DISCLOSURE This site is operated by costsegsmart.com — it is an editorial reference, not an independent publication.
REFERENCE · AUDIT CASES · TAX COURT RECORD

Cost segregation audit cases — what survives, what fails.

The outcome of a cost segregation audit depends almost entirely on whether the underlying study had engineering methodology — not on the study fee, not on the firm's brand, not on the property's value. The Tax Court record makes this distinction visible.

Reading time · 13 min Updated · 05 / 26 Source · Tax Court + IRS ATG
01

The pattern

Engineering-based studies survive examination. Residual-method "studies" don't. The IRS treats those as two different things.

"Has anyone ever had an IRS audit involving a cost segregation study?" is one of the most common pre-purchase questions investors ask. The honest answer is yes — it happens, and the outcome depends almost entirely on whether the study had engineering methodology behind it.

The public Tax Court record bears this out. Studies grounded in engineering analysis — site evidence, component-level cost basis, classification rationale tied to Rev. Proc. 87-56 class-life tables and the IRS Cost Segregation Audit Techniques Guide — survive. Studies that are actually a percentage applied to purchase price, dressed up in cost-seg vocabulary, don't.

This page walks through the foundational cases on both sides, then translates the pattern into what to look for in any study you receive.

02

Cases where the study survived

Two foundational decisions establish what an engineering-based study looks like — and how the courts evaluate it.

SURVIVED EXAMINATION
Hospital Corporation of America v. Commissioner
109 T.C. 21 (1997) · The foundational case

Often called the "HCA case," this is the decision that legitimized engineering-based cost segregation as we know it today. Before HCA, taxpayers used component depreciation methods that the IRS challenged routinely. The Tax Court held that engineering-based reclassification of building components into shorter MACRS asset classes was acceptable when supported by proper engineering analysis.

What HCA actually established: components properly classified as tangible personal property under §1245 — even if physically attached to a building — can be depreciated over the shorter recovery periods specified in Rev. Proc. 87-56, not the 27.5- or 39-year building life. The decision also accepted the use of standard cost-engineering reference data (RSMeans, Marshall & Swift) as a basis for component cost allocation when original construction documents weren't available.

Why this matters for any study you order: HCA is the framework every defensible study still operates within. A study that doesn't reference HCA-style component classification — and instead applies a flat percentage to the building basis — is operating outside the precedent that protects it.

SURVIVED EXAMINATION (PARTIALLY)
AmeriSouth XXXII, Ltd. v. Commissioner
T.C. Memo 2012-67 · The multifamily refinement

The AmeriSouth case is where the line between §1245 personal property and §1250 structural building components got tested for residential multifamily. The taxpayer's cost segregation study reclassified a long list of items — kitchen plumbing, bathroom accessories, mirrors, special-purpose electrical, site improvements — into shorter recovery periods. The IRS pushed back on a subset.

The court split the items. Where the taxpayer's engineering analysis demonstrated that a component's primary purpose was non-structural — for example, special-purpose electrical wiring serving a specific appliance, not general illumination — the reclassification held. Where the analysis amounted to "this is a kitchen item, kitchens are personal property" without component-level engineering support, it didn't.

Why this matters: AmeriSouth is the case multifamily and residential cost-seg providers are tested against. A residential study that reclassifies, say, decorative lighting as 5-year property without showing why that specific lighting is decorative-not-functional doesn't have AmeriSouth-style support. A study that does is on solid ground.

03

Where studies fail

Three failure patterns recur across the audit record. None are about study fee. All are about methodology.

FAILURE PATTERN 01
"Percentage of basis" templates without engineering analysis

The most common failure mode. The "study" is a single page showing the building basis multiplied by a fixed percentage to produce the 5-year, 7-year, and 15-year reclassification amounts. There's no asset detail. There's no classification rationale. There's no source cost data.

This isn't a study under the meaning of Pub 5653. It's an estimate dressed up as a study, and an examiner who opens the workpapers can tell within seconds. Reclassifications under this kind of "study" don't survive review because there's nothing in the record to defend.

FAILURE PATTERN 02
Residual-method "DIY" studies without engineer involvement

Variations on the residual method exist in tax software and in spreadsheet templates circulating in real-estate-investor communities. The owner classifies items themselves, applies generic percentages, and produces a worksheet that looks structured but has no engineer in the loop.

The IRS Cost Segregation Audit Techniques Guide is explicit that the preparer's qualifications matter. Pub 5653 lists "preparer qualifications" as one of the six characteristics of a quality study. A study without a qualified preparer's name and methodology disclosure attached fails this test on its face.

FAILURE PATTERN 03
Engineer-prepared but property-blind

The most subtle failure mode. The study has engineering methodology on paper. There's a credentialed engineer's name on the cover. But the engineer never reviewed property-specific photos, never looked at construction documents (when available), and never validated the component list against the actual property. The classifications are generic boilerplate keyed to property type.

Under examination, this kind of study fails the "reproducibility" test in the ATG: another competent preparer, working from the same workpapers, couldn't arrive at the same result because the workpapers don't actually reflect the property. The classifications might survive on items that are obviously property-type-generic. They don't survive on anything else.

04

What the IRS actually looks for

The IRS Cost Segregation Audit Techniques Guide is the document examiners are explicitly instructed to use. It's public, it's specific, and it's the right place to start.

The IRS Cost Segregation Audit Techniques Guide (Publication 5653) isn't a regulation — it's the field manual examiners follow when reviewing a cost-seg study. The Guide names six characteristics of a quality study. In practice, these four carry most of the weight in audit defense:

  1. Engineering-based classification per Rev. Proc. 87-56 class-life tables. Components classified into 5-, 7-, 15-year asset classes per the published class-life schedule, not by author judgment. Rev. Proc. 87-56 is the table.
  2. Property-specific evidence. Closing documents, the property's tax assessment, dated property photos, construction documents where available. The study has to reflect the actual property, not a property-type template.
  3. Detailed cost basis for each reclassified component. Per item or per category. Not lump percentages. The workpapers need to show how the component's cost was derived (invoice, contractor bid, RSMeans line item, etc.).
  4. Methodology disclosure that's reproducible. Another competent preparer, given the same workpapers, would arrive at the same result. The methodology section in the report describes the approach in enough detail that a CPA — or an examiner — can replicate the math.

For a deeper walk-through of these characteristics applied to a study you've received, see the Defensibility Checklist. For documented failure patterns at greater depth, see Failure modes.

05

Study fee is not the variable

A common misreading of the audit risk is that more expensive studies are safer. The audit record doesn't support that.

The IRS doesn't price-discriminate audits. A $495 engineering-based study and a $7,000 engineering-based study have the same audit defense if both are grounded in Pub 5653 methodology. Conversely, a $7,000 "study" that's actually a residual template — and these exist; some traditional firms produce them at premium pricing for small-property work because the unit economics of doing real engineering at that scale don't pencil — is more dangerous than no study at all, because it generated a reclassification position that the workpapers can't support.

What's changed in the last several years is that engineering-based methodology — site visit (or structured photo capture), component take-offs, RSMeans-aligned cost data, classification rationale per asset class, methodology disclosure — has been automated for the residential and small-commercial segments. The same engineering analysis that took 20-40 hours of engineer time in the 2000s now runs in under an hour for a typical residential property because the cost database, the component classification logic, and the QC checks are codified. The methodology hasn't changed; the unit economics did.

This is why the price spread on engineering-based studies has collapsed for residential and small-commercial. It's not because the small-property version is a worse study. It's because the cost structure that made traditional firms charge $5,000-$8,000 per study (engineer hours billed) doesn't apply when the engineering analysis is codified.

06

How to evaluate a study you receive

Three questions to ask before signing off on any cost-seg report — yours or a client's.

Does the report have a CPA Review Page?
A summary section that translates the engineering output into the language a CPA needs to file Form 4562 (or Form 3115 for lookback adjustments). If the CPA Review Page isn't present, the CPA has to do that translation work themselves and may push back on the study.
Does the report disclose its methodology?
A dedicated methodology section that names the data sources (RSMeans / Marshall & Swift / equivalent), the component classification framework (Rev. Proc. 87-56 class-life tables), and the engineering review process. If the methodology section is generic boilerplate, the study won't survive an examiner who reads carefully.
Will the provider show you the methodology section of a sample report before you order?
Engineering-based providers will. Residual-method providers usually won't, because the methodology section is the part that reveals which kind of study they actually produce. A provider that won't share a sample report's methodology section before you commit is a hard pass.
The summary in one sentence

Cost segregation audits go badly when the study had no engineering analysis behind it; they go well when the study followed Pub 5653 methodology — regardless of what the study cost.

For the practical "before-you-order" version of this analysis, see the Defensibility Checklist. For the IRS's own manual that examiners use, see Pub 5653.

NEXT

Run a study you've received through the checker.

Six yes/no questions, calibrated against the same Pub 5653 criteria that examiners use. For documented failure patterns at the workpaper level, see Failure modes. For a sample 40+ page engineering-based study with the methodology section fully disclosed, see a sample Cost Seg Smart report.

→ DEFENSIBILITY CHECKER