The fund model treated permitting time as a constant. It isn't.

Passive position, $150k into a small fund building detached ADUs in a permissive metro and selling them to homeowners as a turnkey product financed by the homeowner. Two year projected hold, quarterly distributions after month nine.

Distributions started at month sixteen and have come in at roughly a third of the projected rate. The fund is not in trouble, the sponsor communicates well, and the units are selling. The model was wrong in one specific place and I should have caught it during diligence.

The pro forma had a permitting line of 10 weeks applied uniformly to every unit in the pipeline. I read that as an average and moved on. It was not an average, it was the sponsor's best observed case, taken from their first three units, all of which were in the same submarket with the same plans reviewer and all of which were on flat lots inside an existing pilot program. As the fund scaled into other jurisdictions in the same metro, permitting ran 14 to 34 weeks. Capital sat in half-built inventory for months at a time and the sponsor's carry on the construction facility ate the spread.

What I should have done during diligence: asked for the distribution of permit durations, not the number in the model. One question. "What were the actual permit timelines on every unit you've completed, listed individually?" A sponsor who has done nine units can answer that in an email. If the answer is three units with an asterisk, I know I'm underwriting a projection dressed as a track record.

The second thing I'd change is that I read the operating agreement carefully on waterfall and fees and skimmed the section describing the pipeline assumptions, which was in an appendix to the PPM rather than in the agreement. The economics I checked were fine. The operational assumption that drove the economics was in a document I treated as marketing.

I'm still in it. The units sell, the thesis on ADU demand holds, and I think we get to something respectable at month thirty six. My return on this will be materially below what I underwrote, and the cause is a single unexamined constant.