What a cost segregation study hands back, and what has to happen to it next

The study comes back. It is a bound report, often over a hundred pages, with photographs, statutory citations, an engineer's methodology section and — somewhere past the middle — the thing the return actually needs: a schedule breaking one building into a long list of components with different recovery periods.

A property that entered the fixed asset system as a single line now has 90 or 180 lines behind it.

The analysis is finished at that point. The data entry has not started, and that second half is where the study's value either lands on the return or sits in a PDF on somebody's desktop until April.

The study produces a schedule, not a deduction

Worth being precise, because the marketing around cost segregation talks in terms of tax savings and the deliverable is a classification document.

An engineering-based study examines the components of a building and identifies the portions of its cost that qualify as shorter-lived property rather than as part of the 27.5 or 39-year structure — personal property such as specialised electrical, dedicated plumbing, cabinetry and fixtures, and land improvements such as paving, fencing and site lighting. Land itself is not depreciable and is not reclassified.

What comes back is an allocation: original capitalised cost, split across recovery periods, with the engineer's support for each classification. The deduction follows only once those components exist as assets in the depreciation system, each with its own cost, in-service date, method, life and convention.

Nothing about that step is intellectually difficult. It is 180 rows of transcription with no room for a mistake, which is a different kind of hard.

The report is a PDF and the tax software wants rows

This is the gap the industry does not discuss, and it is odd that it does not, because every study runs into it.

The firm that performed the study delivers a report. Some will provide a spreadsheet on request; many deliver only the bound document, and the component schedule inside it is a printed table with its own column order and its own headings — asset description, class life, cost, method, and often a placed-in-service date that differs from the building's when improvements were made later.

The destination wants those rows in its own import layout. Drake's 4562 import is 104 positional columns where the order is the contract. Lacerte's fixed asset import reads CSV and XLS but not XLSX. ProConnect takes 17 columns and truncates descriptions at 25 characters, which matters more than it sounds when a study names components at length.

So a study that cost several thousand dollars ends with someone reading a table off a PDF and typing it into a grid, which is exactly the operation that drops a row or reads a figure out of the adjacent column — and unlike a missing 1099, nothing downstream rejects it.

Applying a study to a building already in service

Where a study is done in the year a property is acquired, the components are simply set up correctly from the start.

Where it is done on a building that has been depreciating as a single 39-year asset for several years, the correction is retroactive, and the mechanism is a change in method of accounting rather than a set of amended returns. The cumulative difference between the depreciation taken and the depreciation that would have been taken under the reclassification is a §481(a) adjustment, and a taxpayer-favourable adjustment is deducted in full in the year of change — which is what Form 3115 is for.

That structure is the reason look-back studies exist at all. It also means the computation reaches backwards through the building's whole depreciation history, so the accuracy of the existing schedule stops being a bookkeeping matter and becomes an input.

What the study changes at the other end

Two consequences that arrive years later and are properties of how the assets were set up now.

Recapture character. Components reclassified as personal property are §1245 property, and §1245 recaptures gain as ordinary income up to the full depreciation allowed or allowable. The same dollars inside an undivided 39-year building would have been §1250, where ordinary recapture is usually zero because MACRS real property is straight-line. The study accelerates the deduction and changes what the disposal looks like — a trade rather than a free gain, and one worth having recorded accurately in the file.

Component-level disposals. Once a roof or an HVAC unit exists as its own asset, replacing it is a disposal of that asset rather than an unidentifiable part of a building. That is only available if the component list survived intact into the fixed asset system, with its own basis and in-service date.

Both of those depend on the rows being right, and both surface long after anyone remembers keying them.

Checking a component schedule before it goes in

Two arithmetic checks, and they take minutes.

The allocation should close. The reclassified components plus whatever remains in the building structure should equal the depreciable basis the study started from. If they do not, either a row was missed in transcription or the total being compared against includes land — which is not depreciable and should not be in the depreciable basis at all.

The subtotals should tie. Study schedules print totals per class life. Sum the transcribed rows within each class and compare against the printed subtotal at a dollar's tolerance for rounding. Anything larger than a dollar is a row, not rounding. This is the same tie-out that verifies any extracted schedule, and it is the only check that catches a duplicated or dropped component before the file is imported.

Then read the class lives against the report's own narrative. A study's summary table and its detail schedule occasionally disagree where the report was revised, and the detail is what gets keyed.

What we do with it

A cost segregation report is a printed schedule like any other, and the conversion treats it as one: the rows come off the page into the import format the destination accepts, and the file is not downloadable until the summed rows match the totals the document printed.

Nothing is classified, inferred or computed. If the report states a component's life, that life is transcribed; if the page does not state a convention, the field comes back empty and flagged rather than filled with a plausible one, and the flag blocks the download until a preparer resolves it. The tool has no opinion about whether a component belongs in a 5-year or a 15-year class — that is the study's determination and the preparer's to accept.

For the version of this problem that is just about getting a printed table into a workable grid, see depreciation schedule PDF to Excel.

FAQ

What does a cost segregation study actually deliver?

A classification report: the building's capitalised cost allocated across recovery periods, with engineering support for each component. The deduction follows only after those components are set up as individual assets in the depreciation system.

Can a study be applied to a building bought several years ago?

The catch-up is handled as a change in method of accounting with a §481(a) adjustment rather than by amending prior returns, which is why look-back studies are viable. Whether a particular property and history qualify is a determination for the preparer and the study provider.

Why does a study make disposal more expensive?

Reclassified components are generally §1245 property, which recaptures gain as ordinary income up to the full depreciation taken. Inside an undivided building those same dollars would usually have been §1250, where ordinary recapture is often zero. The acceleration is real and so is the trade.

The study came as a PDF. How do the components get into the tax software?

By transcription, unless the provider supplies a spreadsheet. The report's table has its own column order and headings, and every destination expects its own layout — Drake positional columns, a Lacerte CSV, ProConnect's 17-column sheet with descriptions capped at 25 characters.

How do I check the component schedule is complete?

Confirm the reclassified components plus the remaining structure equal the depreciable basis the study began with, then sum the transcribed rows within each class life and compare against the printed subtotal at a dollar's tolerance.

Does the study change the total depreciation over the life of the building?

It changes timing rather than the total. The same basis is recovered, earlier, with the disposal consequences that follow from the reclassified components being §1245 property.