What Is Cost Segregation in Real Estate? How It Works for Multifamily, What It Saves, and What It Costs Later

What cost segregation is

When you buy an income property, the purchase price is split between land, which is never depreciated, and everything else, which is. The default rule under IRC §168 treats “everything else” as a single asset: residential rental property depreciated straight-line over 27.5 years, nonresidential real property over 39 years. A $10 million apartment acquisition with $2 million of land therefore produces about $291,000 of depreciation a year, every year, for 27.5 years.

But the building is not a single asset. It is a structure plus thousands of components that the tax law has always classified separately when they are bought separately. Appliances are 5-year property. Carpet is 5-year property. A parking lot is a 15-year land improvement. The Tax Court settled in 1997, in Hospital Corporation of America v. Commissioner (109 T.C. 21), and the IRS accepted in its subsequent guidance, that a taxpayer may identify those components inside an acquired building and depreciate each over its own recovery period. A cost segregation study is the engineering and accounting analysis that does the identifying: it inspects the property, takes off quantities, assigns a cost to each component using construction-cost data or the actual construction records, and classifies each one under the asset classes of Rev. Proc. 87-56 (reproduced in IRS Publication 946). The IRS publishes its own Cost Segregation Audit Techniques Guide describing what it expects such a study to contain.

What the study does not do is create deductions. Total depreciation over the life of the property is identical with or without it. It moves deductions forward in time — and because of bonus depreciation, it moves a great many of them into year one.

The four buckets: 5-year, 15-year, 27.5-year, and land

5-year property (and some 7-year) is tangible personal property under §1245: items that are not structural components of the building. In an apartment community this means appliances, carpet and other removable floor coverings, window treatments, cabinetry and countertops in many studies, decorative lighting, the electrical and plumbing that serves specific equipment rather than the building generally, security and access-control systems, clubhouse and fitness equipment, signage, and furniture in common areas. The distinction between a structural component and personal property is the heart of every study and most of the audit controversy; the test is whether the item is permanently attached and essential to the building's operation as a building, or whether it serves a particular use and could be removed without damaging the structure.

15-year property is land improvements: depreciable things attached to the land rather than the building. Parking lots and drives, sidewalks, curbs, retaining walls, fencing, landscaping with a determinable life, irrigation, exterior lighting, swimming pools, playgrounds, dog parks, and the site utilities that run from the street to the buildings. On a suburban garden-style community spread across ten acres, this bucket is large.

27.5-year property is what remains: the foundation, framing, roof, exterior walls, windows and doors, elevators, HVAC that conditions the building, the plumbing and electrical that serve the building as a whole, fire protection, and interior walls, ceilings, and permanent floor coverings such as tile. For nonresidential property the same bucket is 39 years. Mixed-use buildings are classified by the 80% test: if 80% or more of gross rents come from dwelling units, the whole building is residential.

Land is not depreciable, and the study does not change the land allocation; that comes from the purchase agreement, an appraisal, or the county assessor's ratio. Owners occasionally hope a study will shrink the land number. A credible one will not.

Bonus depreciation in 2026: 100%, permanent, with a date cutoff

Bonus depreciation under §168(k) lets a taxpayer deduct a percentage of the cost of qualified property — property with a recovery period of 20 years or less — in the year it is placed in service, before regular depreciation begins. The 5-year and 15-year buckets qualify. The 27.5-year building does not. Cost segregation and bonus depreciation are therefore a pair: without the study, almost nothing in an acquired apartment building is bonus-eligible; with it, a quarter or more of the basis is.

The percentage has moved. The 2017 tax law set it at 100% for property acquired and placed in service after September 27, 2017, then phased it down — 80% in 2023, 60% in 2024, with 40% scheduled for 2025. The 2025 tax legislation reversed the phase-down and made 100% bonus depreciation permanent for qualified property acquired after January 19, 2025 and placed in service after that date. Property acquired on or before January 19, 2025 remains under the old schedule, so a building that went under a binding contract in late 2024 and closed in March 2025 generally gets 40%, not 100%; the acquisition date, determined under the binding-contract rules, controls. The IRS issued interim guidance on the new rules in Notice 2026-11 pending regulations. For a buyer closing on multifamily today, the practical statement is simple: every dollar the study moves into 5-year or 15-year property is deductible this year.

Two cautions. Bonus depreciation is an election out, not in — it applies automatically unless you elect out by class, which you might do in a low-income year. And many states decouple: California allows no bonus depreciation at all, and a number of others require addbacks, so the state return may show a very different picture from the federal one.

Why apartments segregate so well

Published studies and the firms that perform them generally place garden-style apartments in the 20% to 35% range of depreciable basis reclassified to short-life property, with the exact figure driven by vintage, finish level, and site. The reasons are structural. Apartment buildings have a very high ratio of interior finish to shell — every unit has its own kitchen, bathroom, flooring, and fixtures, so the appliance-and-cabinet bucket is replicated hundreds of times. Garden communities are low-density, so the land improvement bucket is unusually large: a 250-unit community on twelve acres has acres of parking, miles of sidewalk and fence, extensive landscaping and irrigation, a pool, and separate utility laterals to each of fifteen buildings. And apartment buildings have amenity spaces — clubhouses, fitness rooms, leasing offices — full of furniture and equipment that is unambiguously personal property.

The number falls for high-rise and mid-rise product, where the cost is concentrated in structure, elevators, and central systems and the land improvements shrink to a courtyard, and it falls for older, lightly finished buildings where the components have little remaining value. It rises for newly built or recently renovated communities with heavy finish packages. A study performed on a value-add acquisition can also be paired with a second pass after renovation, since the renovation dollars are new basis and the removed components can be written off under the partial-disposition rules.

A worked example: the $10 million garden community

A buyer closes in 2026 on a 180-unit garden-style community for $10,000,000, with $2,000,000 allocated to land and $8,000,000 of depreciable basis. The study reclassifies 30% of the depreciable basis: $1,400,000 to 5-year property and $1,000,000 to 15-year land improvements, leaving $5,600,000 in the 27.5-year structure. Mid-month and half-year conventions are ignored for clarity.

Rounded. The $2.6 million year-one figure is a deduction, worth about $963,000 of federal tax at the top rate if the owner can use it — see the next section. Annual depreciation in later years is lower because the short-life basis is gone. The study fee for a property of this size commonly runs in the low-to-mid five figures and is itself deductible.

Who can actually use the loss

A $2.6 million deduction on a property that produces $500,000 of net operating income creates a $2.1 million rental loss. Who deducts it depends entirely on the passive activity rules of IRC §469, and this is the part the short-form videos leave out.

Rental losses are passive by default, which means they offset only passive income. For a W-2 earner with no other passive income and more than $150,000 of adjusted gross income, the loss is suspended on Form 8582 and carried forward. It is not lost; it is parked. It will offset future rental income from the property, income from other passive investments such as syndication K-1s, or the gain when the property is eventually sold in a taxable sale. That is still valuable, but it is not a $963,000 check in April.

The loss is usable now in three situations. Real estate professional status, covered in its own guide, removes the per-se passive label for a taxpayer with more than 750 hours in real property businesses and more hours there than anywhere else, provided they also materially participate in the rentals; for a married couple one spouse must meet the hour tests alone. Short-term rentals with an average stay of seven days or less are not rental activities at all, so a materially participating owner's losses are nonpassive without the professional-status tests. And other passive income — from syndications, from a profitable rental portfolio, from a business you own but do not run — absorbs the loss directly, which is why experienced passive investors deliberately pair loss-generating acquisitions with income-generating ones. Even for a qualifying taxpayer, the §461(l) excess business loss limitation caps business losses against non-business income at $626,000 for joint filers in 2025, indexed, with the excess carried forward as a net operating loss, so a $2.1 million loss is deducted over more than one year regardless.

What it costs later: recapture

“Do you have to pay back cost segregation?” is the right question, and the honest answer is that you do not repay it, but you do settle up. Every dollar of depreciation reduces your basis, and basis is what you subtract from the sale price to compute gain. Take $2.6 million of depreciation instead of $291,000 and your gain on sale is $2.3 million larger. How that gain is taxed depends on which bucket produced the deduction.

Gain attributable to depreciation on §1245 personal property — the 5-year bucket — is recaptured as ordinary income, at rates up to 37%. Gain attributable to depreciation on the building and on land improvements that are §1250 property is unrecaptured §1250 gain, taxed at a maximum of 25%. Gain above the original basis is long-term capital gain at up to 20%. The 3.8% net investment income tax can apply to all of it. Our calculator separates these layers for any sale.

Three things make the trade worthwhile anyway. First, time value: a $963,000 tax deferral at a 37% bracket, invested for five or seven years, earns real money before any of it is repaid. Second, rate arbitrage: the 5-year property was deducted at 37% and much of the gain comes back at 25% or 20%, and if you sell in a lower-income year the ordinary recapture is taxed at a lower rate too. Third, you may never sell for cash. A building held until death receives a §1014 basis step-up and the recapture vanishes with the gain. A building exchanged under §1031 defers the gain — mostly. That “mostly” is the next section.

Cost segregation and the 1031 exchange

Exchangers who have segregated their building face a question the general explainers never reach: what happens to all that §1245 property in a like-kind exchange? Two rules interact.

The first is good news. The 2020 regulations at Treas. Reg. §1.1031(a)-3 define real property for exchange purposes by reference to the property's nature — land, inherently permanent structures, and structural components — and state expressly that classification as §1245 property for depreciation does not by itself disqualify an item from being real property for §1031. Carpet and cabinetry that a study classified as 5-year property can still be part of the real property you exchange; you do not have to carve them out and sell them separately as personal property, which was the fear when the 2017 law removed personal property from §1031.

The second is the catch. §1245(b)(4) limits the nonrecognition of §1245 recapture in a like-kind exchange: recapture is deferred only to the extent of the §1245 property you receive in return. If you relinquish a building carrying $1,400,000 of fully depreciated 5-year components and acquire a building whose 5-year components are worth $400,000, roughly $1,000,000 of §1245 recapture is recognized as ordinary income in the year of the exchange — even though the exchange is otherwise fully deferred and you received no cash. The practical response is to have a cost segregation study performed on the replacement property as part of the exchange, both to document the §1245 property received and because the replacement's own short-life components are what shelter the recapture. Buyers trading from a heavily finished garden community into a net-lease box or a DST interest with little personal property should model this before closing, not after.

The third point concerns depreciating the replacement. Under Treas. Reg. §1.168(i)-6, the basis carried over from the relinquished property continues to be depreciated on the old schedule as if the exchange had not happened; only the excess basis — the new money you added — is treated as newly placed in service and eligible for a fresh cost segregation and bonus depreciation. A taxpayer may elect out of this treatment and depreciate the entire replacement basis as new property, which restarts the clock and can be attractive when the carried-over basis is small or the old property was mostly depreciated. Either way, the exchange and the study are complements: the exchange defers the gain, the study on the excess basis generates the new deductions, and the §1245(b)(4) analysis tells you whether any recapture leaks out in between.

What a study involves, what it costs, and look-back studies

A quality study follows the IRS's Audit Techniques Guide: a site visit or detailed photographic survey, review of the closing statement, appraisal, and any construction drawings or contractor invoices, quantity take-offs for each component, cost assignment using actual costs where available or published construction-cost databases where not, classification of each component with citations to the authority supporting it, and a report that reconciles to total basis. The “detailed engineering approach from actual cost records” is the gold standard for new construction; the “detailed engineering cost estimate approach” is what acquisitions get. Studies that rely on rules of thumb or sampling without a property-specific analysis are the ones the guide flags.

Fees scale with property size and complexity rather than with the tax saved. Full engineering studies on an apartment community commonly run from the mid four figures to the low-to-mid five figures; lower-cost modeled or software-driven studies for small residential rentals exist for a fraction of that and are appropriate when the basis is small. The fee is a deductible expense. Reputable providers also include audit support in the fee, and the answer to the Reddit question about how to find a quality provider is the one the thread itself gave: ask the CPA who will defend the return which firms' reports they have seen hold up.

Look-back studies. A building bought in a prior year and depreciated entirely over 27.5 years can still be segregated. The IRS treats the correction as a change in accounting method eligible for automatic consent (Form 3115, designated change number 7), filed with the current-year return. The depreciation you should have taken from the acquisition date, minus what you did take, is a §481(a) adjustment deducted in full in the year of change — no amended returns. Bonus depreciation applies at the rate that governed when the property was originally acquired, so a 2021 acquisition catches up at 100% and a 2024 acquisition at 60%. For an owner who bought several years ago in a strong market and has since qualified as a real estate professional, the look-back study can produce the largest single deduction of their investing life.

When not to do it

The videos titled “why cost segregation can backfire” are describing real situations. A short hold is the clearest: if you will sell for cash within two or three years, you accelerate deductions at your bracket and recapture them at your bracket shortly after, having paid for the study and possibly pushed yourself into the §461(l) cap for nothing. A low-bracket year wastes the deduction, which is why the election out of bonus by class exists. No ability to use the loss is a reason to wait rather than a reason never to do it; a look-back study later captures the same deductions when they are worth more. State nonconformity can leave you tracking two depreciation schedules forever for a modest state benefit. A planned 1031 exchange into property with little personal property is the §1245(b)(4) problem above. And a partnership with partners in different tax positions — a real estate professional alongside passive investors — takes one depreciation method for everyone, which is a governance conversation before it is a tax one.

For passive investors: how it shows up on a K-1

If you invest as a limited partner in a multifamily syndication, the sponsor almost certainly commissions a cost segregation study on every acquisition, and the result arrives on your Schedule K-1 as a large box-2 rental loss in the year of purchase — often 50% to 80% of your invested capital in year one under 100% bonus, depending on leverage and the study's allocation. For you the loss is passive. It offsets distributions from the same deal, income from your other passive investments, and ultimately the gain on sale; it does not offset your salary unless you independently qualify as a real estate professional and materially participate, which a limited partner by definition rarely does. Sponsors who market “paper losses” as a feature are describing a real benefit for investors with passive income to shelter and a deferred one for everyone else. When the property sells, the K-1 will report the §1245 and unrecaptured §1250 recapture separately, and if the sponsor exchanges rather than sells, the §1245(b)(4) question is theirs to manage — and worth asking about before you wire funds. Our underwriting guide covers how to read the depreciation assumptions in a sponsor's projections.