What is the IRS Cost Segregation Audit Techniques Guide?

The ATG is the IRS's internal training manual for examining cost segregation studies, first issued in 2004 and updated since (a substantially revised edition was released in 2022). It is publicly available, describes the legal framework, ranks study methodologies by reliability, lists the elements of a quality study, and flags commonly misclassified components. It is guidance to examiners, not law.

What does the IRS consider a quality cost segregation study?

The ATG lists thirteen principal elements, including preparation by someone with construction and tax expertise, a detailed methodology description, actual site inspection, use of actual cost records where available, listing of each asset with its cost and classification, legal citations supporting classifications, reconciliation of all costs to total basis, and treatment of indirect costs. Studies missing several elements draw proportionally more scrutiny.

Which cost segregation method does the IRS prefer?

The detailed engineering approach from actual cost records ranks highest — tracing contractor payment records to specific components. The detailed engineering cost-estimate approach, used on acquisitions where records don't exist, is also accepted when well documented. Survey/letter approaches, residual estimation, and rule-of-thumb percentages rank lowest and invite the examiner to challenge the entire allocation.

What components does the IRS most often challenge in cost segregation studies?

Recurring targets include general-use electrical and plumbing allocated to short lives without load studies, HVAC classified as personal property without meeting the sole-justification test, entire wall and partition systems treated as movable, site utilities that serve the building's operation, and aggressive allocations of indirect and soft costs to short-life property.

Reading a cost segregation study the way an IRS examiner would

The IRS published its playbook. The Cost Segregation Audit Techniques Guide — first issued in 2004, revised most recently in a substantially rewritten 2022 edition — is the manual revenue agents use to examine studies, and it is sitting on irs.gov for anyone to read. For a CPA deciding whether to sign a return built on a study, the ATG is the single most useful vetting document in existence, because it says, in the government's own words, which studies get respected and which get shredded.

This piece walks through what the ATG tells examiners to look for: the methodology hierarchy, the elements of a quality study, the components most often misclassified, and how to read a study report critically before — not after — the deduction goes on the return. It assumes familiarity with the basics of what a study is.

What the ATG is, and what it isn't

An Audit Techniques Guide is training material for examiners, not substantive authority. It cannot be cited as law, and taxpayers are not bound by it. But it performs two functions that make it indispensable in practice. First, it frames how an examination will actually proceed — the information document requests, the questions, the components pulled for testing. Second, it collects the governing law in one place: the Section 1245/1250 distinction, Hospital Corporation of America v. Commissioner, 109 T.C. 21 (1997), the Whiteco permanence factors (65 T.C. 664 (1975)), Rev. Proc. 87-56's class lives, and the regulations under Section 1.48-1 defining structural components.

The ATG's posture is notable: it does not attack cost segregation as a concept. It accepts that properly supported component classification is the law. Its skepticism is aimed at quality — unsupported percentages, credential-free preparers, and classifications that ignore the case law's actual factors.

The methodology hierarchy

The ATG describes six approaches and is explicit that reliability varies. In descending order of examiner confidence:

How the ATG ranks study methodologies:

Rank	Approach	Description	Examiner posture
1	Detailed engineering from actual cost records	Traces contractor pay applications, invoices, and change orders to specific components	Most reliable; testing focuses on classification calls
2	Detailed engineering cost estimate	Components identified from drawings and inspection, costed from published data (e.g., RSMeans), reconciled to basis	Accepted for acquisitions lacking cost records, if documented
3	Survey / letter approach	Contractors asked after the fact to estimate component costs	Weaker; hearsay-quality costing
4	Residual estimation	Short-life items costed; everything unexamined dumped into 39-year — or worse, the reverse	Scrutinized; shortcuts invite full reallocation
5	Sampling / modeling	Study of sample properties extrapolated across a portfolio	Acceptable only with statistically defensible design
6	Rule of thumb	"Hotels are 30% five-year property"	No documentation; effectively an invitation to disallow

The practical lesson: on new construction, insist the provider use actual cost records — they exist, and estimating around them is a red flag in itself. On acquisitions, a detailed engineering estimate with a real site visit is the standard; a purely remote "desktop" study is a residual-method product wearing better branding.

The thirteen elements of a quality study

The ATG lists principal elements of a quality study and principal elements of a quality report. Condensed, the study should show:

Preparation by qualified individuals — construction/engineering expertise plus tax-law competence, with credentials stated.
A described methodology — which of the approaches above, and how applied.
Use of appropriate documentation — cost records, drawings, specifications, appraisals.
Site inspection — with date, inspector, and photographs.
A complete asset listing — every unit of property, its cost, and its assigned class life, not category-level lump sums.
Legal support for classifications — citations to Rev. Proc. 87-56 asset classes and the relevant cases for contested items.
Cost reconciliation — component costs sum exactly to depreciable basis; nothing double-counted, nothing orphaned.
Treatment of indirect costs — architect fees, permits, general conditions allocated on a stated, consistent basis.
Land and non-depreciable items excluded and separately stated.
Identification of Section 1245 vs. 1250 rationale for borderline items specifically.
Consistency with the taxpayer's other filings — purchase price allocations under Section 1060, appraisals, insurance schedules.
A statement of assumptions and limiting conditions.
Preparer signature and firm identification.

None of these is individually dispositive, but the ATG trains examiners to treat gaps as depth-of-audit signals. A report with no photographs, no asset-level detail, and no legal citations will be examined as if every classification is wrong until proven otherwise.

The components examiners are trained to challenge

The ATG devotes chapters to specific problem areas. The recurring ones:

Electrical distribution. Only wiring, panels, and outlets that serve specific items of Section 1245 equipment qualify for short lives. Studies frequently allocate a percentage of the entire electrical system to 5-year property using a load study — acceptable in principle, but the examiner will ask for the load study. General building lighting and convenience outlets are structural components, full stop.
Plumbing. Kitchen and process plumbing serving equipment can be 5-year (that is the HCA holding); restroom plumbing is a structural component. Blanket percentages that sweep in domestic water and sanitary systems get reversed.
HVAC. The default is structural component. Machinery whose sole justification is meeting equipment or process temperature/humidity requirements (the §1.48-1(e)(2) sole-justification test) can qualify — a server-room unit, a kitchen hood makeup system — but comfort-conditioning capacity is 39-year even if it incidentally helps equipment.
Walls and partitions. Only genuinely movable, reusable partition systems (the Whiteco factors: movability, design permanence, damage on removal) escape structural-component treatment. Drywall on metal studs does not become 5-year property because a tenant might someday remove it.
Site work. Land improvements (15-year) must be distinguished from land itself (never depreciable) — clearing and general grading are land; parking lots are improvements. Site utilities serving the building's general operation belong with the building.
Indirect and soft-cost loading. Allocating a disproportionate share of architect fees, interest, or general conditions to short-life property inflates the reclassification quietly. The ATG tells examiners to test the allocation basis.

How to read a study report critically

A practical vetting sequence for the CPA who receives a finished report:

Check the reconciliation first. Do the components sum to depreciable basis? Is land excluded and stated? Arithmetic failures predict everything else.
Identify the methodology honestly. If the property was new construction and the report used estimates anyway, ask why. If it was an acquisition and no one visited the site, discount accordingly.
Pull the five largest short-life line items and read their justifications. Are there citations and facts, or adjectives? The examiner will sample the same way.
Scan the electrical, HVAC, and plumbing allocations against the problem list above — these are where aggressive studies hide their aggression.
Compare the reclassification percentage to the norms for the building type (see asset classes and recovery periods for the class framework). A warehouse at 35 percent short-life is not a triumph; it is a finding waiting to happen.
Confirm the downstream mechanics. If this is a look-back, the Section 481(a) computation and Form 3115 filing need to match the study; if this is a placed-in-service-year study, check bonus eligibility dates against the acquisition contract (the January 19, 2025 seam matters — see the bonus interaction).

The exposure math

If an examiner reallocates components back to 39-year, the immediate cost is tax on the disallowed acceleration plus interest. Accuracy-related penalties under Section 6662 are in play for substantial understatements, though a taxpayer who reasonably relied in good faith on a professionally prepared, apparently qualified study has a reasonable-cause argument under Section 6664(c) — an argument that gets stronger the more the study resembles the ATG's quality checklist and weaker the more it resembles a rule-of-thumb letter. The preparer's own diligence obligations run in parallel.

The asymmetry is the point. A quality study and a cut-rate study claim similar deductions; only one of them holds the deduction when tested. The ATG is the government telling you, in advance and in writing, which is which. Read the study against it before the return is signed — and if the study cannot survive that reading, neither can the position. For owners whose facts are marginal in the first place, the better question may be the one posed in when cost segregation doesn't make sense.