Almost every vacancy-rate explainer online was written for an apartment complex. They open with the same formula: vacant units divided by total units. That works when you own 200 doors and three of them turn in a given month. It falls apart the moment you own three units total, because the formula can only ever read 0%, 33%, 67%, or 100%. A single turnover swings your “rate” by a third, and a fully leased year reports a perfect zero that means nothing.
For a 1-10 unit landlord the useful question is not how many units are empty on a snapshot date. It is how many rentable days you lost over the year, and how many rent dollars went with them. Those are two different measurements, physical vacancy and economic vacancy, and the gap between them is where small portfolios quietly bleed. This page is the formula and the worked numbers. The argument for why you should normalize vacancy when you underwrite, rather than trust a single year, lives in the companion piece on pro forma vs actuals.
Physical vacancy, measured in days
Physical vacancy is the share of available rental time a unit produced no paying tenant. The small-portfolio version counts days, not units:
Physical vacancy rate = vacant days ÷ rentable days
Rentable days is the number of days the unit was actually available to rent, which usually means the full period minus any time you took it offline for a major renovation you chose to do. Say you own a triplex and one unit sits empty for 30 days between a move-out and the next lease. Over a 365-day year that unit had 365 rentable days. Across all three units you have 1,095 rentable unit-days and 30 vacant unit-days. Your physical vacancy is 30 ÷ 1,095, or about 2.7%. The vacant-units formula on the same event would have told you 0% on every day except the snapshot, then a sudden 33%. The day-based number is the one you can actually plan around.
Counting days also forces you to be honest about turnover. The empty stretch between tenants is not a clean 30 days. It is the days from the old lease ending, through the move-out inspection and the make-ready work, through showings, to the day the new lease actually starts paying. If you track that window per unit, your vacancy number stops being a guess.
Economic vacancy, measured in dollars
Physical vacancy still flatters you, because a unit can be physically occupied and still not paying full rent. Economic vacancy fixes that. It measures lost rent against the most the property could have collected:
Economic vacancy rate = (gross potential rent − rent actually collected) ÷ gross potential rent
Gross potential rent is every unit at market rent for the whole period, as if the building were full and everyone paid on time. Take that same triplex, three units at $1,500, so $4,500 a month or $54,000 a year of gross potential rent. Now layer in a real year: one unit empty 30 days (about $1,500 lost), a half-month free-rent concession to land a good tenant fast ($750), and one tenant who skipped a final month you never recovered ($1,500). That is $3,750 of lost rent. Economic vacancy is 3,750 ÷ 54,000, or 6.9%. The physical-vacancy number for the same year was only 2.7%. Same building, and the honest figure is more than double.
The pieces physical vacancy misses are exactly the ones that hurt:
- Concessions. A free half-month or a discounted first month is rent you gave up to fill the unit. The lease says $1,500 but the collection says less.
- Turnover gaps. The days a unit is technically available but earning nothing while you clean, repair, and re-lease. These compound with the cost of turnover itself.
- Loss to collections. Rent that was charged, came due, and never arrived. On paper the unit was full. In the bank account it was empty.
- Below-market rent. A long-term tenant paying $1,300 in a $1,500 unit is a quiet $200-a-month economic vacancy you may choose to keep, but you should at least see it.
Where vacancy lands in NOI and cap rate
Vacancy is not a footnote. It sits near the top of the income statement, right after gross potential rent, and everything below it inherits the number. Subtract vacancy and credit loss from gross potential rent and you get effective gross income. Subtract operating expenses from that and you get net operating income. Divide NOI by price and you get the cap rate you are paying. Understate vacancy by three points on that $54,000 triplex and you have invented about $1,620 of phantom NOI, which at a 6% cap rate is roughly $27,000 of price you are overpaying. Vacancy errors do not stay small.
This is also why vacancy belongs in the operating expense ratio conversation even though it is technically a deduction from income, not an expense. Both are levers a seller's pro forma loves to set to their most flattering position. If you are building a small rental portfolio, a consistent vacancy assumption across every deal is what lets you compare them honestly instead of rewarding whichever seller had the luckiest twelve months.
What a good vacancy rate actually is
There is no single right number, and anyone who quotes one without asking where your property sits is guessing. A commonly used healthy band for stabilized small residential rentals is 3% to 7%, but that is a starting point, not a rule. The honest source for market-level rates is the U.S. Census Housing Vacancy Survey, which publishes rental vacancy by region. Metro-specific figures move year to year, so check current data rather than trusting a number you read once.
The trap is treating one good year as proof. A 0% vacancy over twelve months on a three-unit building is a sample size of one, not a track record. The next move-out, the next concession, the next slow re-lease will pull you back toward the mean. Underwrite for the mean, not the best year you have seen.
How to track it without a spreadsheet that lies
The reason most landlords never compute economic vacancy is that the raw material is scattered. Lease start and end dates live in one place, the rent you actually collected lives in another, and the days a unit sat empty between tenants live in your memory until they do not. A spreadsheet can hold all of it, but only if you update it the day something happens, and a turnover is the busiest week to ask yourself to do bookkeeping.
I built rents.ai because that reconciliation kept slipping through my own spreadsheets. Its rent roll records payments per unit, so the vacant days and the lost rent fall out of data you already entered, and its investor-grade analytics normalize vacancy at 5% so one lucky full year does not inflate the NOI and cap rate it reports back to you. The honest limitation: it does not pull a live market-rent feed, so the gross potential rent it compares against is the market rent you type in per unit, no better than your own number for the area. The math is only as good as the rent you tell it the unit should earn.