How Does a Cost Segregation Look-Back Study Work?

The computation, step by step

The look-back mechanism sounds abstract until you see the arithmetic behind it. There are really three steps.

1. Classify. The engineering team walks the building and assigns every component to its correct recovery period: 5-year (carpet, most flooring, decorative lighting, cabinetry, appliances, window treatments, certain electrical and plumbing serving equipment), 7-year (certain fixtures and furniture), 15-year (paving, fencing, landscaping, site utilities, outdoor lighting, pools and patios), and everything else stays on the 39-year (commercial) or 27.5-year (residential) structural shell.
2. Recompute. For every year the property has been owned, the team calculates what depreciation would have been claimed on each component had it been classified correctly from the placed-in-service date forward, including the bonus depreciation rate that applied to that component's original acquisition date.
3. Subtract. That recomputed total is compared against what was actually claimed under the original, less accurate classification. The difference between the two is the section 481(a) adjustment, and it is entirely a catch-up, not a projection. It reflects real depreciation the property was always entitled to and never claimed.

What lands on the return

That adjustment, whether it favors more depreciation (almost always the case in a look-back, since the whole point is correcting an over-conservative original classification) does not go on an amended return. It goes on Form 3115, filed with the current year's tax return, which is the IRS's mechanism for an automatic consent to change an accounting method. The 481(a) adjustment computed in the study becomes the number the CPA reports on that form.

This is a meaningfully different filing event than most owners expect. There is no reopening of 2019, 2020, or 2021 returns. There is one current-year filing that captures every year of the correction at once. That structure is exactly what makes a look-back study practical years after the fact instead of a multi-year cleanup project, and it is the same reason the strategy is described in detail on our Form 3115 look-back guide.

What our engineering team owns versus what the CPA owns

The division of labor is clean and it matters. Our engineering team classifies the building, computes the section 481(a) adjustment, and stands ready to take technical questions directly from the owner's own CPA about methodology, cost allocation, and class-life assignments. The CPA reviews that work, prepares the Form 3115, and files it along with the rest of the return.

We never file returns ourselves. That is a deliberate boundary, not a limitation. The engineering behind a study and the tax filing built on top of it are two different disciplines, and keeping them separate is part of what makes the study hold up if it is ever questioned, since each piece was built and is defended by the people who actually did that piece of the work. See our audit defense guide for how that division plays out if an examiner has questions.

How this differs from a same-year study

The engineering itself is identical to a study run in the year of purchase. The building gets classified into the same recovery periods using the same methodology. What changes is only the mechanism for claiming the result: a same-year study simply becomes the property's first depreciation schedule, while a look-back study computes a catch-up on top of years of prior depreciation and delivers it through Form 3115 instead.

One number worth understanding: the benchmark studies our engineering team has delivered describe first-year deductions as the section 481(a) catch-up (when a study reaches back) plus the current year's increased depreciation from the newly identified faster schedules. On a $2,804,440 restaurant building basis, one delivered study identified $599,678 in first-year deductions against a $9,000 fee, a 66.6 to 1 ratio. That figure is the combination of both pieces working together, not one or the other.

When the CPA has not handled one of these before

A section 481(a) catch-up filed through Form 3115 is a routine filing category, but any individual CPA may not encounter one every year, since most of their clients simply have not had a look-back study done. That unfamiliarity is normal and it is exactly what the technical relationship between our engineering team and the owner's CPA is built to solve. The CPA does not need to have handled a cost segregation catch-up before to file one correctly; they need the computation done right and a team available to answer specific questions about it.

In practice this looks like a short technical exchange: the CPA reviews the delivered computation, asks about specific classification decisions or how a particular component's cost was allocated, and our engineering team answers those questions directly, the same team that built the classification in the first place. That is a materially different experience than being handed a report with no one available to explain how a specific number was reached.

What to bring to the conversation

An owner considering a look-back study benefits from having the property's placed-in-service date, the original purchase price allocation between land and building, and a general sense of any capital improvements made since. None of that is required to start, since a free Preliminary Benefit Estimate can model the likely result before any of it is assembled, but having it ready shortens the path from estimate to delivered report.

It also helps to loop in the CPA early rather than after the engineering is finished. A short conversation up front about how the Form 3115 filing fits into that year's broader return, alongside any other income, loss, or planning items already on the table, tends to make the eventual filing smoother than introducing the whole mechanism for the first time right before a deadline.

Frequently asked questions

Is the 481(a) adjustment always positive for the owner?

In the overwhelming majority of look-back studies, yes, because the original classification was almost always more conservative (39-year or 27.5-year straight-line across the board) than a correct component-level classification would be. Whether that pattern holds on a specific property is a determination the study itself makes, not an assumption made in advance.

Does filing Form 3115 increase my chances of an IRS exam?

Form 3115 is a routine, automatic-consent filing used for many accounting method changes beyond cost segregation, and the IRS's own Audit Techniques Guide describes exactly how a look-back study using this form should be documented. There is no published data suggesting the form itself raises exam risk when the underlying study is properly engineered.

How long does a look-back study take from start to finish?

Engineering turnaround generally runs 4 to 6 weeks during tax season, and typically 2 to 3 weeks in January and February. The Form 3115 filing timeline on top of that depends on the CPA's own return preparation schedule.

Can I do more than one look-back study on the same property over time?

A look-back study captures all missed depreciation up to the point it is performed. A later capital improvement to the same property would generate its own new component to classify, but it would not reopen or duplicate the earlier look-back's catch-up.