline.

How fast do solar panels degrade?

The reference work is NREL's Compendium of Photovoltaic Degradation Rates, published in Progress in Photovoltaics in 2016. It aggregated more than 11,000 degradation rates from roughly 200 studies across 40 countries. For crystalline silicon it reports a median in the 0.5% to 0.6% per year range and a mean in the 0.8% to 0.9% range. HIT and microcrystalline silicon land nearer 1% a year, behaving more like thin-film products.

That gap between median and mean is the part usually left out. A mean well above the median means the distribution has a long tail: most systems cluster near the median, and a minority degrade considerably faster and drag the average up. Quoting 0.5% as the expected outcome is defensible; quoting it as a ceiling is not.

The earlier Jordan and Kurtz analytical review, published in 2013 from around 2,000 rates spanning 40 years of field testing, put the median at 0.5% a year and remains the source most consumer material is echoing. Both works aggregate published studies rather than sampling American rooftops at random, and their authors flag bias related to system size and measurement accuracy. Separately, NREL reported in 2017 that systems installed after 2000 showed fewer failures and different degradation modes than older ones, so an older median may understate how modern hardware behaves.

What a degradation rate means in practice

Two methods circulate, and they do not give the same answer. Keep them apart.

• Linear remaining output = 1 − (degradation rate × years). Simple, and the form many warranty schedules resemble.
• Compounded remaining output = (1 − degradation rate)^years. Each year's loss applies to what is left, not to the original figure.
• Future annual production = first-year production × remaining output factor.
• Annual energy loss = first-year production − future annual production. Annual value = future annual production × value per kWh.

At the same rate, the compounded method always leaves more than the linear one, because the percentage bites into a shrinking base. Over 25 years at 0.6% the difference is about one percentage point; at 1.0% it widens to nearly three. Throughout this guide the factor (1 − r)^N describes output once N full years have passed, so the first year carries no degradation at all.

Output after 10, 20 and 25 years

What degradation does to solar savings

Performance warranty is not production

Two separate documents usually come with a module, and conflating them causes most of the disappointment in this area.

A performance warranty promises that a module, tested a particular way, will still make a stated fraction of its rated power at a given year. It says nothing about how many kilowatt-hours your array delivers, because that depends on weather, shading, soiling, availability and the inverter — none of which the module maker controls. Claims typically require proof of ownership, serial numbers, measurement by the specified method, and sometimes inspection by an approved installer, and the costs of removing, shipping and reinstalling a module are often not included.

NREL frames the end of the term plainly: modules are typically warrantied for 20 to 25 years, after which they can still produce electricity but the output is no longer guaranteed, and there are no end-dates as such for productive life. Well-built modules have demonstrated performance for 30 years outdoors.

Why production can fall faster than the panels age

Slow annual degradation cannot produce a step change. If output drops noticeably between one month and the equivalent month a year earlier, the cause is almost certainly something else.

Weather is the reason a single weak month proves nothing. Irradiance varies year to year, high cell temperatures reduce output, and smoke, snow and cloud all cut production without anything being wrong. Comparing raw totals from two different years mixes those effects with degradation. Weather normalization — adjusting measured output for the sunlight and temperature actually experienced — is what separates the two, and it needs several years of data before a trend of a few tenths of a percent is visible at all.

How to tell normal aging from a real problem

1. Compare like with like: the same month against the same month in earlier years, not against last month.
2. Normalize for weather as far as your data allows, or at least note unusual cloud, smoke, snow or heat before drawing conclusions.
3. Read the monitoring alerts. Most step changes are already flagged there.
4. Compare strings or microinverters against each other. One area behaving differently is diagnostic.
5. Look from the ground for new shade — a grown tree, a neighbor's addition, a new antenna — around midday.
6. Call the installer if the drop is abrupt, persistent, or confined to part of the array. That is a service call, not a maintenance task.

Does degradation change solar payback?

It lengthens it, modestly. Annual value = future annual production × value per kWh, and a system delivering 94% of first-year output in its eleventh year is delivering 94% of the savings too. Over a payback period measured in years rather than decades, a few tenths of a percent compounding has limited room to act.

What moves payback more is everything degradation is often blamed for: the value of a kWh where you live, whether excess generation is credited at retail or wholesale, how much you consume on site, an inverter replacement partway through, and financing. If your quote assumed a degradation rate at the low end of the published range, it is worth re-running it at the mean instead and seeing whether the case still holds. If it only works at 0.5%, it is a thinner case than it looked.

What to do next

1. If production is stable year over year: nothing to do. Note your first-year figure so future comparisons have a baseline.
2. If it is drifting down by a few tenths of a percent a year: that is consistent with normal module aging.
3. If it dropped abruptly: check monitoring alerts and per-string data, then contact the installer. Do not investigate the hardware yourself.
4. If you are near a warranty threshold: read the schedule, the test method and the claim conditions before assuming a claim exists.
5. If the system is old but working: production at 20 years is typically still around 88% to 90%, which rarely justifies replacement on energy grounds alone.