
20 years in infra & renewables, and 2+ GW developed & financed
Kevin Feldman - 6 January 2026, ParisThe same spreadsheet, different assumptions. Why renewable energy financial models ‘sound’ different depending on whether you are in the US or Europe.The assets are similar. The spreadsheets look familiar. But beneath the surface, US and European renewable energy financial models reflect very different incentives, structures, and ways of thinking about risk.
Two big US features often seem like “quirks” to Europeans:
The US has historically stimulated investment through the tax code rather than through direct subsidies or regulated tariffs. Federal tax credits (the ITC/PTC and their IRA-era cousins under the Inflation Reduction Act) are dollar-for-dollar reductions of federal tax liabilities, and over time they became a core driver of renewable deployment. This changes both the timing and presentation of cash flows in the model.
The problem is that many developers and sponsors don’t have enough taxable income to use those credits efficiently. So they bring in a tax equity investor whose job is to monetize those tax attributes upfront. That creates an extra layer in the capital stack and introduces such fun concepts as:
•Flip dates and target tax-equity IRRs,
•DROs and tax capital accounts that must stay compliant with IRS rules,
•Basis step-ups, ITC recapture, HLBV, etc.
All of that sits below EBITDA, largely in the tax schedules. Most European projects have nothing comparable.
In many US models, the cash flow available for debt service (CFADS) used for debt sizing is pre-tax, and tax is pushed below debt service in the waterfall. That removes circularity from the pure banking case.
A big reason is structural: project SPVs are typically LLCs taxed as partnerships, so they don’t pay corporate income tax. Tax is borne not at the project level but instead by the partners and tax-equity investors. The lenders naturally want to be paid first out of pre-tax cash.
European models often feature the opposite: tax is deducted inside CFADS and inside the debt-sizing loop. Combine that with reserves and sculpting and you get a natural circular reference – usually neutralized with copy/paste macros.
A pattern I’ve seen several times in US models looks something like: size debt as the minimum of:
•a P50 case at target DSCR, and
•a P99 one-year case at 1.0x DSCR.
In Europe, debt sizing is more often anchored on a P90 DSCR test as the main constraint, with P50 used more as a sense-check than as the sizing base.
Both sides will also cap leverage with a maximum gearing ratio, which can end up setting the final debt size depending on how rich the project cash flows are.
If you strip it down, both approaches are trying to answer the same question: “what happens when things go wrong?” But they do it with slightly different “accents”:
•US: “We’ll run economics on P50 but make sure the far tail (P99) can still pay the bank back.”
•Europe: “We start from a conservative ‘bankable’ case (P90) and size from there.”
Take a very simplified example (purely illustrative numbers assuming a typical wind or solar resource distribution, where P90 and P99 are roughly 90–95% and 80–88% of P50, respectively):
•P50 CFADS = 100
•P90 CFADS = 93
•P99 CFADS = 86
If a US lender targets 1.30x DSCR on P50, the P50 test allows c. 100 / 1.30 ≈ 77 of annual debt service, while the P99 1.0x test allows 86 - so in practice the P50 test binds.
A European lender sizing on P90 at, say, 1.20x would allow c. 93 / 1.20 ≈ 78 of annual debt service.
So in this stylized case, P50 at 1.30x (with a P99 tail check) and P90 at 1.20x give almost the same debt size - different “accents”, similar underlying risk appetite.
Behind that sits another structural nuance. In the US, there is often an implicit or explicit expectation of refinancing - mini-perms, back-levered refinancings once projects are de-risked, or tax-equity plus shorter-tenor bank debt that gets taken out later. European lenders are, on average, more “take-and-hold”: they underwrite as if they will sit in the deal for the full term and spend more time worrying about long-term tail risk and back-ended volatility.
Numerically the approaches are not miles apart - a P90 DSCR in the 1.15x-1.25x range will usually give you a debt amount similar to a P50 DSCR of 1.30x with a P99 tail test - but to a credit committee they feel different. One sounds like “robust in the base with hard tail checks;” the other sounds like “conservative by design.”
European project finance models almost always carry a proper balance sheet and P&L that actually matter, because:
•Retained earnings and legal reserves can constrain distributions;
•Some financing docs reference accounting concepts directly;
•IFRS-based reporting often pushes sponsors to care about how the project shows up in the accounts as well as in cash.
In many US models, the focus is mainly on cash-flows, with a lighter treatment of the balance sheet. The project vehicle often sits inside a larger US GAAP environment, but for day-to-day decision-making on a single asset, cash is king. In practice, distributions are constrained by covenants and structure, not by statutory reserve mechanics, so accounting equity matters less in the project-level model.
In the US, I’ve met plenty of people who are fairly relaxed about carefully managed circular references (“Excel will iterate, we’re fine”), but quite wary of macros.
In Europe, circularities are treated as a big no-no, so people tend to push loops into macros instead: debt sizing, sweeps, sculpting, you name it.
So the US version might have iterations turned on and everything visible on the sheet; while the European version looks “clean” but hides the moving parts in VBA. None of this makes one side more “conservative” than the other - it’s just where you park the moving parts.
Send the US version to a European bank and you get “please turn iterations off;” send the macro-heavy European file to a US sponsor with locked-down laptops and half the team can’t run it.
One market difference shows up very clearly in financial models: basis risk.
In the US, most power markets are nodal or zonal with locational marginal pricing (LMP). That means a project’s revenue depends not just on the power price, but on where the power is injected into the grid. The spread between a project node and a hub price is real, volatile, and often material - and US lenders expect it to be explicitly analyzed and modeled.
As a result, US models almost always include node-to-hub basis assumptions, historical basis analysis, and stress cases for transmission congestion and curtailment-driven spreads.
In much of Europe, market coupling and zonal pricing mean this concept barely exists in day-to-day discussions. Many European practitioners have never even heard the term “basis risk,” because it historically hasn’t been a primary driver of project revenues.
But when European investors look at US assets, that blind spot matters. A risk that is invisible in one modeling culture is front and center in the other.
This is another example of how identical Excel tools can encode very different assumptions about what “risk” actually means.
None of the above is about “right” and “wrong”. It’s just tax systems, legal frameworks, accounting regimes and banking habits showing up in Excel.
The risk in cross-border deals is that people talk past each other. You think you’re arguing about numbers and assumptions when you’re really arguing about conventions and structure. Ironically, much of project finance debt on both sides of the Atlantic is ultimately provided by the same European and Japanese banks - but with different local house styles and credit lenses.
At SNOW, we operate in both the US and Europe, with an active transatlantic team supporting renewable energy transactions across markets. We advise on tax-equity-heavy US deals as well as classic European project finance, and we also bring hands-on experience in solar and wind project development - not just “pure investment banking.”
If you’re a US developer targeting European investors (or lenders), or a European investor looking at US tax-equity structures, that “modeling accent” is more than a cosmetic issue. It affects how much debt you can raise, how quickly you get to IC approval, and how comfortable people feel with risk.
Part of our job at SNOW is to make sure the model sounds “native” to the people being asked to put capital at risk.