Investing in Short-Term Rentals

On September 5, 2023, New York City began enforcing Local Law 18, and short-stay inventory in the largest hotel market in the country, 22,246 listings in August 2023 by Lodgify’s count, fell to about 4,000 within months (contemporaneous reports put the initial drop at more than 90%). Nothing had changed about the apartments; what changed was a registration rule: no whole-unit stays under 30 days, the host physically present, no more than two guests. Two years later, in December 2025, the City Council’s housing committee declined even to vote on a bill that would have loosened the rule for one- and two-family homes. The owners who bought Brooklyn brownstone floors in 2021 on the strength of $300 nights are still waiting.

That is the shape of this market in one scene. A short-term rental earns hotel revenue on a house, and for two years after the pandemic that looked like the easiest money in real estate: US listings grew 22.3% in 2022 alone, according to AirDNA, and revenue kept up because travel demand was recovering faster still. Then supply caught demand. Revenue per available listing, the number that pays the mortgage, fell in most markets through 2023, and the “Airbnbust” argument of that summer, in which one data vendor showed revenue down 47.6% in Sevierville, Tennessee and another showed it down 9.4%, was really about how much of the trade’s 2021 return had been a one-time reopening effect.

Investing in Luxury Real Estate, the hub’s flagship, covers the trophy house and its carry; Investing in 1031 Exchanges covers deferring the gain when you sell; Buying a Second Home Abroad takes the same purchase across a border, where the letting rules that squeeze a domestic host are often the reason a residency permit exists at all; and Investing in Opportunity Zones is the one regime that pays you to hold rather than to let. This one is about the house that is run as a business: the record from 2019 to 2026, the regulation map, the cost stack, the tax rules, the loan, and a $600,000 house run three ways.

Strip away the app and a short-term rental (STR) is a furnished dwelling let by the night to transient guests, usually for stays of under 30 days, and taxed and regulated as lodging rather than as housing. The owner is paid not for shelter but for a hotel room with a kitchen, and that distinction runs through the whole asset: the revenue swings with travel demand and seasons; the cost base is three to five times a landlord’s; and the licence to operate is a hospitality licence, which a city council can withdraw.

The venues are two and a half. Airbnb reported 148 million nights and seats booked in the second quarter of 2026, up 10% on a year earlier, on $27.2 billion of gross booking value, up 16%; its average daily rate rose 5%. Vrbo, owned by Expedia, is the second platform; Booking.com is the half, larger abroad than in the US. Around the platforms sit the data vendors, AirDNA and Key Data most prominently, and the managers, from Vacasa and Casago at the full-service end to Evolve and RedAwning at the marketing-only end.

On the revenue side, the other party is the traveller, whose alternative is a hotel: Key Data’s Vacation Rental Market Index for the second quarter of 2025 found short-term rentals earning revenue per available unit about nine percentage points ahead of hotels across 13 million tracked listings; the two products are priced against each other. On the asset side, the other party is every other host in your zip code. Supply is not fixed the way hotel rooms are: any homeowner with a spare bedroom can enter in a week, and several hundred thousand did in 2021–2023.

Price discovery is nightly and algorithmic: pricing tools reprice against comparable listings and hotel rates, so a host has little pricing power beyond the quality of the listing. Revenue per available listing (RevPAR: nightly rate multiplied by occupancy) is the number that matters, not either component alone. A market can post a rising average daily rate and falling revenue at once if occupancy falls faster, which is what much of the country did in 2023.

Measure against 2019. AirDNA puts pre-pandemic US occupancy at about 57%, and 2019 is the last year in which supply and demand grew at the same pace. Then 2020 removed 16.1% of demand, by AirDNA’s count at the time, and 2021 gave it all back and more: demand in nights passed 2019 levels in April 2021, and occupancy reached an all-time high of 60.8% for the year because supply had not yet caught up. That is the year the marketing still quotes; it was a reopening effect, not a run-rate.

Supply answered. AirDNA recorded available listings above 1.39 million in July 2022, a record at the time, and supply growth of 22.3% for the year; in 2022 US short-term rentals earned more than $62 billion and more than two million listings received at least one booking. The count of available listings peaked at 1.64 million in September 2023. Demand grew but could not match that; occupancy fell back through 2022 and 2023 toward its pre-pandemic level and revenue per listing fell with it.

AirDNA’s 2024 outlook, published in December 2023, called the result a “new equilibrium” after 2023’s 4.9% RevPAR decline, the first since 2014. Its 2025 outlook, published December 5, 2024, put 2024 supply growth at 6.9%, down from the 22.3% peak, and forecast 2025 demand growth of 4.9% against supply growth of 4.7%, enough to lift RevPAR 2.9%. That is roughly what happened: through 2025 the monthly reviews had year-to-date demand growing 5.7% against listings growing 4.6%, occupancy at 54.9% for the first half and 55.5% by autumn, and August 2025 setting a RevPAR record on nightly rates despite a slight dip in occupancy. October’s review was titled around uncertainty and high costs: revenue growing at the rate of inflation, fixed costs faster.

AirDNA’s 2026 outlook, released December 16, 2025 under the headline that 2026 would be the best year to invest since 2021, forecast listing growth of 4.6%, average daily rates up 1.5% and occupancy easing by about 1%. A midyear update released July 8, 2026, “Steady Demand and Slower New Supply”, cut supply growth to 2.7%, matched by demand, forecast RevPAR up 2.9% as nightly-rate growth accelerated from 0.7% in January to about 3% by spring, and put 2026 occupancy at 57.4% against a 57.0% pre-pandemic average. In AirDNA’s telling, supply slowed because renewed inflation and higher energy prices pushed mortgage rates back above 6% and would-be hosts stopped buying, not because the trade got worse.

The monthly reviews (54.9–55.5%) and the outlook (57.4%) measure occupancy on different bases; treat 55% as the working national occupancy for underwriting and 57% as the ceiling for a market without a supply problem.

Key Data’s series, built from professionally managed portfolios rather than scraped listings, tells the same story from the top of the market. Its second-quarter 2025 index had short-term rentals beating hotels on RevPAR in every region, with gains of 11% in the Mid-Atlantic, 10% in New England, 9% in the Rocky Mountains and 6% in Hawaii, and it described large managers as deliberately trading occupancy for rate. That is the professional response to a glut, and it is why national averages flatter the owner of one house, who cannot run a portfolio’s yield strategy.

What the record does not show is a total return: AirDNA and Key Data measure operating revenue, not owner returns after costs, debt and taxes, and neither captures the value of the house. The worked example builds that return and compares it with the same house on a lease.

On June 27, 2023, Nick Gerli of Reventure Consulting posted a chart under the words “The Airbnb collapse is real”, showing revenue per available listing down more than 45% year over year in Phoenix, Austin and Sevierville, Tennessee, and warning of forced selling by over-leveraged hosts. The chart used data from AllTheRooms. Within days AirDNA published its own May 2022 to May 2023 comparison and found the average market down 3.6%, not 40.3%: where Gerli’s chart had New Orleans down 37.1%, AirDNA had it down 1.2%, and Sevierville was down 9.4% on AirDNA’s count against 47.6% on AllTheRooms’. The vendors argued over how AllTheRooms counted listings that had never been booked and how it handled the 2022 comparison base.

Who was right? On the aggregate, AirDNA: there was no wave of forced sales in 2023 or 2024, and the national revenue series recovered to a record in August 2025. On the individual host, Gerli had a point the averages hid. A market down 9.4% on average contains listings down 40% (the 2022 entrants with no reviews, priced by an algorithm into a crowded field) and listings up 5% (the established houses with hundreds of reviews). Supply gluts do not lower the average much; they redistribute revenue from new entrants to incumbents, which is the Key Data occupancy-for-rate finding seen from the other end.

Two lessons survive. Ask any vendor how it defines “available”, and put 2019 next to any “down 30% from peak” figure, because the 2021 base was a reopening anomaly. And the vendors are participants: AirDNA sells data to buyers and lenders and benefits from a healthy-looking market; a housing-crash analyst benefits from the opposite. Use both, date both, underwrite on the lower.

An owner is paid three ways: operating income, which the worked example puts at 2.6% to 4.9% of price for a well-chosen house and below zero for a badly chosen one; the change in the house’s value, which has nothing to do with the bookings; and the tax treatment, which has quietly become the reason high earners buy these and which the tax section sets out. The middle source is the one that has stopped helping.

Through 2020–2022 the house’s price did most of the work: a host who bought in 2020 could show a 30% gain before a single cleaning fee was counted. That is over. On our own tape, the Invest Alternative housing sub-index (Parcl Labs’ national price feed, a 12% weight in our composite) stood at 97.746 on September 8, 2026, up 1.89% over 30 days but down 2.25% since we began recording it on June 30, 2026, ten weeks earlier. The Parcl Labs USA price-per-square-foot series behind it went from $463.57 on June 30 to $453.12 on September 7, 2026, a 2.3% fall in ten weeks. Our Zillow top-tier series, the 65th to 95th percentile of US home values, read $715,426 in July 2026. The national house is flat to slightly down; a short-term rental bought in 2026 should be underwritten on operating income and tax alone, with appreciation as an unpaid-for option.

Both are national series, and a cabin in a resort county can move against them in either direction. We carry them because a national series at −2.25% over ten weeks removes the most flattering assumption from every pro-forma an agent will show you.

Regulation is the one risk in this asset that takes revenue to zero without any change in demand. Six jurisdictions since 2023 show the forms it takes.

New York: the near-ban that held

Local Law 18, passed by the City Council in 2022 and enforced from September 5, 2023, requires hosts to register, forbids whole-unit stays under 30 days, requires the host to be present, and caps guests at two. Short-stay listings fell from 22,246 in August 2023 to about 4,000 by May 2024, an 82% fall on Lodgify’s one-year count. The bill to soften it, Intro 1107, introduced by Council Member Farah Louis on November 13, 2024 with Airbnb’s backing, would allow under-30-day stays in one- and two-family homes with no host present. The resistance is stronger: Skift reported on December 18, 2025 that the housing committee would not bring the bill to a vote, it was filed at the end of the session without a hearing, and as of spring 2026 no successor had advanced. A ban, once passed, is hard to unwind, because the hotel industry, the housing advocates and the neighbours all prefer the new status quo.

Dallas: the ban that the courts stopped

In June 2023 the Dallas City Council passed two ordinances removing short-term rentals from single-family zoning, about 90% of the city’s inventory. The Dallas Short-Term Rental Alliance sued and won an injunction from District Judge Monica Purdy. The Fifth District Court of Appeals ruled against the city in February 2025 and, on July 18, 2025, affirmed most of the temporary injunction (reversing it in part) on the owners’ constitutional property-rights claim; the city’s further attempts to lift it failed, and on October 16, 2025 Dallas petitioned the Texas Supreme Court, arguing it needed the ban in force before the FIFA World Cup matches of June 2026. The rentals kept operating throughout. Texas is the friendliest large state for hosts, as the next section explains; Dallas shows the price, three years under an injunction a higher court could lift.

Barcelona: the phase-out with a date

On June 21, 2024, Mayor Jaume Collboni announced that Barcelona would not renew any of its roughly 10,100 tourist-apartment licences and that all would expire by November 2028. The People’s Party challenged the underlying Catalan decree; Spain’s Constitutional Court rejected the appeal in ruling 64/2025 of March 13, 2025. By 2026 the city had issued no new licences and more than 3,500 apartments had returned to the residential market. Nationally, in May 2025 Spain’s Ministry of Social Rights, Consumer Affairs and the 2030 Agenda ordered Airbnb to block 65,935 unregistered or mislabelled listings; the Madrid High Court refused Airbnb’s appeal in June 2025, the platform had removed them by the end of July, when the ministry named a further 55,000, and in December 2025 the ministry fined Airbnb €64 million. Barcelona is the template: a phase-out on a fixed date with court approval.

Florida: local control by veto

Florida’s 2011 statute (Fla. Stat. §509.032(7)(b)) pre-empted local bans but grandfathered ordinances adopted on or before June 1, 2011, which is why Miami Beach can still prohibit the product while most of the state cannot. Senate Bill 280 in 2024 would have moved regulation to the state; it passed the Senate 23–16 and the House 60–51, and Governor DeSantis vetoed it in June 2024, writing that it “prevents local governments from enforcing existing ordinances or passing any new local measure which would exclusively apply to vacation rentals.” Counties and cities still write the day-to-day rules and the state licenses the unit through the DBPR. Florida remains the most STR-friendly large coastal state, and the one where insurance has done what regulation could not.

Hawaii: the phase-out with 7,000 units in it

Honolulu’s Bill 41 (Ordinance 22-7, 2022) confines rentals under 90 days to resort zones, with rentals lawfully operating on a 30-day minimum when it passed grandfathered. Hawaii County, the Big Island, adopted Bill 47 (Ordinance 25-50), which from July 1, 2026 requires every short-term rental to register, at $250 a year hosted and $500 unhosted, with fines of up to $10,000 for operating unregistered.

Maui went further. On December 15, 2025 the County Council passed Bill 9 by 5–3 and Mayor Richard Bissen signed it the same day as Ordinance 5909, ending transient vacation rentals in apartment-zoned districts, the so-called Minatoya list of roughly 7,000 units (6,208 of them actively let short-term as of May 2024): West Maui by January 1, 2029 and the rest of the island by January 1, 2031. A proposed class action, Lynam v. County of Maui, was filed on December 22, 2025 arguing a regulatory taking, and in June 2026 the council created new hotel districts (Bill 88, Ordinance 6008) that let about 2,056 of the units keep operating, over the objection of all three planning commissions (Star-Advertiser, July 5, 2026). A Kihei or Kaanapali condo’s exit window is now set by that litigation and that rezoning, not by the tourism cycle.

The 2025–2026 pattern elsewhere

Los Angeles has limited hosting to a primary residence with a 120-night cap (more only with an extended home-sharing permit) since July 2019, and New Orleans moved in 2023 to resident-operator rules and per-block limits. The 2025–2026 ordinances follow the same three templates: primary-residence-only, resort-zone-only, or a licence cap. A city with a housing shortage and a hotel lobby will regulate; a resort county whose economy is the rentals will license and tax.

The private regulator most buyers forget is the homeowners’ association. The Community Associations Institute’s 2025 statistical review puts 35.2% of the nation’s housing stock, home to about 78 million people, inside a community association, and in the Sun Belt subdivisions and resort condominiums where short-term rentals cluster it is far higher. The declaration of covenants runs with the land and can do what a city cannot: prohibit short-term letting outright, require a minimum lease term, or cap the number of rented units. The city’s permit is irrelevant if the covenants say no.

Texas is the reference point because of Tarr v. Timberwood Park Owners Association, decided unanimously by the Texas Supreme Court on May 25, 2018. Kenneth Tarr let his San Antonio house by the night against covenants restricting it to “residential purposes” and single-family occupancy. The court held that so long as the occupants used the home for a residential purpose, “no matter how short-lived,” neither their use nor Tarr’s renting violated the deed restrictions: a generic residential-use clause does not ban short-term rentals. What followed is the part that matters: associations across Texas and beyond amended their declarations to bar leases under a stated term or require owner occupancy, by the supermajority vote their documents specify, and courts have upheld such amendments against owners who bought before the vote.

You can therefore buy a house whose covenants permit nightly rentals and lose the right two years later at a meeting you did not attend, with no compensation. The diligence is not “does the HOA allow it” but: read the declaration for any lease-term restriction; read the amendment clause for the vote threshold (two-thirds with an active board is a live risk; 90% among absentee owners is not); read three years of minutes for “short-term”, “Airbnb” or “transient”; and ask the manager in writing whether a rental amendment has been proposed. Florida condominium boards can also adopt “reasonable” rules by board vote, without any owner vote at all.

The gap between gross bookings and net income is where most first-time projections fail. These are the lines between the guest’s payment and your bank account.

Platform fees

Airbnb has spent 2025 and 2026 moving hosts from its split fee, under which the host paid about 3% and the guest paid a service fee of roughly 14.1–16.5% at checkout, to a single host-only fee of 15.5% of the booking subtotal (nightly rate plus cleaning and guest fees, before taxes), deducted from the payout. Software-connected hosts were moved from October 27, 2025 and fully migrated by April 13, 2026; the rest move by September 15, 2026 outside the European Economic Area and October 13, 2026 inside it. The two structures cost the guest about the same, so the migration is neutral if you raise your listed price to match and expensive if you do not: a host who leaves the rate unchanged gives up about 12.5 points of revenue.

Vrbo’s pay-per-booking model charges a 5% commission on the rent and mandatory fees plus a 3% payment-processing charge on the whole payment, about 8% all in, per its published help page (2026). Direct bookings cost only the card-processing fee, which is why experienced operators aim their marketing at the repeat guest.

Management

Full-service management (listing, pricing, guest messaging, cleaning coordination, maintenance calls and owner statements) has a 2026 national average of roughly 20–25% of gross bookings, and the large brands run higher: Vacasa does not publish a rate, and 2026 comparisons put its contracts at 25–35% and higher once add-on services are counted. The “half-service” tier (Evolve, RedAwning: marketing and booking only, with cleaning and maintenance left to you) runs 10–15%; Evolve publishes 10% and 15% tiers, RedAwning 10%, 15% and 18%. Individual co-hosts charge 10–25%. Management is the largest controllable cost in the stack; self-management, roughly five to ten hours a week per property, changes the return by more than any other decision.

Cleaning, supplies and turnover

The guest pays a cleaning fee and the owner pays the cleaner; the two rarely match, and the fee carries the 15.5% platform charge on the way through. Linens, consumables and replacing what guests break add a few thousand dollars a year. Utilities, internet and, in a resort, hot-tub or pool service are the owner’s, not the tenant’s: the largest structural difference from a lease.

Furnishing

A three-bedroom house furnished to the standard that photographs well costs $25,000–$45,000, and the furniture lasts three to five years under guest use. Treat it as a $35,000 cheque at the start and a reserve of about 10% of it a year.

Insurance

A homeowner’s policy excludes business use and a standard landlord (DP-3) policy covers premises liability only, so both leave a host uninsured for the claims that happen: guest injury, guest-caused damage, and lost income while the house is unlettable. A purpose-built policy from a specialist such as Proper Insurance starts from about $2,000 a year, and 2026 guides put the typical host at $1,500–$3,000 before wind or flood cover in coastal states, where the luxury guide’s insurance section applies. Airbnb’s AirCover is a host guarantee with exclusions, not a policy.

Property tax, occupancy tax and the rest

Property tax follows the state: the Tax Foundation’s 2026 table of effective rates runs from 0.29% in Hawaii to 1.88% in New Jersey and Illinois, with Texas at 1.40%, and a non-homestead Florida property reassesses under a 10% cap rather than the 3% homestead cap. Lodging taxes, charged to the guest and remitted by you, are state sales tax on transient stays plus a county or city occupancy tax, commonly 10–15% of the nightly rate together (Florida: 6% state sales tax on transient rentals under §212.03, a county tourist development tax of up to 6% under §125.0104, and a discretionary county surtax of 0.5–1.5%). The platforms remit in many jurisdictions and not in others; the registration and filing obligation is yours either way.

Before debt, a managed short-term rental in a good market keeps about 23 cents of every dollar and a self-managed one about 43; a lease on the same house keeps about 50 cents of a smaller dollar.

Federal tax treatment is the reason a high earner buys a short-term rental rather than a duplex, and the same rules, misapplied, are the reason the IRS audits them. Four rules matter.

The 14-day rule

Section 280A(g) of the Code lets you rent a dwelling for fewer than 15 days in a year and exclude the income entirely, with no deductions. Above 14 rental days the property is a rental, and if your own use exceeds the greater of 14 days or 10% of rental days it is a “residence” under §280A(d)(1): deductions are limited to the rental income, with expenses allocated between personal and rental days under §280A(e). A family member’s stay is personal use unless it is for substantially full-time repairs. Owners who want the treatment below keep personal use under that ceiling; the 1031 guide’s vacation-home safe harbor (Rev. Proc. 2008-16) uses the same 14-day and 10% tests.

Schedule E or Schedule C

Rental income is reported on Schedule E and is not subject to self-employment tax. If you provide “substantial services” for the guests’ convenience, the sort a hotel provides, the activity is a trade or business on Schedule C and the net income carries the 15.3% self-employment tax up to the Social Security wage base and 2.9% above it. The IRS’s own examples (Publication 527) are regular cleaning, changing linen and maid service during the stay; cleaning and fresh linens between guests, utilities and Wi-Fi are generally treated as not substantial, but the line is factual, so get your preparer’s view in writing. Do not let a preparer put you on Schedule C to look like a business; the passive-activity result below does not require it.

The 7-day rule and the passive-loss regime

Section 469 treats rental activities as passive by definition: losses (including depreciation) can only offset passive income unless you qualify as a real estate professional, a test few salaried owners meet. Treasury Regulation §1.469-1T(e)(3)(ii)(A) carves out an activity in which the average period of customer use is seven days or less: it is not a “rental activity” at all for §469 purposes. Paragraph (B) does the same for an average stay of 30 days or less where significant personal services are provided. You compute the average by dividing rental days by the number of stays for the year, and a single month-long booking can push a property over seven.

Falling outside the rental definition does not make the losses deductible by itself: the activity is then an ordinary trade or business, and its losses are non-passive only if you materially participate under one of seven tests in §1.469-5T. The three that matter for an STR are more than 500 hours in the year; more than 100 hours and more than any other individual; or “substantially all” of the participation. A contemporaneous log is the evidence. A property under full-service management, where the manager’s staff plainly spend more hours than the owner, fails the 100-hour test, which is why the tax strategy and the management decision are the same decision.

Cost segregation and 100% bonus depreciation

A house is depreciated over 27.5 years as residential rental property, or over 39 years if it is transient lodging under §168(e)(2)(A)(ii); preparers split on which applies to a nightly-rented house. Either way the building is slow. A cost-segregation study reclassifies the parts of the property that are not the building (appliances, furniture, finishes, fencing, landscaping, driveways) into 5-, 7- and 15-year property, typically 20–30% of the depreciable basis on a furnished house. The One Big Beautiful Bill Act, signed July 4, 2025, restored 100% bonus depreciation permanently for qualified property acquired and placed in service after January 19, 2025 (property under a binding contract before January 20, 2025 stays on the old phase-down, 40% for 2025). Everything reclassified into 20-year-or-shorter property can be deducted in full in the year of purchase.

Put the three rules together and you have the “STR loophole”: segregate a quarter of the basis into short-life property, deduct it in year one, and if your average stay is seven days or less and you log more than 100 hours and more than anyone else, the loss offsets salary at your marginal rate. The worked example runs it in dollars ($45,360 of federal tax not paid in year one on a $600,000 house at 35%). The promoters say less about the rest: the study costs several thousand dollars, the deduction is a deferral recaptured when you sell, and the participation log is the first thing an examiner asks for.

Recapture and the sale

When you sell, gain attributable to depreciation on 5-, 7- and 15-year property is §1245 recapture at ordinary rates; gain attributable to straight-line depreciation on the building is unrecaptured §1250 gain at a maximum 25%; the rest is long-term capital gain at 0%, 15% or 20%. The 3.8% net investment income tax under §1411 applies to a passive activity and not to a business in which you materially participate. Deferring all of it through a like-kind exchange is the ground of the 1031 guide.

Conventional mortgages underwrite the borrower; a debt-service-coverage-ratio (DSCR) loan underwrites the house. The lender divides expected annual rent by annual debt service and lends if the ratio clears its floor, without tax returns, which is why DSCR loans, made by non-bank lenders and sold into securitisations, are the default for investors with more than one or two properties. As of mid-2026 the published range is 6.5–8% for most residential investment property, with the best-qualified borrowers in the low-to-mid 6s (one lender exchange quoted 6.12–6.49% in June 2026), against a Freddie Mac 30-year conforming average of 6.71% on September 3, 2026 (6.50% a year earlier). Short-term rentals carry an add-on of about 0.25–0.50%, because the lender haircuts projected STR income. Down payments are 20–25%; most lenders want a DSCR of 1.0–1.25, and sub-1.0 programmes exist at lower loan-to-value and a higher rate, one lender qualifying down to 0.75.

Two clauses matter more than the rate. The first is how the lender counts STR income: an appraiser’s long-term rent estimate (Form 1007), which fails the ratio on any resort house; twelve months of platform statements; or a discounted AirDNA projection, the vendor’s optimism lent back to you. The second is the prepayment penalty. DSCR loans price lower with a five-year step-down (typically 5% of the balance in year one, falling a point a year) than with a three-year or zero penalty, and an owner who must sell in year two because the city changed the rules pays it on top of everything else. Buy the shorter penalty wherever there is a regulatory question; it costs less than the exit.

A second-home conventional mortgage is cheaper for a first property, but its occupancy affidavit says you will use the home yourself for part of the year and will not put it under a management agreement that controls occupancy; a second-home loan on a plainly full-time rental is misrepresentation. The honest choice is the DSCR loan at its higher rate, and the arithmetic below uses it.

The screen has four filters, in the order the risks bite: regulation, then supply, then demand, then the house.

The first filter is legal and binary. A market is buyable only if a whole-home, non-owner-occupied rental under 30 days is permitted in the zoning where the house sits, the permit transfers or can be obtained, and no ordinance, ballot measure or lawsuit is pending. That rules out New York City, most of Los Angeles, Honolulu outside resort zones, Barcelona and apartment-zoned Maui; it puts Dallas and Austin in a grey category that pays a higher return for litigation risk; and it leaves the historical vacation markets (the Smoky Mountains, the Gulf Coast, the Ozarks, the Blue Ridge, the ski towns, the desert resorts), where the rentals predate the platforms and the county’s tax base depends on them. Buy where the product is the local economy, not the local controversy.

The second filter is supply against demand. For the submarket, not the metro, you want the twelve-month change in available listings from AirDNA or AirROI, the twelve-month change in revenue per available listing, and 2019 occupancy next to the current figure. A market whose listings grew faster than its revenue is one where you will be the marginal new entrant priced below the incumbents; a market whose listings shrank while revenue held is one where regulation or exhaustion has done your work for you.

The third filter is the shape of demand. Drive-to markets within three hours of two large metros held up through every shock since 2020; fly-to and single-event markets show the revenue concentration Key Data flagged in 2025, with the year’s income arriving in a handful of weekends. Seasonality decides cash flow: a cabin at 80% occupancy in October and 30% in March needs a reserve for March.

The fourth filter is the house, and here the STR buyer’s criteria invert the homebuyer’s: bedrooms count more than square feet, because the platforms sort by guest count; a hot tub or a view moves the nightly rate by more than a renovated kitchen; a two-minute walk to the beach beats a large lot. Test every dollar spent on the house against the nightly rate it adds, over the nights actually booked, after the 35 cents of platform and management on every dollar.

One purchase price, three uses: a short-term rental in a resort market, the same in a saturated metro suburb, and a twelve-month lease, with every cost and the tax overlay.

The inputs

The house costs $600,000 in each case, bought with 25% down ($150,000) and a $450,000 DSCR loan at 7.0% over 30 years, which costs $2,994 a month or $35,926 a year, of which $31,355 is interest in year one. Closing costs are $9,000. The short-term rental cases add $35,000 of furnishing, so cash in is $194,000 against $159,000 for the lease. Property tax is 1.0% of price ($6,000), insurance $2,500 on an STR policy and $1,800 on a landlord policy, maintenance 1% of price ($6,000); the lease case adds a $1,500 leasing and turnover allowance, so its fixed costs are $15,300. The STR cases add utilities and internet ($4,800), supplies ($1,800), a furnishing reserve ($4,000), software ($600) and, for the resort house, hot-tub and yard service ($2,400; $1,200 in the suburb). The platform fee is Airbnb’s 15.5% host-only fee and management is 20% of gross where used.

Market A is a drive-to resort market, a three-bedroom cabin an hour from a large metro: an average daily rate of $310 and occupancy of 60%, which is 219 booked nights and $67,890 of gross bookings. Market B is a four-bedroom house in a Sun Belt suburb where listings grew faster than demand: an ADR of $195 and occupancy of 52%, 189.8 nights (call it 190) and $37,011 gross. The lease is $3,300 a month, $39,600 a year, with 5% vacancy and 8% management. These rates and occupancies are our underwriting assumptions, set around AirDNA’s 2025 national occupancy of 55% with a resort premium and a saturated-market discount; they are not any vendor’s figure for a named market.

Year one, before tax

• Market A, managed: $67,890 gross; less platform $10,523 and management $13,578; less fixed costs $28,100; net operating income $15,689, or 2.6% of price; after debt service, cash flow of −$20,237.
• Market A, self-managed: the same $67,890 less platform $10,523 and fixed $28,100; NOI $29,267, or 4.9% of price; cash flow −$6,659.
• Market B, managed: $37,011 gross; less platform $5,737 and management $7,402; less fixed $26,900; NOI −$3,028; cash flow −$38,954. Self-managed, NOI is $4,374 and cash flow −$31,552.
• Long-term lease: $39,600 less vacancy $1,980 is $37,620; less management $3,010 and fixed $15,300; NOI $19,310, or 3.2% of price; cash flow −$16,616.

Three things stand out. Every case loses cash after a 7% mortgage at 75% loan-to-value, because none yields more on price than the debt costs. The resort STR beats the lease only if you run it yourself; managed, it trails the lease by $3,600 a year while carrying $35,000 more capital and a business’s worth of work. And the suburban STR should not be an STR at all: it grosses $2,600 less than the lease and costs $11,600 more to run before management.

The tax overlay

Now the part that makes the trade. Assume the owner earns a salary taxed at a 35% marginal federal rate, buys the resort cabin after January 19, 2025, commissions a cost-segregation study that reclassifies 27% of the $480,000 depreciable basis (land is 20%) into short-life property, and keeps average stays under seven days and a log of more than 100 hours. The short-life property, $129,600, is deducted in full in year one under the restored 100% bonus rule, and because the seven-day exception takes the activity out of §469’s rental definition, the resulting loss offsets salary. At 35% that is $45,360 of federal tax not paid in year one, more than twice the year’s cash shortfall. The remaining $350,400 of building depreciates at $8,985 a year on a 39-year life (or $12,742 on 27.5), worth about $3,100 a year at 35%. The $35,000 of furniture is left out of the shield to keep the example conservative; it is itself 5- and 7-year property, and expensing it under the same rule would add roughly $12,000 of year-one tax saved at 35%, recaptured on the same terms at sale.

The lease gets none of this against salary: its losses are passive under §469 and carry forward, unless the owner’s income is low enough for the $25,000 allowance under §469(i), which phases out entirely at $150,000 of adjusted gross income. The managed STR gets none of it either, because the manager’s staff out-hour the owner. So after tax in year one the self-managed resort STR is roughly $58,000 ahead of the lease (an after-tax gain of about $41,800 against the lease’s $16,600 shortfall).

Year five: the exit

Sell the resort cabin after five years at 2% annual appreciation, for $662,448, less 6% in commission and closing, netting $622,702; after the $423,592 loan balance the owner walks away with $199,110 before tax against $194,000 invested, plus five years of operating cash flow. Adjusted basis is $609,000 less $129,600 of bonus depreciation and $44,925 of straight-line, or $434,477, so the gain is $188,225. Of that, $129,600 is §1245 recapture at ordinary rates (35%: $45,360), $44,925 is unrecaptured §1250 gain at 25% ($11,231), and the $13,702 remainder is long-term gain at 20% ($2,740), a federal bill of $59,331.

The year-one shield was $45,360 and the five years of straight-line shield about $15,700, so the tax comes back almost exactly, five years later. That is what the loophole is: an interest-free loan from the Treasury, sized at your marginal rate times your short-life basis, repaid on sale unless you exchange under §1031.

Five failures take the whole investment rather than a year’s return.

The rule change

Every regulatory case in this guide is an owner whose revenue went to zero while the mortgage did not: the New York host in September 2023, the Maui condo owner with a date certain, the Dallas operator with a court calendar. The mitigants are the legal filter above, a prepayment penalty short enough to permit a sale, a house that works as a long-term rental at the same mortgage, and the discipline to sell when an ordinance is introduced rather than when it passes.

The platform

Airbnb can suspend a listing or an account for a guest complaint or a policy change, and the appeal process is opaque. A host with one listing on one platform has one customer, and the 2025–2026 fee migration shows that customer can change its terms on notice. Vrbo, direct bookings and a repeat-guest list are the diversification; none replaces the largest platform’s demand.

The counterparty

The two largest corporate failures in the sector fell in one year. Vacasa, which went public through a merger with TPG Pace Solutions in December 2021 at a $4.5 billion valuation, was sold to Casago on April 30, 2025 for $5.30 a share, about $130 million; the combined firm manages more than 40,000 homes, and owners saw fee schedules and service levels change. Sonder, which master-leased about 7,500 apartment and hotel rooms from landlords and let them by the night, announced an immediate wind-down on November 10, 2025, the day after Marriott terminated its licensing agreement, and filed for Chapter 7 liquidation in Delaware on November 14, leaving its landlords with unpaid rent and empty furnished units. If your income runs through a manager or a master tenant, their balance sheet is your risk, and the two largest in the sector could not make the economics work at scale.

The house

A short-term rental is a public accommodation: guests are injured, damage property and hold parties, and a homeowner’s or landlord’s policy denies the claim. Resort markets bundle location risk: the Gatlinburg wildfire of November 28, 2016 killed 14 people and destroyed some 2,500 structures in the Smoky Mountains’ largest rental market; Hurricane Ian on September 28, 2022 removed much of Fort Myers Beach’s rental stock; and Florida’s insurers have spent 2023–2026 repricing the rest. A resort house needs a named-storm or wildfire deductible you can actually pay and a reserve for the season it cannot be let.

The promoter

The STR loophole has a course-selling industry attached, and the failure mode is a buyer who takes the depreciation on a managed property, fails the participation test on examination, owes the tax with penalties, and sells into a saturated market with a prepayment penalty. There is no fraud in the rule; the fraud is in the pitch that it works without the hours. The related “rental arbitrage” pitch, master-leasing apartments to sublet by the night, is Sonder’s model at retail scale and ended the same way.

The cleanest exposure to short-term rental demand is Airbnb’s stock, which is not exposure to the asset at all: Airbnb takes its fee on $27.2 billion of quarterly bookings whether or not any host makes money, and its interest in host supply is that there be more of it, the opposite of an owner’s. Expedia holds Vrbo the same way. Neither gives you a house, appreciation or a depreciation deduction, and both are the better investment in a supply glut for that reason. For pooled ownership of the building itself, Investing in Real Estate Syndications and Private REITs on this hub covers the sponsor’s fees and the redemption gates, and Investing in Net-Lease Commercial Property covers the single-tenant building that pays a cap rate.

Fractional and crowdfunded vacation rentals exist and have a poor record. Here Collection LLC sold Regulation A shares in individual vacation rentals at $100 minimums in 2022–2023, per its SEC filings, and shut down on January 4, 2024, listing its houses for sale and returning the proceeds. Arrived, better known for single-family rentals, offers vacation-rental shares at $100 minimums with a 3.5–6% sourcing fee, property-management fees of 15–25% of gross on its vacation homes, expected holds of up to fifteen years, and a secondary market (launched November 2025) only after six months, per 2026 summaries of its terms. In every fractional structure the sponsor takes the management fee, the platform fee and a slice of the appreciation, and the investor takes the passive-loss regime with none of the seven-day benefits.

Co-hosting, running other owners’ listings for 10–25% of gross, is the way to learn the business with no capital. Master-lease arbitrage is the way to lose someone else’s capital. The syndications that hold portfolios of vacation rentals belong to the hub’s syndication guide; read their PPMs as hotel deals, because that is what they are.

Ranked honestly: own the house yourself in a permitted resort market and run it, or own Airbnb stock, or do neither. The managed middle, the fractional platforms and the master-lease operators are where the fees live.

In the order that saves the most money if you stop partway.

1. Decide whether you will run it. If the answer is no, stop here and buy a long-term rental or Airbnb stock; the worked example shows the managed STR trailing the lease, and the tax treatment is unavailable without the hours.
2. Pick three candidate submarkets through the legal filter first: whole-home rentals permitted by right, permit transferable or available, no pending ordinance or litigation, and two years of council minutes read for the word “short-term”.
3. Buy the data, not the pitch: an AirDNA or AirROI market report for each submarket showing twelve-month change in listings, revenue per available listing, 2019 occupancy against current, and seasonality by month. Reject any submarket where listings grew faster than revenue.
4. Underwrite a specific house with the cost stack in this guide, require net operating income above 4.5% of price self-managed, then run the same house as a twelve-month lease and require the STR to win.
5. Check the association before the inspection: declaration, amendment threshold, minutes, and a written answer from the manager on any rental amendment proposed.
6. Get the loan right: a DSCR loan on twelve months of statements or a discounted projection, the shortest prepayment penalty you can afford, and no second-home affidavit on a full-time rental.
7. Buy the policy, a purpose-built short-term rental policy with liability, guest damage and loss of income, plus wind or flood where relevant, before the first guest.
8. Register everything: the state licence, the county and city permits, the lodging-tax accounts, and the platform’s registration field.
9. Commission the cost-segregation study in the year of purchase if your marginal rate justifies it, and start the hours log on the day you close.
10. Run it for a year before judging it: the first year carries the setup, the second year is the business.

A short-term rental is a good small business and a mediocre investment. The operating return on a well-chosen, self-run house is 4–5% on price before debt, better than a lease and worse than a Treasury bill once the work is counted; the appreciation case is absent on our tape, with the national housing series down 2.25% between June 30 and September 8, 2026; and the tax case is real, large and temporary, worth about 7.5% of the price in year-one federal tax at a 35% rate and repaid at exit. That suits a high earner who wants to run a second business and will hold through a regulatory scare, and not the passive buyer the management industry is built to serve.

What would change the view, with thresholds:

• Supply. AirDNA’s national listings growth back above 6% for two consecutive quarters would return the market to the 2022–2023 pattern; below 2%, with demand above 3%, incumbents’ pricing power becomes the base case. The mid-2026 forecast is 2.7% for both.
• Occupancy. National occupancy (AirDNA monthly basis) below 53% for a full year is a glut; above 57% for a full year is 2021-style pricing power. It ran 55.5% by autumn 2025.
• Rates. A Freddie Mac 30-year average below 6% (it was 6.71% on September 3, 2026) would bring the second-home buyer back and restart supply growth within two quarters. DSCR pricing for short-term rentals below 6.5% closes about a quarter of the gap in our example; the self-managed resort cabin is still about $4,900 short at 6.5% on its $450,000 loan and breaks even at about 5.1%, so a DSCR quote near 5% is the reading that makes it cash-flow positive.
• Dallas. A Texas Supreme Court ruling for the city on the June 2023 ordinances would be the first appellate green light for a single-family-zone ban in a large Sun Belt city and would be copied within a year; a ruling for the operators would settle the Texas question for a decade.
• New York. A floor vote on the one- and two-family-home amendment to Local Law 18 would be the first reversal of a major ban; as of spring 2026 it has not left committee.
• Maui. The takings litigation over Bill 9 (Lynam v. County of Maui) and the hotel rezoning under Bill 88 set the value of every apartment-zoned unit on the island; a court striking the phase-out would reprice them by a third or more overnight.
• Airbnb’s fee and the tax rules. The split-fee retirement completes September 15, 2026; if fourth-quarter national ADR does not rise roughly 10% against trend, hosts absorbed it and the cost stack here is 12 points worse. Any bill narrowing the seven-day exception or reinstating a bonus-depreciation phase-down removes the reason the marginal 2025–2026 buyer buys.