Which of Your Rentals Actually Qualify for the STR Tax Treatment (Probably Not All of Them)
The short-term rental exception is not a status you have. It is a test each property either passes or fails, every year, on one number: the average length of a stay. I have yet to look at a mixed portfolio where the owner guessed that number correctly, and the guess is almost always too generous.
What the rule actually says
Rental real estate is passive by default under §469, which is why rental losses usually cannot offset W-2 or business income. The exception people call the STR loophole comes from Temp. Reg. §1.469-1T(e)(3)(ii)(A): if the average period of customer use is seven days or fewer, the activity is not a rental activity for §469 purposes. It is treated as a business, and a business loss you materially participate in is not passive.
Read that again, because the operative words are "average period of customer use." Not "is it on Airbnb." Not "do you self-manage." Not "is it furnished." The platform is irrelevant. The furniture is irrelevant. It is arithmetic on your own booking data.
How the average is calculated
Total nights booked, divided by the number of bookings. Per property. Per tax year.
That denominator matters more than people expect. Two hundred nights across eighty stays is 2.5 nights and clears the test comfortably. The same two hundred nights across five stays is forty nights and fails it outright. Identical occupancy, identical revenue, opposite tax treatment.
Cancelled bookings do not count. Owner stays are not customer use. And a stay that runs past year end is the kind of edge your CPA should see rather than something to resolve yourself at 11pm.
Where mixed portfolios go wrong
Most operators who have been at this a few years are not running one thing. They have some nightly rentals, a couple of mid-term units rented to travelling nurses or insurance placements, and maybe a long-term lease or two that came with a building.
The nightly units usually pass easily. Mine run between two and four nights. The mid-terms never do. A thirty-day minimum is, by definition, an average period of customer use of at least thirty days, and thirty is not seven or fewer. A twelve-month lease is not close to the line. It is not on the same map as the line.
I worked recently with an operator running fourteen doors across two entities. They described the portfolio as short-term rentals, because that is the business they are in and that is how they think of it. When the bookings were run through the test, six properties came out under four nights and qualified. Three mid-term units averaged over seventy nights. The remaining five were long-term and had no booking data at all, because a lease never passes through a property management system.
Six of fourteen. The other eight are perfectly good rentals producing perfectly good income. They are just not short-term rentals for the purpose of this rule, and treating them as though they were is how a return gets built on a number nobody checked.
The test you still have to pass after that one
Clearing seven days does not finish the job. It only moves the property out of "rental activity" and into "business." To take the loss against other income you also have to materially participate, under one of the seven tests in Reg. §1.469-5T. The one most operators lean on is 500 hours, or the 100-hour test where nobody else does more than you.
These are separate hurdles and both have to clear. I have seen the 7-day test treated as the whole loophole often enough to think the shorthand is doing real damage. Passing it and then failing material participation leaves you with a suspended passive loss, which is not nothing, but it is not the outcome anyone was planning around in October.
Grouping, briefly, and why I am not going to tell you what to do
Reg. §1.469-4 lets you group activities that form an appropriate economic unit, which can change how both tests apply across several properties. There is also the §1.469-9(g) election, which covers rental real estate for real estate professional status and specifically cannot include activities that are not rental activities, which is what your sub-seven-day properties have become.
That interaction is genuinely tricky and the facts matter. It is your CPA's call, and anyone on the internet telling you confidently which way to go, including me, does not know your return.
What to do before year end
Run the number on each property. Not the portfolio. Each property, this year, total nights over total stays.
If something is sitting just over seven, you still have about nine weeks of calendar to change, and minimum-stay settings are the lever. If something is at thirty or ninety, stop planning around it qualifying, and get it off the mental list of properties you expect the exception to cover.
And if you cannot produce the number at all, that is the finding. The data is in your booking history, which means either your tools compute it or you are going to reconstruct a year of stays by hand in March. We built RentReel to calculate it from the reservations an operator already has, per property, because it is the single number that decides the treatment and it was the one thing nothing else in the stack would tell me.
I checked mine this month. I would check yours.