I just interviewed the person who helped write the IRS cost seg rulebook. Here is what he said about audit risk.
James C. Peacock spent 38.5 years as a General Engineer at the IRS. He was among the first IRS engineers to examine cost segregation studies. He contributed to the original Cost Segregation Audit Techniques Guide, developed 2000 through 2002 and released publicly in 2004, and every major update since, including the February 2025 edition. Per James, he trained approximately 200 new-hire IRS engineers on cost segregation and Section 179D before retiring in September 2025.
If anyone can tell you what actually happens when the IRS looks at a cost segregation study, it is James Peacock.
Quick Answer: Cost segregation does not trigger IRS audits. James Peacock, a 38.5-year IRS engineer, puts the business audit rate at roughly 0.07%. With a cost segregation study, roughly 0.078%. The IRS Data Book shows S-corps and partnerships around 0.1% in TY2021. The real risk is study quality.
The Actual Numbers: Where the 0.4% Figure Comes From (and Why It's Wrong)
You may have seen a statistic that the IRS audit rate is around 0.4%. That number reflects individual returns across all income brackets, not business entities holding rental property.
James gave me the numbers he observed from inside the IRS. Businesses get audited at roughly 0.07%. Add a cost segregation study to your return, and that rate moves to roughly 0.078%.
"Two out of a thousand instead of one out of a thousand," he said. "It doubled. But it's still very, very, very low."
That is the whole story. The delta is 0.008 percentage points. The IRS Data Book (Publication 55B, Table 3-1) corroborates the order of magnitude: S-corporations and partnerships, the entity types most likely to hold investment real estate, show audit rates around 0.1% in the most recent settled data (Tax Year 2021). James's professional observation from inside those examinations tracks.
Here is what makes this even cleaner: the IRS does not segment audit data by deduction type. There is no published table for "businesses with cost segregation studies." The IRS tracks audits by entity type and income size, not by depreciation strategy. James's numbers come from his professional observation inside the IRS, not a published breakdown that doesn't exist.
The IRS's own Cost Segregation Audit Techniques Guide adds the final point: a well-prepared study "greatly expedites the Service's review, thereby minimizing audit burden." Quality documentation doesn't raise audit risk. It reduces friction if an audit opens for any other reason.
The 0.4% figure you may have seen is not wrong. It describes a different population. It is the wrong number for this question.
Cost Seg Does Not Trigger Audits. It Gets Discovered Inside Them.
This is the most important thing James said in our entire conversation, and it reframes the entire question.
"Cost seg studies don't get discovered until there is already an audit open," he explained.
IRS audits are initiated through DIF scores (Discriminant Information Function scores) and specific IRS examination campaigns, not by scanning for depreciation deductions. When an agent is reviewing a return that was already flagged, they may encounter large depreciation figures and pull in an IRS engineer. That is when the cost seg study surfaces.
So the question should not be "will doing cost seg get me audited?" The question should be "if I am already in an audit, how does my cost seg study hold up?"
Those are very different problems with very different solutions.
What the IRS Actually Checks First
When an IRS engineer does review a cost segregation study, James told me there are a small number of things they go to immediately.
Land Allocation
"If a property is purchased for a million dollars and the study adds up to a million dollars, we go, where's the land?"
Land cannot be depreciated. Full stop. If a study is allocating 100% of a purchase price to depreciable property with nothing set aside for land, that is an automatic adjustment. No negotiation. The engineer catches it in the first ten minutes.
This is not a subtle mistake. It is a structural error that a qualified provider simply does not make. But James saw it repeatedly in studies from less rigorous firms.
RS Means Code Specificity
James told me he goes to the back of a report looking for RS Means codes. RS Means is the industry-standard construction cost database. When a study references "RS Means mechanical" without a 12 or 16-digit code, that generates an Information Document Request (IDR). Vague references to a cost database are not support. Specific line-item codes with units and quantities are support.
Square footage models get the same treatment. "They tell the IRS they used averages and didn't look at the property," James said. Engineers flag those studies immediately and request contractor documentation to verify.
The 1250 Property Problem
James reviewed a study on a high-rise building. The 1245 personal property calculations looked reasonable. But when he examined the 1250 real property section, there was no steel framework accounted for. No concrete slabs. "Tens of millions of dollars just gone. That was an automatic adjustment right there."
1250 property covers the structural components of a building. Studies that race to maximize 1245 classifications sometimes treat 1250 as an afterthought. James said he saw this pattern often. The IRS engineer will catch it, and the adjustment is not small.
The LUQ Framework and Why the IRS Will Never Publish Thresholds
IRS engineers use a framework called LUQ: Large, Unusual, Questionable items. When something in a study qualifies as LUQ, it gets scrutinized.
James gave a practical example: a standard warehouse where HVAC accounts for 50% of total property cost. That is unusual for a building that is not temperature-controlled. It draws attention.
Here is the important thing he told me about IRS thresholds: "The IRS deliberately never published percentage thresholds because as soon as we say X percent is acceptable, everyone's going to claim X minus 1 percent."
There is no published bright line. The IRS keeps it that way on purpose. This means any provider telling you their allocation percentages are "safe" because they fall below some known threshold is either guessing or misleading you.
The IRS flag is proportionality to the specific property. A legitimate engineer builds allocations based on the actual building and documents why each figure is what it is.
The Kitchen Cabinet Problem and the Whiteco Test
Two of the trickier classification questions in cost segregation involve things that look movable but may not be.
Kitchen Cabinets and the Amerisouth Ruling
In Amerisouth v. Commissioner (2012), the Tax Court ruled that kitchen cabinets, counters, and sinks are 1250 property, not 1245 personal property. The court's reasoning: they serve "the operation and maintenance of the building."
To classify them as 1245 property (5-year depreciation), you must document two things: (a) they were actually removed from the property and (b) they were actually reused or disposed of. Both. Not theoretically removable. Actually removed.
James told me this distinction matters because many studies classify fixtures as personal property based on removability theory without documenting actual removal. That will not hold up.
The Whiteco Test
The Whiteco case established a related principle: personal property has to be movable and must in fact be moved, not just be capable of being moved.
The case involved roadside billboards that were physically dismantled and relocated. The key was the actual movement, not the theoretical possibility of it.
Self-storage facilities are a practical example. Metal partitions screwed into walls can qualify as 1245 property if the owner can prove they are removed and relocated when tenants change. If they stay permanently fixed, the classification does not hold.
IDR: What Happens When the IRS Asks Questions
An IDR is an Information Document Request. It is a formal document the IRS sends when they need documentation to evaluate something on a return.
When a cost seg study is pulled into an audit, the first-round IDR typically asks for the purchase agreement and the cost seg study itself. If the responses raise further questions, additional IDRs follow.
"The least amount of IDRs, the easier the audit goes," James told me.
What generates more IDRs: vague RS Means references, missing land allocations, square footage models, and 1250 property that looks incomplete. What generates fewer IDRs: a complete study with specific cost codes, proper land allocation, a contractor's final application for payment, and plans and specs.
The IRS engineer will typically ask for the contractor's final application for payment and the property's plans and specifications. These are the primary source documents. A good study is built from them. A weak study cannot be reconciled against them.
James also flagged contingency fee arrangements as something that draws scrutiny. The ATG directs examiners to "closely scrutinize" studies where the provider's fee is based on the tax benefit generated. The IRS will ask for the engagement agreement in the first IDR. This does not mean contingency fee studies are automatically invalid, but it means they face extra examination from the start.
"It's the Support, Not the Report"
This phrase came up repeatedly. James attributed it to a colleague, and it was something he carried throughout his career.
The report is the deliverable you receive. The support is everything behind it: the site visit notes, the cost calculations, the RS Means codes, the contractor invoices, the documentation of what was found and how it was measured.
A professional-looking 80-page PDF means nothing if it cannot be reconciled against the underlying property. IRS engineers do not grade on presentation. They ask for documentation and trace it back to the source.
This is also where AI-generated cost seg studies fail entirely. James was direct about it: AI studies that skip physical inspection "go against every IRS rule." There is no remote observation substitute for a qualified engineer walking the property.
That said, James acknowledged AI has a legitimate supporting role. "We would get cost seg studies where there was so much emphasis on getting the 1245 property estimated that the 1250 property was an afterthought." AI can function as a completeness check to make sure 1250 property is not overlooked. It cannot do the classification work itself. See our breakdown of legitimate studies versus scam operations for more on how to spot the difference.
Short-Term Rentals and the 39-Year Rule Most Investors Get Wrong
One classification error James flagged is particularly common with short-term rental properties.
STR properties with average stays under 30 days are classified as 39-year non-residential property, not 27.5-year residential property. The residential classification requires average stays above 30 days.
This changes the base depreciation but it does not eliminate the cost seg benefit. Interior improvements to non-residential property qualify as Qualified Improvement Property (QIP), which is 15-year property eligible for 100% bonus depreciation. Land improvements, pools, landscaping, parking, are 15-year property regardless of rental type.
The gain from cost segregation on STR properties comes from the 15-year component acceleration, not from the building classification itself. Providers who advertise the residential 27.5-year rate on under-30-day STR properties are applying the wrong base.
The New Class of IRS Engineers
Per James, he trained approximately 200 new IRS engineers in cost segregation and Section 179D before retiring.
He noted that the IRS is expanding its use of AI in the agent force, including a contractor program for return selection. New tools are being gradually introduced.
His assessment of what this means for audit exposure: "The audit rate might not go down. If you get caught in a trap, it's still a trap."
More trained engineers plus improved return selection tools means studies with the quality problems described above face greater examination risk over time, not less. The investors who get audited on cost seg are not audited because they did cost seg. They are audited because something else flagged their return, and then their study could not survive review.
The study quality question is not just about worst-case scenarios. It is the primary variable you control.
Choosing a Provider Based on What You Now Know
The IRS flags specific things: missing land allocation, vague or absent RS Means codes, square footage models, underbuilt 1250 property sections, and classifications that rely on removability theory without documentation of actual removal.
A legitimate provider does not cut corners on any of these. Their fee reflects the cost of a qualified engineer doing a physical inspection, building a property-specific cost model, and producing documentation that can withstand IDR scrutiny.
If a quote looks significantly cheaper than the market rate, ask why. A study that saves you $80,000 in taxes but cannot survive a first-round IDR has cost you more than you saved.
For more on provider selection, see our guide to choosing a cost segregation provider and our comparison of top firms.
FreeCostSeg has a detailed technical breakdown of what an IRS engineer looks for during examination, sourced from the same interview: What an IRS Engineer Looks for in a Cost Segregation Audit.
Frequently Asked Questions
Does the IRS target cost segregation studies specifically?
No. Cost segregation does not appear as a standalone audit trigger. Returns are selected through DIF scoring and specific IRS examination campaigns. If a return is already under audit, large depreciation deductions may bring in an IRS engineer to examine the underlying study. The study itself does not initiate the audit.
What is the actual audit rate for businesses that have done cost segregation?
The IRS audits a very small fraction of pass-through business returns, well under 1% annually. There is no published data showing cost segregation raises that probability. James described the overall audit risk as "very, very, very low" and was clear that cost seg itself does not trigger audit selection.
What is the most common mistake in cost segregation studies that triggers IRS adjustments?
Missing land allocation is the most automatic. If a study allocates 100% of a purchase price to depreciable property with nothing for land, it is an immediate adjustment. After that: vague RS Means codes, square footage models, and 1250 property sections that are incomplete or treated as an afterthought.
Should I avoid cost segregation to reduce audit risk?
No. The IRS audit rate for business entities is very low and there is no published data connecting cost segregation studies to higher audit probability. What you should focus on: choose a provider who performs physical inspections, uses property-specific RS Means codes, properly allocates land, and builds documentation that can withstand an IDR. The strategy is not the risk. Study quality is the risk.
About the Expert
James C. Peacock spent nearly 39 years at the IRS as a General Engineer and Subject Matter Expert in the LB&I Division. He was among the first IRS engineers to examine cost segregation, contributed to the Cost Segregation Audit Techniques Guide from 2004 through the 2025 update, and served as the IRS's primary technical expert on Section 179D from 2014 through his retirement in September 2025. Per James, he trained approximately 200 new-hire IRS engineers on cost segregation and 179D before retiring. He holds a degree in Architectural Engineering from The University of Texas at Austin (1983).
Wondering if cost segregation makes sense for your property? Use the free calculator at FreeCostSeg.com to estimate your potential tax savings in under two minutes.
