The Biggest Risk to Your Vacation Rental Isn’t a Party. It’s an Empty House in November.
Every owner conversation about a high-value home eventually arrives at the same fear, and it’s always a party. Somebody rents the house, forty people show up, and the place gets destroyed.
It’s a reasonable fear and I’ll spend most of this post on it. But I want to be honest about something first, because it took me a while to see it clearly: for a Pacific Northwest property, the party is not your largest uninsured exposure. An empty house in November is. I’ll get to why, and I think it’s the part of this that almost nobody writes about.
A note before I start. This post quotes statutes and insurance policy language. It isn’t legal advice and it isn’t insurance advice. Policy forms vary by carrier and by individual policy, and the legal questions here are genuinely unsettled in places. Talk to your own counsel and your own broker.
What the platforms actually do, and where they stop
Airbnb made its party ban permanent in June 2022, after a temporary ban in August 2020. The current policy language prohibits “open-invite parties, disruptive gatherings, or other community disturbances,” and notably applies “regardless of size.” That’s a stronger framing than a headcount cap, and it’s the one to know.
Behind the policy is a machine learning system, running year-round in the US and Canada since August 2022. It looks at review history, how long the guest has been on Airbnb, trip length, distance from the guest to the listing, weekday versus weekend, and other signals. Airbnb layers heightened restrictions on high-risk dates, and they publish the numbers: over 20,000 people redirected over the July 4 weekend in 2025, roughly 38,000 over Halloween 2024, over 20,000 in the US on New Year’s Eve.
That’s real, and it works. Airbnb reports that fewer than roughly 0.06% of US stays in 2025 resulted in a party report.
Now the limits, which matter more if you own the house.
Blocked guests are redirected, not removed. Airbnb’s own wording: guests who can’t book an entire home through this system “will still be able to book a private room or a hotel room through Airbnb.” Someone screened off your listing goes somewhere else. Possibly the unscreened house down the street.
It screens the booking, not the stay. The model scores a reservation at the point of booking. It cannot see a legitimate five-night reservation that becomes a party on night three, or a compliant-looking person booking on someone else’s behalf.
It’s Airbnb-only. It doesn’t touch your Vrbo bookings, your direct bookings, or Booking.com. And Vrbo’s equivalent is considerably weaker. Their own launch announcement says it plainly: “Vrbo does not block or cancel bookings, that decision can only be made by the host or booking party.” Vrbo emails you a risk score and lets you cancel penalty-free. Which means someone has to be reading those emails.
Airbnb also says something about its own system that I’d take at face value: “no system is perfect.”
What guest screening actually catches
Beyond the platform, professional screening runs through services like Truvi or Autohost. What they check: identity verification with a liveness selfie compared against government ID, watchlist checks against guests who’ve previously caused damage and refused to pay, disposable email and burner phone detection, payment and chargeback fraud signals, device and network behaviour, and criminal background checks.
That last one comes with limits worth stating plainly, because the industry oversells it.
Criminal screening is US guests only. Truvi runs against 2,400+ national and federal criminal databases with a seven-year lookback, and against the US National Sex Offender Registry. International guests get no criminal screening at all. Nor does anything older than seven years show up, and county-level coverage in aggregated databases is incomplete.
More to the point, and I’ll quote Autohost against their own industry here because it’s the most honest thing I read in this research:
“Criminal background checks were designed for long-term tenancy relationships. The risks are centered around identity fraud, property misuse, payment fraud, and behavioral issues like unauthorized parties, none of which are predicted or prevented by a criminal history report.”
That’s right. A clean record doesn’t predict whether someone invites forty people over. What actually correlates with party risk is behavioural: local bookings, last-minute reservations, one and two-night stays on holiday weekends, third-party bookings, and guest count relative to occupancy limits. That’s what a good screening process weights.
The fair housing part, which most operators get wrong
Here’s where I see confident nonsense written, so let me be careful.
You’ll read that the federal Fair Housing Act doesn’t apply to short-term rentals. Don’t rely on that. The FHA governs “dwellings,” defined as property “occupied as, or designed or intended for occupancy as, a residence.” Whether a nightly stay by someone who has a home elsewhere sits inside that definition is genuinely unsettled and fact-dependent. Courts haven’t resolved it for STRs.
And it barely matters, because other law fills the gap.
Title II of the Civil Rights Act of 1964 covers “any inn, hotel, motel, or other establishment which provides lodging to transient guests,” excepting only owner-occupied properties with five or fewer rooms. A professionally managed whole-home rental is not getting through that carve-out.
ADA Title III reaches “places of lodging” where stays are primarily 30 days or less, occupants have no right to return to a specific unit, and the operation has hotel-like characteristics: off-site management and reservations, housekeeping or linen service, no lease or security deposit. A professionally managed STR hits nearly every one of those.
And for anyone operating in Washington, the state law settles it. RCW 49.60.040 defines a place of public accommodation to include any place kept for gain or hire for “the entertainment, housing, or lodging of transient guests.” That’s explicit, and it sidesteps the federal “dwelling” argument entirely. RCW 49.60.215 makes it an unfair practice to refuse anyone “lodging” in such a place.
Washington’s protected class list is also substantially broader than federal law. It includes race, creed, colour, national origin, citizenship or immigration status, sex, marital status, sexual orientation, gender identity, age, honourably discharged veteran or military status, status as a mother breastfeeding her child, sensory, mental or physical disability, and use of a trained service animal.
So the working assumption for any Washington operator should be that you’re running a place of public accommodation under state law, with a wider protected class list than the federal one.
The principle that follows is simple to state and harder to hold to: screen the booking, not the person. Identity match, payment integrity, prior damage history, stated purpose, guest count against occupancy, third-party booking prohibition, risk patterns. Applied identically to every booking, every time. Anything that varies by who the guest appears to be is a problem regardless of intent.
There’s a further wrinkle I’d flag rather than pretend to resolve. Running criminal background checks through a consumer reporting agency can pull you into Fair Credit Reporting Act obligations around permissible purpose, disclosure, and adverse action notices. Whether a nightly booking constitutes a permissible purpose under FCRA is not settled. It’s a question for counsel, not for a blog post.
Noise monitoring, and why Washington law shapes the hardware
Detection covers what booking-time screening can’t: the stay that turns.
The devices in common use are Minut, NoiseAware, and Party Squasher, and they work differently. Minut and NoiseAware measure sound pressure and use duration thresholds rather than instant triggers. Minut’s default is 75 dB in daytime and 70 during quiet hours, sustained continuously for ten minutes before the first alert, with the timer resetting if it drops. Party Squasher doesn’t use sound at all. It counts nearby mobile phones, several times a minute.
Party Squasher publishes a caveat worth knowing: it’s “less suitable for townhouses, apartments, or condos, especially in multi-unit wood-framed buildings,” because it can’t reliably tell your space from your neighbour’s. In detached homes it works well.
Now the Washington-specific part, which explains why every credible device in this category is engineered the way it is.
Washington is a two-party consent state, and it’s a criminal statute. RCW 9.73.030 makes it unlawful to intercept or record any “private communication” or “private conversation” without the consent of all participants. Consent is obtained by announcing to all parties, “in any reasonably effective manner,” that recording is about to occur.
A device that derives only a decibel level, and never captures, stores, or transmits audio content, isn’t intercepting a conversation, because there’s no communication content to intercept. That is precisely why the industry builds these things without audio capture and says so explicitly in their documentation.
I’d label that as the reason the market is built this way rather than as settled law. I found no Washington appellate decision applying RCW 9.73.030 to decibel sensors.
What I would treat as a clear warning: the moment a device captures audio content, the analysis changes completely, and RCW 9.73 carries criminal penalties. Smart speakers, video doorbells with audio, anything that buffers or streams sound. That’s real exposure in Washington and it’s easy to install without thinking about it.
Airbnb has its own rules layered on top. Indoor security cameras and recording devices are banned entirely, even when switched off or disconnected. Decibel monitors are permitted indoors because they don’t record audio, but they cannot go in bedrooms, bathrooms, or sleeping areas, and they must be disclosed. Washington’s voyeurism statute, RCW 9A.44.115, uses a reasonable-expectation-of-privacy standard rather than listing rooms, which is broader and less predictable than a bedroom-and-bathroom rule.
Our practice: common areas only, decibel and duration thresholds rather than anything with audio, disclosed in the listing and in the rental agreement, with the device’s no-recording specification on file.
AirCover is a guarantee, not insurance
This is where I’d most want an owner to read the actual terms rather than the marketing page.
AirCover for Hosts advertises $3M in host damage protection and $1M in host liability insurance. The Host Damage Protection Terms, last updated 5 February 2026, say some things the marketing doesn’t:
“These Host Damage Protection Terms are not an insurance contract. The Host Damage Protection guarantee is not insurance.”
Your recourse is “to the Responsible Guest in the first instance.” Airbnb pays after the guest doesn’t. The filing deadline is 30 days from the responsible guest’s checkout. You’ll see 14 days quoted all over the industry, including on pages that should know better. That reflects archived terms. The current terms say 30.
And then the exclusions, which are the whole reason for the next section:
- Acts of nature, including “weather-related events”
- Mold and microbial growth, specifically “any loss relating to mold, mildew, fungus, spores”
- Normal wear and tear
- Damage from lack of utilities due to external factors
Read that twice if you own a house in the Pacific Northwest. AirCover excludes weather events and mold. Those are the two things most likely to damage your property, and they’re the two things it structurally cannot cover, because AirCover is guest damage protection. There’s no guest involved in a November pipe burst.
Separately, Washington requires operators to carry primary liability insurance of not less than $1 million in the aggregate, under RCW 64.37.050, unless the platform provides equal or greater coverage. And RCW 64.37.040 requires platforms to warn operators that their personal policy “might not provide liability protection, defense costs, or first party coverage” for short-term rental use. That’s Washington’s own code acknowledging the gap between a homeowners policy and an STR operation, and I’ve almost never seen it cited.
The empty house in November
Here’s the part I think is genuinely under-covered, and it’s where I’d spend an owner’s attention if I only got one section.
Standard policy forms treat an empty house differently, and the trigger language is not what you’d expect.
Look at a standard homeowners freezing exclusion. Here’s the language from a Safeco form, and near-identical wording appears across the industry:
“freezing of a plumbing, heating, air conditioning or automatic fire protective sprinkler system, or of a household appliance, or by discharge, leakage or overflow from within the system or appliance caused by freezing, while the dwelling is vacant, unoccupied or under construction… This provision does not apply if you have used reasonable care to: a. maintain heat in the building; or b. shut off the water supply and drain the system”
Note “vacant, unoccupied or under construction.” A furnished rental sitting between bookings is generally unoccupied even though it isn’t vacant, and this clause reaches both.
A separate clause in the same form excludes vandalism and malicious mischief “if the dwelling has been vacant for more than 60 consecutive days immediately before the loss.” On the commercial side, the standard ISO form CP 00 10 defines a building as vacant unless at least 31% of its square footage is rented and actively used, and after 60 consecutive days of vacancy it excludes vandalism, sprinkler leakage, glass breakage, water damage and theft outright, and cuts payment on everything else by 15%.
I want to be careful here, because this is where blogs oversimplify. “Vacant” and “unoccupied” are not the same thing, they’re treated differently, and the distinction gets litigated. Some clauses say vacant, others say vacant or unoccupied. Don’t take my word for what your policy does. Have your broker read your actual form.
But notice what the freezing exclusion gives back: it doesn’t apply if you used “reasonable care” to maintain heat or to shut off and drain the water. A manager with documented winterisation, logged temperature monitoring, and scheduled physical inspections is generating exactly the evidence that defeats that exclusion. That’s not a sales point, it’s the actual mechanism.
Why turning the heat down is the wrong instinct
The reflex when a house sits empty is to drop the thermostat and save money. In this climate that’s backwards, for two separate reasons.
Freeze risk. The commonly cited threshold, from Building Research Council work distributed by IBHS, is that uninsulated pipes in unconditioned spaces become at risk when outside temperatures hit 20°F or below. I’d treat that as a rule of thumb rather than a PNW-calibrated figure, since it was developed largely around southern building stock where plumbing sits outside the insulation envelope. It’s still relevant here, because a lot of older Puget Sound housing has plumbing in unheated crawl spaces and uninsulated exterior walls. Same failure geometry.
Tacoma Public Utilities recommends keeping the thermostat “no lower than 55°F” while away. IBHS’s position is that draining beats heating: shut the main, open all fixtures until water stops.
And mold, which is the quieter one. King County Public Health’s guidance is direct: “Heat all the rooms in your home to keep moisture from forming on the walls and other surfaces of unheated areas.” They frame the risk window as “the wet seasons of fall, winter, and spring,” and they also recommend flushing the air two or three times a day by opening doors and windows.
Which is impossible in an empty house. That’s the gap.
The EPA puts the target at 30% to 50% relative humidity, below 60% at minimum, and gives the response window: material dried within 24 to 48 hours of a leak usually won’t grow mold. An unmonitored leak in a vacant home in November has weeks.
So dropping the thermostat to 45°F in a humid maritime climate produces condensation on cold interior surfaces, in exactly the closets and exterior walls King County warns about, while simultaneously moving the building closer to the freeze exclusion’s trigger. Both failure modes at once, and AirCover covers neither.
The dates, which are not what Seattle owners assume
This is the single most useful fact I can give an owner with a property outside the city.
| Location | 50% chance of first 32°F |
|---|---|
| Seattle | 17 November |
| Leavenworth | 30 September |
| Bend, Oregon | 20 September |
All three are NOAA National Centers for Environmental Information probability dates, so they’re directly comparable. Seattle hits a 10% chance by 1 November and 80% by 29 November. Leavenworth is at 10% by 18 September and 80% by 8 October.
Freeze risk in the mountain and eastside markets begins roughly seven to eight weeks earlier than it does in Seattle. An owner in Leavenworth or Bend who winterises on a Seattle mental calendar is exposed through the whole of October.
Even inside western Washington it varies. A Snoqualmie Valley or North Bend property freezes earlier than a waterfront Ballard one, because Puget Sound moderates the shoreline.
And here’s the psychological trap. Seattle’s average winter low never actually drops below freezing. October averages 47°F, November 41°F, December 37°F, January 38°F. Freeze damage here isn’t a seasonal baseline, it’s an episodic cold snap. Which is exactly why owners in this region under-prepare relative to owners in genuinely cold climates. Nothing happens for four winters and then one week in January costs $40,000.
For scale: water damage and freezing accounted for 27.6% of homeowners claims in 2022, with average claim severity of $13,954 across 2018 to 2022, per ISO and Verisk data via the Insurance Information Institute.
What a shoulder season protocol actually looks like
Not a checklist for its own sake. This is what generates the documentation that matters if you ever claim.
- Thermostat held at a minimum, not turned off, with remote temperature monitoring and alerting
- Humidity monitored, with mold-risk alerting. The same sensor that handles noise usually does this
- Outside faucets shut and drained, hoses disconnected, irrigation winterised
- Whole-home leak detection with automatic shutoff where the property warrants it
- Physical inspection on a schedule, logged with dates, not “when someone’s passing”
- Every fixture run periodically so P-traps don’t dry out and release sewer gas
- Gutters and downspouts cleared, because needle fall here is relentless and water at the foundation is how this starts
- The whole record kept, because “reasonable care” is a thing you have to be able to show
That last point is the one owners skip. The exclusion is forgiven if you took reasonable care. Reasonable care you can’t evidence is worth much less when a carrier is reading the file.
Common questions
Does Airbnb prevent parties?
Partly. Airbnb runs a machine learning system year-round in the US and Canada that screens entire-home bookings for party risk, with heightened restrictions on high-risk dates, and it blocks tens of thousands of bookings around major holidays. Its limits: it screens at booking rather than during the stay, blocked guests are redirected rather than removed from the platform, and it doesn’t cover Vrbo, Booking.com or direct reservations. Vrbo’s equivalent flags risk to the host but, in Vrbo’s words, “does not block or cancel bookings.”
What does guest screening actually check?
Identity verification against government ID with a liveness check, watchlists of guests with prior unpaid damage, disposable email and burner phone detection, payment and chargeback fraud signals, and criminal background checks. The criminal check is US guests only with a seven-year lookback, so international guests get none. Behavioural signals like local bookings, last-minute one-night stays and third-party reservations predict party risk better than criminal history does.
Can I refuse a booking I’m not comfortable with?
You can decline for booking-related reasons applied consistently to everyone. You cannot decline on protected characteristics. In Washington that list is broader than federal law, because RCW 49.60.040 expressly covers “lodging of transient guests” as a place of public accommodation. Worth knowing separately: Airbnb counts how often you decline requests as a search ranking factor.
Are noise monitors legal in a Washington rental?
Devices that measure decibel levels without capturing audio are what the industry uses, and that design exists because Washington is an all-party consent state under RCW 9.73.030, a criminal statute. Airbnb permits decibel monitors indoors but bans all indoor cameras and recording devices, prohibits monitors in bedrooms, bathrooms and sleeping areas, and requires disclosure. Anything that captures actual audio is a different legal question entirely. This isn’t legal advice.
Does AirCover cover a burst pipe?
No. AirCover’s Host Damage Protection is not insurance, it’s a guarantee against guest-caused damage, and its terms exclude weather-related events and mold explicitly. A freeze or a moisture failure in an empty house has no responsible guest, so there’s nothing for it to attach to. That’s what your property policy is for, subject to whatever the vacancy and freezing clauses in your specific form say.
When should I winterise a Pacific Northwest vacation rental?
Earlier than most owners think, and it depends heavily on where the property is. Average first freeze is around 21 November in Seattle, around 30 September in Leavenworth, and around 20 September in Bend. An owner applying a Seattle calendar to an eastside or mountain property is exposed for roughly two months.
If you own a home that sits empty through the shoulder season and you’re not certain what’s actually being done to protect it, that’s worth a conversation. We’re happy to walk through what we’d do, whether or not you end up hiring us.
Recreation Stays manages vacation rental properties in Seattle and select Pacific Northwest markets. For a property-specific income estimate, visit our rental income calculator or reach out directly.
Schedule a Consultation | Get Your Revenue Analysis | See Our Portfolio
With 25+ years in luxury hotels and vacation rentals, Adam has led operations for brands like Fairmont and St. Regis and built high-performing hospitality businesses from the ground up. Today, as Founder & CEO of Recreation Stays, he brings that same expertise to helping owners unlock maximum returns while delivering five-star guest experiences. He’s also the host of The Proven Principles Hospitality Podcast, where industry leaders share what works in modern hospitality, and was recently recognized as one of the Top 100 Most Powerful People in US Hospitality.