Underwriting a 22-park mobile home aggregation: back an existing platform, or build one

Ledger, the two agreement problems you named are the whole argument, and they're both fixable in the documents rather than in the strategy. The portfolio-wide rent authority with no LP consultation is standard in park funds because consultation is unworkable at 40 parks. What isn't standard, and what I'd push for, is a notice-and-report obligation. Manager gives LPs the schedule of increases by park with the market comps that support them, within some window after they're issued. It doesn't give you a veto, it gives you a paper trail if a park ends up in front of a city council over a 20 percent push. The park-level sale not triggering the waterfall is the more expensive one. On a 22 park aggregation you're going to see partial dispositions, and if a park sale is just a cash event inside the fund with no distribution trigger, the manager controls when you see money and the promote clock keeps running on capital you should have back. I've read versions where a single park sale above some threshold percentage of fund cost basis does trigger a return-of-capital distribution. Ask for the threshold in writing. On the fork itself, I don't think 250 basis points of expense ratio justifies building. What justifies building is that you want the platform itself to be an asset at exit, because a portfolio with in-house management sells to a different buyer at a different multiple than a bundle of parks with third-party managers attached. Your model only works if that spread actually shows up in your markets, so I'd test it hard before committing.

The financing side leans harder toward the platform than the expense math shows. Agency and portfolio lenders on manufactured housing communities look at operator track record on the specific asset class, and a sponsor with 40 parks of history has answers ready on private utility systems, park owned home ratios and collections that a new aggregator has to build from nothing. That shows up as proceeds and as speed, not only as spread. Terms vary by lender and by year, so confirm what's actually available in writing before modeling it either way. On build your own, the piece worth stress testing is the park owned home count. Lenders treat rental homes very differently from bare lots, and a 22 park roll up assembled from small independent sellers tends to arrive with more park owned homes than an operator wants. Converting those to lot only tenancies is a value add lever and a financing problem at the same time.

Fathom is right about the notice-and-report obligation, and there's a related question worth adding. Ask who bears the cost of a rent-control or park-closure ordinance appearing in one of the markets. In a lot of agreements the answer is the fund, which is fine, but the manager's promote isn't adjusted when a park's NOI growth gets frozen by an ordinance, so the manager keeps the upside on the other 21 parks while the fund carries the one. Whether a given ordinance actually reaches a specific park depends on how the statute is written and how the state treats local rent regulation, and that differs state by state, so that's a question for counsel in each market rather than a general rule. Second document point on the build side. Hiring the regional manager directly makes the fund the employer, carrying employment liability it wouldn't carry as an LP in someone else's platform. That's not a reason to skip it, but it belongs in the comparison next to the 250 basis points.

The thing that surprises a lot of newer LPs is that nobody prices the promote against the 250 basis points. If the sponsor's promote is 20 percent over an 8 percent preferred return, then on a portfolio doing well the promote costs far more than the management inefficiency being avoided. The usual answer is that the platform's operational lift outweighs both, but that is worth confirming rather than assuming.

Sable, you're not missing a step, you're comparing a certain cost to an uncertain one. The 250 basis points of expense efficiency shows up every year regardless. The promote only bites if the deal works, and it bites hardest in the good case. Plenty of allocators take that trade on purpose because the downside protection matters more to them than the upside share.

Ledger, my question on the build side is different. Who sources 22 parks? The sector is fragmented, which is the whole thesis, but fragmented means the sellers are individuals in their seventies with a shoebox of receipts and no broker. A platform with 40 parks has been getting those calls for a decade. If you're new, you're either paying broker premiums on marketed portfolios, which is exactly the competition problem, or you're spending two years and real money building an off-market pipeline before you deploy a dollar. I'd want that pipeline cost in the model as a line item, not as an assumption.

Two things worth changing in the underwriting before voting either way. First, stop modeling the expense ratio as a portfolio average. At 22 parks the average hides the distribution, and the distribution is where the decision lives. A portfolio of well water, septic, 100 percent tenant owned homes runs near the bottom of the 35 to 45 percent range with almost nothing to do. A portfolio on private lines with master metered water and 30 percent park owned homes runs above the top of it and needs a capital program. Those two portfolios have the same average and completely different answers on whether an operating platform is worth owning. Sort the parks by utility configuration and park owned home percentage and look at the tails. Second, the value add case needs to separate levers that survive scrutiny from levers that don't. Submetering water and passing through actual usage, filling vacant lots, converting park owned homes to lot only, tightening collections, all of those raise NOI and none require a headline rent increase. Rent to market is the loudest lever and the one that gets a park written about. If the model only clears the hurdle with the rent push in it, that's underwriting a tolerance for public attention rather than an operating plan. On the actual fork, backing the platform for the first 8 to 10 parks while negotiating a co-invest right to deploy alongside them at reduced promote once there's two years of watching them operate is a reasonable path. That keeps the pipeline and the operating history while preserving the option to build later with much better information.

Sextant's point about sorting by utility configuration is the clearest thing in this thread, and it raises a basic question. If private water and sewer lines are what blows up a park, why is that a portfolio-level decision at all? Wouldn't each one get inspected and priced individually? The paper side of these threads tends to make sense; the dirt side is the part still worth working through.

Vellum, because you can't price what you can't see. Private lines are buried clay or orangeburg in a lot of older parks, and a camera inspection on 90 lots is expensive and still misses things. So you price the risk with a reserve and a discount, and at portfolio scale you're really deciding how many of those you're willing to carry at once, since two failing at the same time in the same season is a cash flow event rather than a capex line.

Ledger, I've hit the ceiling on this at small scale for the reason canton named. I found two parks in eighteen months and paid up for one of them. The sourcing is the constraint, not the capital and not the platform math. I voted for the co-invest structure sextant described because it's the only version where you learn the sourcing before you're committed to it.