The vacancy number on a boarding house does not behave like apartment vacancy

The distinction you are drawing between average vacancy and clustered vacancy is real, and it does not get enough attention in boarding house underwriting.

The standard pro forma treats vacancy as if residents leave at perfectly spaced intervals, which keeps the annual number smooth. What you are describing is variance, meaning the actual spread of when those vacancies hit. A 12 percent annual rate means roughly 1.4 rooms vacant at any moment across a 12-room house. If those vacancies land in the same quarter, you might have four rooms empty for six weeks while the rest of the year looks fine. The year-end figure averages that away.

Your turn-cost point matters too. The gap between one resident leaving and the next paying rent is not just lost rent. You have cleaning, inspection, possible repairs, and intake processing before the room is rentable again. If your licensing requirements include a compliance check before re-occupancy, that gap can extend further than most operators plan for. I would not quote a specific number because it varies by municipality and operator, but checking what your local licensing framework actually requires at turnover is worth doing before you build any assumptions into a cash flow model.

Tracking vacant room-days per period rather than an annual percentage gives you a much clearer picture of when your cash position is under pressure, not just whether it is, averaged across the year.

A thought on the underwriting side: stress-testing a pro forma with clustered vacancy, say four rooms empty simultaneously for 45 days, tells you whether the property can service its debt through a bad quarter. That is a different question from whether the annual numbers work.

What software or spreadsheet setup are you using to track room-days currently? Knowing what you are working with would help me point to something more useful.

Per-room-days wins when you carry debt; annual percentage wins only for quick comparisons across properties.

The risk I do not see priced anywhere in these pro formas is the correlation between departures and the reason for departure. In a standard apartment building, turnover causes are scattered, job relocation, lease expiration, life change, and they do not travel together. In a boarding house serving residents with unstable housing histories, the causes cluster because the population shares common stressors, a benefit disruption, a seasonal income dip, a local shelter or program opening nearby that pulls residents toward a different option. When three rooms empty in the same 30-day window, the operator is often looking at a shared trigger, not bad luck, and underwriting built on independent vacancy events cannot capture that. The per-room-day method you described gets the measurement right, but measurement alone does not solve the problem unless the operator is also asking what caused the cluster, because the next cluster will likely share the same shape. What does the intake data on your 14-room house show about why residents leave, and is there any pattern in the timing relative to month-end or benefit cycles?

What market are you operating in, because seasonal clustering hits differently in a Rust Belt city versus somewhere with stable year-round demand?

Per-room-days wins on cash flow timing; annual percentage wins only when comparing properties of different sizes.

The intake process you described is doing more work than the vacancy number, and the part that compounds it is that boarding house operators almost never price the intake delay into their minimum rent-ready reserve. On a 14-room house, if two rooms go vacant inside the same 30-day window and the operator is running a 14-day turn on each, that is 28 room-days of lost revenue before a single new resident pays anything. At $600 per room, that is $560 in lost rent, which looks manageable until you add the cost of the turn itself, and then the real problem appears: the operator who set reserves at 10 percent of gross designed that reserve for a smoothed annual vacancy, so when three departures stack in Q3, the reserve is already partially drawn from the Q2 softness and the math stops working in the same period the debt service is still due.

The related piece that changes the picture here is how this interacts with security deposit structure. Most boarding house leases run month-to-month or week-to-week, which is why the vacancy clusters in the first place, but the deposit collected is usually one week or one month, and some operators in states with strict deposit accounting rules cannot touch that deposit until the room is formally inspected and damages are documented, which adds time before the reserve is replenished from that source. The operator who tracks per-room vacant days per quarter and also tracks the average days between deposit release and next move-in has a real picture of liquidity stress. What is your current lease term structure, month-to-month or something longer, because that single variable determines whether the clustering risk is structural or situational?

The two-rooms-same-month scenario is what broke my storage facility assumptions before I ever touched residential. At the storage place I could absorb a cluster because unit turns were fast and mostly cosmetic. A boarding house turn is slower and the intake process, background check, intake interview, whatever paperwork the operator has built, adds days that nobody pencils in as a hard cost. If your intake runs 5 to 7 days on top of the physical turn, you just added another $100 to $140 per room at $600/month before you have touched a paintbrush.

What I want to know from you is whether your debt service is monthly or quarterly, because that is what makes the clustering problem catastrophic versus just annoying. A lender who gets paid monthly does not care that December looked fine.

Clustering risk has a second layer most pro formas miss entirely.

Take that 14-room house at $600 per room. If the operator runs a 14-day turn and two vacancies land in the same month, the lost revenue is roughly $1,400 per room for the vacant days plus the turn period, so $2,800 total on just two rooms. Now assume those two rooms share a common trigger, a conflict between residents, a local shelter closing, a rule change that pushed out a subset of residents at once. Correlated departures are more common in this resident profile than in market-rate multifamily, because the residents often share a case manager, a subsidy program, or a social network, and when one thing destabilizes it destabilizes a cluster. That means the variance around your expected vacancy is wider than a simple 10 percent assumption implies, and wider variance on a property with thin margins and monthly debt service is a different risk than the same variance on a property with strong reserves.

The standard vacancy rate assumes independence between rooms the way a coin flip assumes independence between flips. Boarding houses do not give you that.

What is the debt service on that 8-room house relative to the revenue from a fully occupied month, because that ratio is what determines whether three bad months in a row is a bruise or a default?

Debt service coverage on boarding houses is often calculated on annual net operating income, and a lender who sees 1.25x DSCR on paper may never know that three of those months the property ran below break-even, because the annual figure smoothed it out. That matters the moment you are in a covenant-tested loan or trying to refinance, because the lender will ask for trailing twelve months and the number will look fine while your operating account tells a different story. The vacancy pattern you are describing also compounds with the intake process in a way that does not show up unless you are tracking room-days: if your operator requires background screening, income verification, or a waiting period before move-in, the turn time is not 14 days, it is whatever the slowest step in that intake pipeline takes, and during a cluster departure that pipeline backs up because you are processing multiple applications simultaneously with the same staff capacity.

The figure most pro formas never build is a rolling 90-day cash reserve sized to the property's worst historical quarter, not its average year. On a 14-room house carrying, say, a $4,200 monthly debt service obligation, a quarter where three rooms turn in the same six-week window can put you $6,000 to $8,000 behind before the annual rate moves more than two percentage points. A reserve sized to the annual vacancy assumption will not cover that. The question worth asking before you build the reserve figure is whether your intake process has any single-point-of-failure step, because that step is where the turn time actually gets set, and if it is a third-party screening vendor or a city inspection requirement, you do not fully control it. Do you have a picture of what the worst consecutive 90-day stretch looked like in dollar terms, not percentage terms?

My Cleveland 12-room showed 9 percent vacancy on paper in 2022 and I still had two months where I was short on debt service, which is exactly the math you are describing. The annual figure was fine. The quarters were not. What I started doing mid-2023 was tracking vacant room-days per month per unit, which sounds tedious but it is just a column in a spreadsheet, and the first thing it showed me was that my bad months were almost always February and August, which told me something about my resident population that the year-end number never did.

The line that got me was the one about 12 percent that never had a clean month. That is a property in permanent triage dressed up as a stable asset, and any lender or buyer reading the summary sheet is going to miss it entirely.

What I have not seen anyone price in here is the intake cost itself as a line item. Not just turn days, but whatever screening process the operator runs, because if you are doing income verification or references on transitional residents, that process takes time and sometimes falls through, and the room sits another week. My Cleveland manager charges a flat re-leasing fee per room that I had to negotiate down, and even so it moves my effective vacancy cost 15 to 20 percent above what the raw days would suggest.

Do you have separate line items for turn costs and re-leasing costs, or is it all folded into a generic vacancy allowance?

My cabin in Sevier County taught me this same lesson with STR nights, not rooms, but same math. Three bad weeks in January wiped out what looked like a healthy annual occupancy number.

That 30-day turn figure you're using, is that for a licensed facility with intake paperwork, because my understanding is some county-level programs require a waiting period before re-occupancy that can push a turn past 45 days before you even start cleaning?

My 8-room in Cleveland Heights ran 9 percent annual vacancy last year but had one quarter where three rooms turned inside six weeks.

Debt service coverage on a boarding house should probably be calculated on a rolling 90-day basis rather than annually, because the annual figure smooths exactly the problem you are describing. If a lender or partner sees 1.3x DSCR on a 12-month trailing basis, that looks fine, but the same property can run below 1.0x for a quarter and still produce that annual number, and that quarter is where you lose the property or burn through reserves. The clustering pattern you named also interacts badly with the intake process cost, because a property with a rigorous intake screen will run longer average turns than one that fills fast and churns faster, so the operator who is doing the job well on resident selection can show worse short-term vacancy numbers than the operator who is not. Per-room-day tracking at the quarter level is the right unit, and it also gives you something useful for pricing: if you can see that Q1 consistently runs higher vacant room-days than Q3, you can adjust your renewal incentive timing rather than reacting to the shortfall after it lands. The question worth building into your model is what your cash reserve threshold is per room, because the number that actually protects you is not the vacancy assumption, it is how many vacant-room-days your reserve can absorb before you are making decisions under pressure.

The misconception worth correcting is that turn cost is a vacancy cost. Most operators fold the repaint and intake days into their vacancy rate when they are actually a separate line, and that double-counts the problem in a way that makes the number look worse than it is while still missing the real issue, which is the timing of the cash gap.

The debt service problem you are describing is a liquidity problem, and the fix is not better vacancy tracking, it is a reserve sized to the worst plausible quarter rather than the average year. On a 14-room house with monthly debt service of, say, $4,500, a three-room cluster vacancy in the same 30-day window produces a shortfall of roughly $1,800 in revenue against full obligations, and if that happens twice in a year, an annual reserve built on 10 percent of gross potential ($10,080) barely covers it once, let alone twice with turn costs on top.

What I want to know is whether the $600 per room is fixed or whether any rooms carry a higher rate, because the concentration risk changes meaningfully if two of your highest-paying rooms turn at the same time versus two of the base-rate rooms.

My Akron duplex closes in three weeks and I have been running every number I can find on boarding houses in the same zip code because that is the next thing I want to buy. Every pro forma I have pulled from brokers uses annual vacancy, and until I read this post I did not have clean language for why that felt wrong to me. The 12-percent-across-12-months-but-never-a-clean-month framing is exactly what I was trying to articulate and could not. That sentence is going in my notes verbatim.

The thing I cannot get a straight answer on yet is what a normal reserve line looks like on a 14-room house relative to a standard multifamily. On my duplex I am holding two months of gross rent as operating reserve. On a boarding house with clustering vacancy risk and 14-day turns, that number seems dangerously low, but I have not found a single underwriting template that treats the reserve differently just because the vacancy profile is different. Does anyone running one of these actually size reserves against worst-case quarterly vacancy rather than annual, and if so, what multiple of monthly gross are you actually holding?

Turnover correlation is the part that gets left out of almost every boarding house model I have seen. When one resident leaves under stress, the conditions that made them unstable often apply to at least one neighbor, and a departure in room three raises the real probability of a departure in room seven within the following 45 days. That means your vacancy events are not independent draws from a pool, they are weakly correlated, and a flat annual percentage treats them as independent. The practical fix is to model vacancy in scenarios rather than as a single rate: a base case where departures spread across the year, and a stress case where three or four cluster into a single quarter, then size your operating reserve to survive the stress case rather than the average. The reserve question connects directly to debt service coverage, because a lender underwriting to 1.25x on annual NOI has no idea that the same property can run below 1.0x for six weeks straight without ever showing a bad annual number. What does your debt structure look like, because the clustering risk matters a lot more on a property with a monthly payment due on the first than on one where the debt is paid annually or deferred?