Discount rates and NPV for solar projects

A payback period tells you how many years it takes to recover an outlay. It does not tell you whether that outlay was a good use of capital compared with the alternatives, because it ignores what happens after payback and it ignores the time value of money. Net present value (NPV) fixes both problems by discounting every future cash flow — electricity savings, export income, maintenance costs, an inverter replacement — back to today's euros and summing them against the initial investment. This article works through how to pick a discount rate for a domestic or commercial PV project, how to build the cash flow table, and where modellers most often get the arithmetic or the assumptions wrong.

By the end you should be able to build a simple NPV model in a spreadsheet, defend your choice of discount rate to a sceptical client or lender, and recognise the difference between a project that merely pays for itself and one that has actually created value. This complements the payback calculator and the discussion of true system cost in the honest payback calculation, both of which stop at the break-even year rather than valuing the full asset life.

Why payback is not enough

A PV array with a 25 to 30 year service life keeps generating well past the point where the initial outlay has been recovered. Two systems can have identical eight-year paybacks and very different lifetime value if one degrades faster, needs a second inverter sooner, or is installed on a roof with worse long-term shading. NPV captures all of this in a single number, expressed in today's euros, that can be compared directly against other uses of the same capital — paying down a mortgage, a term deposit, or a different capital project entirely.

The mechanic is simple. Each year's net cash flow (savings plus export revenue minus operating costs) is divided by (1 + r)^n, where r is the discount rate and n is the year number counting from the investment date. Summing these discounted flows and subtracting the initial capital outlay gives the NPV. A positive NPV means the project earns more than the discount rate; a negative NPV means the money would have done better elsewhere at that same rate.

Choosing a discount rate

The discount rate should reflect the opportunity cost of the capital being spent, adjusted for the risk profile of the cash flows. For a private homeowner paying cash, a common reference point is the after-tax return available on a safe, liquid alternative — a savings account or short-dated government bond — plus a margin for the fact that PV savings are not perfectly certain (electricity prices could fall, the household could move house before capturing full value, equipment could underperform). Many owner-occupier models in Europe use a real discount rate in the range of 3 to 6 percent for cash-funded residential systems.

For a business or a project financed with debt, the appropriate rate is usually the weighted average cost of capital (WACC), blending the cost of debt (the loan interest rate, adjusted for any tax shield) and the cost of equity (the return equity investors require given the project's risk). Utility-scale ground-mount projects with long-term power purchase agreements and low technology risk often use real discount rates in the 4 to 8 percent range; smaller commercial rooftop projects with shorter contracts or more uncertain future tariffs are usually appraised at the higher end or above.

A critical and frequently mishandled detail is the distinction between nominal and real discount rates. If your cash flow forecast is in nominal euros — that is, it includes an assumption about future electricity price inflation — you must discount with a nominal rate. If your cash flows are expressed in constant, today's-euro terms, use a real (inflation-adjusted) discount rate. Mixing the two, for example inflating electricity prices but discounting with a real rate, systematically overstates NPV and is one of the most common errors in amateur solar financial models.

Building the cash flow table

A defensible model separates the components of each year's cash flow rather than collapsing them into a single "savings" line. A reasonable structure is:

Degradation should be applied to the generation figures, not bolted on as an afterthought — see degradation assumptions in financial models for how to build a realistic year-by-year degradation curve rather than a flat linear guess.

Worked example

Consider a 6 kWp residential system costing €9,000 installed, generating 6,600 kWh in year one, with 55 percent self-consumption at a retail tariff of €0.30/kWh and the remainder exported at €0.09/kWh. Assume 0.5 percent annual degradation, 2 percent annual tariff escalation, annual operating cost of €80 (real terms), an inverter replacement of €1,200 in year 15, and a real discount rate of 4 percent over a 25-year horizon.

Year 1 cash flow: self-consumed energy 3,630 kWh × €0.30 = €1,089; exported energy 2,970 kWh × €0.09 = €267; gross revenue €1,356; minus €80 operating cost = €1,276 net.

Year 1 discounted value: €1,276 / (1.04)^1 = €1,227.

Repeating this for 25 years, applying 0.5 percent degradation to the energy figure and 2 percent escalation to both tariffs each year, and subtracting the discounted €1,200 inverter cost in year 15 (discounted value roughly €1,200 / 1.04^15 ≈ €666), a spreadsheet sum of the discounted annual flows for this profile comes to approximately €19,800. Subtracting the €9,000 initial outlay gives an NPV of roughly €10,800 at a 4 percent real discount rate. Raising the discount rate to 7 percent — appropriate if you consider the future tariff assumptions optimistic — reduces the sum of discounted flows to roughly €14,200, for an NPV of about €5,200. The project remains positive at both rates in this example, but the sensitivity is large: doubling the discount rate roughly halved the NPV, which is typical for a long-lived asset with cash flows weighted toward later years.

Internal rate of return and payback compared

The internal rate of return (IRR) is the discount rate at which NPV equals zero — it answers "what return does this project actually generate", independent of any external benchmark rate. IRR is useful for ranking competing projects but can be misleading when cash flow patterns are unusual (for example, a large negative flow midway through life from an inverter failure can produce more than one mathematically valid IRR). NPV at a chosen discount rate avoids this ambiguity and is generally the more robust decision metric.

What goes wrong in practice

The most common modelling error, as noted above, is discounting nominal cash flows with a real rate or vice versa. The second most common is applying an unrealistically low discount rate to make a marginal project look attractive — using, say, a mortgage refinancing rate of 2 percent for a homeowner who could otherwise invest at a higher risk-adjusted return elsewhere. A third error is ignoring the inverter replacement cash outflow entirely, which for a 20-25 year model horizon materially inflates NPV; see inverter replacement and lifecycle costs for realistic timing and cost assumptions. A fourth is assuming flat electricity tariffs for 25 years, which understates value if prices rise in real terms and overstates it if they do not — sensitivity testing on this single assumption is essential. Finally, models that ignore self-consumption share entirely, or assume a constant self-consumption percentage across 25 years regardless of household changes (electric vehicle purchase, heat pump installation, children leaving home), tend to be wrong in ways that are large enough to change the investment decision; cross-check assumptions against the self-consumption calculator.

Sensitivity and scenario analysis

Because NPV is sensitive to several compounding assumptions — discount rate, degradation rate, tariff escalation, and self-consumption share — a single-point NPV figure should always be accompanied by a sensitivity table showing NPV under low, central and high assumptions for at least the discount rate and the tariff escalation rate. This is more informative to a client than a false-precision single number, and it demonstrates that the analyst has tested the fragility of the conclusion rather than presenting an optimistic base case as if it were certain.

Standards and references

There is no IEC or EN standard governing financial appraisal methodology for PV; discounted cash flow analysis follows standard corporate finance practice. However, IEC 61724 (PV system performance monitoring) and the degradation guidance referenced by IEA PVPS Task 13 are the relevant technical inputs feeding the generation forecast that underlies the cash flow model — the financial model is only as good as the yield estimate it is built on.

Frequently asked questions

What discount rate should a homeowner use?

There is no single correct number, but a reasonable starting point is the after-tax return available on the safest liquid alternative use of the same capital, plus a small premium for the uncertainty in future electricity prices and equipment performance. Many owner-occupier models use 3 to 6 percent in real terms. Test the conclusion at both ends of that range rather than committing to one figure.

Is NPV better than payback for comparing quotes?

Yes, for comparing two systems with different costs, generation profiles or degradation characteristics, because NPV accounts for the full asset life and the time value of money. Payback remains useful as a simple communication tool, particularly for buyers who want a rough intuition, but should not be the sole basis for a purchase decision.

Should I use a nominal or real discount rate?

Use whichever matches your cash flow forecast. If your electricity price forecast includes inflation or tariff escalation, use a nominal discount rate. If your cash flows are expressed in today's euros with no escalation assumed, use a real discount rate. Mixing the two inflates the result and is the single most common modelling error in this area.

Does a higher discount rate always favour smaller systems?

Generally yes, because it penalises cash flows further in the future more heavily, and larger systems typically have proportionally more of their value concentrated in later years due to the fixed-cost components of installation. A higher discount rate narrows the gap between marginal and strongly positive projects and can flip a marginal project to negative NPV.

How do I account for an inverter replacement in the model?

Treat it as a single negative cash flow in the year it is expected to occur, discounted like any other flow. If the timing is uncertain, model it under both an early and a late replacement scenario and report the NPV range rather than a single point estimate.

Can NPV go negative even with a positive payback?

No — if simple payback occurs within the analysis horizon, the undiscounted cumulative cash flow is positive by definition at that point, but a very high discount rate applied over a long horizon can still produce a negative NPV if most of the value is generated late in the asset's life and the discounting erodes it heavily. In practice this is uncommon for typical residential discount rates but can occur for aggressive commercial hurdle rates above roughly 10 percent.