Buy a house for $350,000 and the price arrives as one number. The tax code does not see one number. It sees a roof, a dishwasher, a driveway and a row of shrubs, and it writes each of them off over a different number of years.
A cost segregation study is the work of pulling that one price apart.
What does a cost segregation study actually do?
The IRS describes the mechanics in its Cost Segregation Audit Technique Guide, published as Publication 5653 and revised in February 2025. To depreciate anything, the guide says, you need the correct method and the proper recovery period for each asset you own, and a property usually holds numerous asset types with different recovery periods. When the actual cost of each item is known, that is simple bookkeeping. When all you have is a lump sum, cost estimating techniques are needed to segregate or allocate the total across the individual assets. That exercise is the study.
Whatever method a study uses, the guide says it should always do three things: classify assets into property classes such as land, land improvements, building, equipment, furniture and fixtures; explain the rationale, with legal citations, for calling each asset section 1245 or section 1250 property; and substantiate the cost basis of each asset while reconciling total allocated costs back to total actual costs.
The guide states one caveat about itself: it is not an official IRS pronouncement and may not be cited as authority. It is written for examiners, which is why it tells you what an examiner will look for.
Which parts of an Airbnb sit on a shorter schedule?
The building is the slow part. Publication 946 sets the recovery periods: 27.5 years for residential rental property, 39 years for nonresidential real property, and 5, 7 or 15 years for the shorter classes. Appliances, carpets and furniture used in a residential rental activity are 5-year property. Office furniture and fixtures are 7-year. Improvements made directly to land or added to it, such as shrubbery, fences, roads, sidewalks and bridges, are 15-year.
Publication 5653 goes further for rentals specifically. Its Exhibit A is a matrix of assets common to residential rental property with the classification and recovery period the IRS recommends for each. These rows show how fine the line gets.
| How the IRS classes it | Recovery period | |
|---|---|---|
| Appliances: electric range, refrigerator, dishwasher, clothes washer, dryer | Section 1245, Asset Class 57.0 | 5 years |
| Awnings and canopies that are readily removable | Section 1245, Asset Class 57.0 | 5 years |
| A false balcony, there for looks and nothing else | Section 1245, Asset Class 57.0 | 5 years |
| Fences, gates, retaining walls, sidewalks, landscaping and shrubbery | Asset Class 00.3, land improvements | 15 years |
| Kitchen cabinetry, counters and sinks | Section 1250 | 27.5 years |
| A balcony occupants can walk out onto | Section 1250 | 27.5 years |
| Interior ceilings, whatever the finish | Section 1250 | 27.5 years |
Read the first and fifth rows together. A refrigerator is 5-year property and the counter beside it is 27.5-year property. That is the argument over cost segregation in two lines, and why studies get written by people who know construction as well as tax law.
Which recovery period the building itself takes depends on a test about income, not about nightly rates. Publication 946 defines residential rental property as any building where 80% or more of its gross rental income for the tax year comes from dwelling units, and it says a dwelling unit does not include a unit in a hotel, motel or other establishment where more than half the units are used on a transient basis. Whether a particular short-term rental lands inside that definition or outside it is a question for a tax professional who can see the property and the bookings.
What did the 2025 law do to bonus depreciation?
It made the 100% write-off permanent for property acquired after 19 January 2025. Treasury and the IRS set out the rules in Notice 2026-11, announced on 14 January 2026, which covers the permanent 100% additional first year depreciation deduction provided by the One, Big, Beautiful Bill. The notice also lets a taxpayer elect 40% instead of 100% for qualified property placed in service during the first tax year ending after 19 January 2025.
The limit is the part that matters here. Publication 946 says property is qualified property for that allowance if it is tangible property depreciated under MACRS with a recovery period of 20 years or less. The 5-year appliances qualify. The 15-year sidewalk qualifies. The building, at 27.5 or 39 years, never does. So bonus depreciation only reaches whatever a study manages to move off the building, which is why the two subjects are always discussed together.
Does any of this change the cash in your pocket?
No. Depreciation is a deduction, not a bill. Here is the same example Airbnb a cost segregation study would be run on: a $350,000 house, 20% down on a 30-year loan at 7%, $25,000 of furnishing, averaging $375 a night and booked 65% of the year, with $4,200 of property tax and $2,400 of insurance.
| Each month | |
|---|---|
| Booking income | $7,415 |
| Platform fee (15.5%) | −$1,149 |
| Management / co-host (20%) | −$1,483 |
| Cleaning | −$300 |
| Supplies | −$150 |
| Maintenance (5%) | −$371 |
| CapEx reserve (5%) | −$371 |
| Utilities | −$350 |
| Property tax and insurance | −$550 |
| Mortgage | −$1,863 |
| Monthly cash flow | +$828 |
These are example figures, not a forecast for any market.
That $828 a month, $9,934 a year, is the money that lands in the account. It takes $105,500 of cash to set the house up. A study changes neither figure. What it changes is the taxable income reported against that cash flow, and therefore the tax owed on it. Two houses can clear the same $828 and owe different amounts of tax.
A study also costs money, and no standard price exists because no standard study exists. Set the fee against the tax it would save in the years you plan to hold the house, not against the size of the deduction.
What makes a study hold up?
Publication 5653 lists 13 principal elements of a quality study. Among them: preparation by someone with expertise and experience, a detailed description of the methodology, appropriate documentation, interviews with the right parties, an explanation of the legal analysis, a reconciliation of total allocated costs to total actual costs, and an identification and listing of section 1245 property. The guide notes there are no prescribed qualifications for preparers, and that a study by a construction engineer is generally more reliable than one by someone with no engineering or construction background.
Exhibit A carries the clearest signal of all. The guide tells examiners that where a return treats these assets consistently with the matrix, they should not adjust the categorization or the recovery periods, and that where a return treats them differently, adjustments should be considered. A study that departs from the matrix is not automatically wrong. It is automatically an argument.
Before you pay for a study
Three things decide whether the study is worth the fee, and none of them are the study itself.
Whether the deduction can offset your other income is the first, and for short-term rentals it usually turns on how long guests stay. The short-term rental tax loophole covers the 7-day test that question runs through. Second is how long you will hold the house, because accelerated depreciation is timing, not forgiveness, and depreciation recapture is what the sale does with it. Third is how much of your spend is already short-lived property: a host who put $25,000 into furniture, linens and a full kitchen has much of that on a 5-year schedule without anyone studying anything.
For how the plain 27.5-year method works before any of this is layered on, see rental property depreciation. Then take the question to a CPA, with the matrix in hand. Nothing above is tax advice, and the facts of one property decide more than any general rule does.