Is P50/P90 still the right way to present yield to a lender?

Linnea WeberFusion Engineering · 18 yrs

It is not just your client set. Leading with the loss tree and relegating exceedance to the summary table is where most of the independent engineers I work with have ended up, for a reason that is more structural than stylistic: P50 and P90 are outputs of the assumptions, so presenting them first invites the reader to interrogate a number whose derivation they have not seen yet. Leading with the itemised losses means the exceedance figures arrive as a consequence of things the reader has already accepted or challenged, and the conversation is much shorter.

The one thing I would keep prominent is the uncertainty breakdown behind P90, separated into interannual variability and model or measurement uncertainty. Those two behave completely differently in a financing context — interannual variability diversifies across a portfolio and across the loan tenor, model uncertainty does not — and a single combined standard deviation hides that. Lenders who are paying attention ask for the split, and providing it unprompted is the single change that has reduced query rounds most in my reports.

Leila NakamuraGrid Storage Engineering · 23 yrs

Verified expertise in Grid Storage Engineering

The reason the loss tree gets all the attention in due diligence is that it is where the recoverable disagreement lives. A technical adviser cannot really argue with your irradiance dataset beyond swapping it for another one and noting the delta, but they can argue that a 2 percent soiling loss in a dusty location with quarterly cleaning is optimistic, and that argument moves the P50 by more than the dataset choice does. Soiling, availability and degradation are the three that carry the disagreement in nearly every review I have been part of, and they are the three where the basis column matters most.

On availability specifically: the assumption has to be reconcilable with the O&M contract, and if it is not, say so explicitly rather than letting the reader find it. A 99 percent availability assumption against a contract with a 97 percent guarantee and no liquidated damages is a finding, and it is much better as a stated limitation in your report than as a discovery in someone else's. The reports that go through cleanly are the ones that have already identified their own weak assumptions.

Thandi KowalskiNuclear Power Engineering · 6 yrs

Counterpoint on dropping exceedance from the lead, from the other side of the table: as a lender's adviser I do go straight to the loss tree, but the headline numbers still do real work. They are the continuity check against the previous report, against the sponsor's model and against the term sheet. If your P50 has moved 4 percent since the last issue I want to see that immediately, and then I want the loss tree to tell me which line moved. Burying it in a summary table makes that first check slower.

So the version that works best for me is neither ordering exactly: a short header block with P50, P90 and the delta against the previous issue, then straight into the itemised loss tree with the basis column, then the exceedance derivation and uncertainty split in full. Three lines at the top, not a page. The complaint about leading with P50 is really a complaint about reports where the headline is the only thing presented with confidence.