Should I do cost segregation? An honest, numbers-first guide.

Cost segregation is a tax study, not a tax shelter. An engineer walks through your property, identifies the components that depreciate faster than the building shell — appliances, flooring, landscaping, specialty lighting — and reclassifies them from 27.5- or 39-year property into 5-, 7-, and 15-year buckets. With 100% bonus depreciation permanently restored under OBBBA for property placed in service after January 19, 2025, every reclassified component generates a large first-year deduction. That's it. That's the whole thing.

Whether the deduction is useful to you is a separate question, and the part most articles skip. The IRS doesn't let you take losses willy-nilly. Passive activity rules §469 wall off rental losses from W-2 income unless you qualify under one of two narrow doors: real-estate-professional status, or the short-term-rental exception. We'll cover both.

The rest of this guide is structured the way an honest CPA would explain it to you: what the study mechanics are, what the math looks like for properties at different price points, the two tax-code doorways that let you actually USE the loss, the lookback option for properties you've already owned for years, the recapture trap on the back end when you sell, and a final section on the cases where cost seg is genuinely not worth it.

When it pays for itself

The break-even is shockingly low. At a $1,495 study fee and a 35% marginal rate, you only need about $4,300 of accelerated depreciation in year one to cover the cost. Most properties over $400K clear that on day one. The real question isn't whether the math works — it's whether your tax situation lets you use the loss.

Six scenarios, six verdicts

Section 03 covers the STR loophole in depth — the rule, its history, and the single sentence in Treas. Reg. §1.469-1T(e)(3)(ii) that high-W-2 doctors and engineers should understand before doing anything else.

The STR loophole, explained

The STR "loophole" isn't actually a loophole. It's an exception buried in a 1986 Treasury regulation that has been there, in plain sight, for nearly four decades. The relevant text is Treas. Reg. §1.469-1T(e)(3)(ii), which says that an activity is NOT a "rental activity" — and therefore not automatically passive under §469 — if the average period of customer use is seven days or less.

Read that again. The default rule under §469 is that all rental real estate is passive. Passive losses can only offset passive income. If you're a W-2 doctor with a long-term rental, your $50,000 cost-seg loss can't touch your W-2 — it sits suspended on your tax return until you have passive income or until you sell the property. That's the trap most articles don't explain.

But if your property has an average customer stay of 7 days or less — Airbnb, VRBO, vacation rental, corporate housing under a week — Treas. Reg. §1.469-1T(e)(3)(ii) says the activity is NOT a rental activity for §469 purposes. It's treated like a trade or business. And losses from a non-passive trade or business CAN offset W-2 income, provided you also materially participate.

Why "average ≤ 7 days" matters

The 7-day test is calculated on average, not on every booking. A property that books for a 14-day stretch in summer and 36 separate 2-night weekend stays during the rest of the year will easily average under 7. A property that rents long-term in winter and weekly in summer probably won't. Run the math on your actual booking history before you commit to the strategy.

Note: Treas. Reg. §1.469-1T(e)(3)(ii)(B) also recognizes a 30-day-or-less average IF significant personal services are provided (think hotels, with daily housekeeping and a front desk). Most vacation rental hosts don't qualify under that prong because they don't provide hotel-level services — but full-service short-term rental businesses sometimes do.

REPS — the 750-hour test

The other doorway out of passive is Real Estate Professional Status under IRC §469(c)(7). Both of these tests must be satisfied:

1. More than half of all your personal services across all trades and businesses must be in real-property trades or businesses in which you materially participate.
2. You must perform more than 750 hours of services per year in those real-property trades or businesses.

Both tests are individual-level (not joint). For a married couple, only ONE spouse needs to qualify, but that one spouse must satisfy both prongs personally. Their hours can't be combined with the non-qualifying spouse's hours.

Why most W-2 professionals don't qualify: the first prong. If you work 2,000 hours per year as a doctor, attorney, or engineer, you'd need to log more than 2,000 hours in real estate — full-time hours on top of full-time hours — to pass the "more than half" test. It's mathematically possible (40 hours/week real estate plus 40 hours/week day job) but it's rare and audit-bait if not meticulously documented.

The two groups for whom REPS works cleanly: full-time real estate professionals (agents, brokers, property managers) and stay-at-home or semi-retired spouses who can dedicate their service hours to the real estate portfolio while the working spouse provides W-2 income to be sheltered.

Form 3115 lookback

If you've owned property for years without doing cost segregation, you can still claim the catch-up. You don't amend prior returns — you file IRS Form 3115 (Application for Change in Accounting Method) and take the entire missed depreciation as a §481(a) adjustment in the current tax year.

Mechanically: you commission a cost-seg study on your existing property. The study reclassifies components retroactively to the date you placed the property in service. The difference between depreciation taken (typically straight-line) and depreciation that should have been taken (cost-seg accelerated) is the §481(a) adjustment — it lands on Form 4562 in the current year as a single deduction.

Real-world example: a $750K rental purchased in 2020. Owner took straight-line for five years. A 2026 cost-seg study identifies $165K of components that should have depreciated faster. After accounting for what was already taken on those components straight-line through 2025, the §481(a) adjustment is approximately $94K. At a 35% bracket, that's a $33K refund check — for property the owner already had.

Depreciation recapture on sale

Here's the gotcha most cost-seg sales pages omit. When you sell, prior depreciation is "recaptured" — added back to your taxable gain. Under §1245, recapture on tangible personal property (the 5- and 7-year stuff cost seg reclassifies most of) is taxed at ordinary income rates, up to 37% federal. Under §1250, recapture on real property is capped at 25%.

So: cost seg accelerates deductions at your marginal rate (could be 22%, 32%, 35%, 37%) on the front end. On the back end, when you sell, the recapture comes back at ordinary rates for §1245 property. If you're in a higher bracket when you sell than when you took the deduction, you can actually lose money on the timing — the dreaded "negative arbitrage."

When recapture wrecks the deal

If you plan to sell within 2–3 years, recapture will claw back most of the upfront benefit. If you have a 5+ year hold, the time value of money on the upfront tax savings still wins comfortably. If you have a 10+ year hold, cost seg is almost always strongly positive even with recapture.

Two strategies that defuse recapture: (1) 1031 exchange into a like-kind property defers the recapture indefinitely; (2) holding until death gets a stepped-up basis under §1014 and erases recapture entirely. Both are real plays for serious investors with long horizons.

When it doesn't work

We'd rather lose your business than sell you a study you don't need. These are the cases where cost seg is genuinely a mistake or a wash:

• Property under $100,000. Even at our $495 entry-level study fee, the math gets thin. Below this floor, just take straight-line.
• You don't materially participate AND you're not REPS. Losses get suspended as passive. They roll forward unused, sometimes for many years. The cost-seg "savings" are paper savings until you have offsetting passive income.
• Your AGI is already low or negative. Deductions can't go below zero. If you can't use the loss this year, you're paying for the study with no immediate benefit. (In some lookback cases this is still worth it, but talk to your CPA first.)
• You're selling within 2–3 years and not doing a 1031. Recapture eats most of the upfront benefit. Long-tail holds win this game.
• You bought 10+ years ago at a low basis. Most of the accelerated depreciation has already been claimed via straight-line. The §481(a) catch-up is small.
• The property is mostly land. Land doesn't depreciate. If your lot value is 60% of the purchase price (common in beach and coastal markets), cost seg has less depreciable basis to work with — though it can still work, the dollar yield drops.

Picking a provider

The cost segregation industry runs from $495 automated engineered studies (Cost Seg Smart) up to $8,000+ traditional Big Four engagements. Pricing variation does NOT track quality variation — it largely tracks overhead and sales-team cost. The IRS Audit Techniques Guide (Pub 5653) is the methodology standard, and any IRS-compliant study follows it whether the firm charges $495 or $8,000.

Three things to check on any provider:

1. IRS Audit Techniques Guide aligned. The provider should explicitly cite Pub 5653 methodology in their report. If they can't, walk.
2. RSMeans or equivalent cost data. Component-level pricing should pull from a recognized engineering cost database, not made-up numbers. RSMeans, Marshall & Swift, or BNi are the three industry standards.
3. Audit defense included. If the IRS questions the study within the statute, the provider should defend it free or at agreed-fee. A provider that disclaims audit defense is selling you a piece of paper.

See costsegregationreviews.com for our disclosed market guide ranking 25 providers, and costsegregationpricing.com for the 2026 pricing survey covering 147 providers.