What a cost segregation study does on a new apartment building

A cost segregation study is a document, built out of records that either exist or do not. This is about the document, and about what changes when the company that built the apartments files the return.

What the study is, and who performs one

A finished apartment community arrives on a tax depreciation schedule as very few lines: land, building, sometimes site improvements, with the building written off over 27.5 years. That is a convenience, not a fact about the property. A building is thousands of assets with different working lives, and many of them are not structural.

A study takes the capitalised cost of the project, splits it into those assets, and gives each the recovery period its class carries. The components that move are largely the ones a builder buys as their own packages: the appliance package, casework, unit flooring, specialty lighting, and much of the site work — paving, site electrical, landscaping, utilities.

The work is part engineering and part tax: people who read construction drawings and price the work, and people who know how a classification is defended. It is not produced by a tax preparer from a closing statement, and the preparer’s credentials are the first thing anyone questioning the report asks about.

Where the component detail comes from when you built the building

On a purchased building, the study starts from a price. The buyer paid one number for a finished property, and nobody in that transaction holds a record of what the cabinets or the site electrical cost. So the estimator reconstructs: walks the property, counts what is there, prices each component from published cost data for comparable construction, and allocates the purchase price across the result. That is a legitimate and accepted method. It is also a reconstruction fitted backwards into a negotiated price.

On a building the sponsor built, the primary records already exist. The general contractor’s schedule of values divides the job into cost codes. The buyout log records which package went to which subcontractor at what price. There are executed subcontracts, monthly pay applications showing the work in place, change orders with pricing attached, submittals naming the equipment actually installed, and as-builts showing where it went. An engineer working from those records is not estimating what a component probably cost. They are reading what it was invoiced at.

On the communities we develop, build and manage, we act as the general contractor rather than hiring one. The trades are still subcontracted, as on any job this size — what stays with us are the contracts and the invoices. What that arrangement changes while a building is going up is set out in what vertical integration changes for an investor. The point here is about paper. Our construction practice has plans and specifications complete at loan closing, aims to have three-quarters of the work bought out by then, and pays subcontractors within thirty days of a pay application. Those commitments exist for schedule and price reasons. Their by-product is a contemporaneous cost record for that building, held by the company that files the return.

Engineering identification, or an allocation

Two approaches sit at opposite ends of this work. One identifies assets and prices them from documentation. The other applies an assumed split, derived from studies of similar buildings, to a total. Both produce a schedule. Only one explains itself.

What an examiner wants is documentation: which assets were identified, how each was quantified, where the cost came from, and who did the work. A study built on subcontractor pricing answers by pointing at the invoice. One built on an assumed share has to defend the assumption.

Neither method is always wrong. On an older acquisition with no surviving cost records, an estimate is the only option. The claim here is narrower: where the sponsor built, the records that make the identification defensible already exist.

What the deliverable is, and who reads it

The output is a report, longer and duller than people expect. Four things are in it.

• A fixed asset schedule. Every component identified, its quantity, the cost assigned, the class life given, the date placed in service.
• A statement of method. Which approach was used, why, and which records were relied on.
• The supporting evidence. Extracts from the cost records, drawings, photographs of the installed work.
• The preparer’s qualifications. Who did the engineering, and who reviewed the classification.

Three parties rely on it, and the investor is not among them. The partnership’s accountant carries the asset schedule onto the depreciation schedule, which produces the depreciation figure on the return and the loss allocated on each K-1. An examiner reads it if the return is examined. Where a fund is audited, its auditor tests the classification.

It is read once more at sale, where the same schedule decides how much of the gain is recaptured and with what character. The investor sees one line on a K-1 and none of the report behind it.

Placed in service, building by building

Depreciation begins when an asset is placed in service — ready and available for its intended use — not when it is paid for, and not when the project is finished.

A ground-up community rarely has one such date. A certificate of occupancy is issued building by building, so the first building can be leasing while the last is still framed; the stages a development runs through puts that in sequence. Each building starts its own clock, and those days can fall in different tax years. Three consequences land on the study.

• Cost has to be attributed to the right building. A code charged to the wrong building puts a deduction in the wrong year. That is construction accounting before it is tax work, and a further argument for records kept by the builder.
• Shared cost has to be split. Site work, utilities and the amenity building serve the whole community, and the study has to state the basis it allocates them on.
• The year the deductions arrive in follows the construction schedule. A building that opens later than underwritten moves its deductions into a later tax year.

The year a loss lands in is a construction fact, not a tax election.

What the study changes, and what it does not

Reclassification moves cost out of the long structural life into shorter classes, where a share may be deducted in the year the asset is placed in service rather than across decades. The deduction is computed on depreciable basis, which is funded by debt as well as equity, so it can be large relative to the cash an investor committed. Why new development produces large paper losses sets out both.

What the study does not do is create a deduction. The depreciable basis is the same before it and after. What changes is which years it is deducted in. Faster deductions early mean smaller deductions later, and a larger share of the gain at sale recaptured as ordinary income rather than capital gain. That is a change in timing and in character, not a permanent reduction.

Whether the loss does anything for a particular investor is a separate question. A passive loss usually cannot reach W-2 income, and that is settled by facts about the investor rather than by the study. The after-tax return calculator runs the year-by-year schedule on your own figures and shows its working.

What a study costs, and when it earns its fee

Where the fee is an expense of the partnership, the investors bear it, and it is incurred whether the classification turns out large or small. It scales with the work rather than with the answer, so the question is whether the basis being studied is large enough for the reclassification to be worth more than the study. A ground-up community generally is; a single small building may not be.

A study also creates ongoing work: the asset schedule has to be maintained, and a replaced component retired off it rather than left sitting beside its replacement. That is the part most often skipped, and the error surfaces at sale.

Questions worth asking about a specific deal

• Who performed the study, and what are their engineering and tax credentials?
• Was it engineering-based, and which records did it read — this building’s construction cost records, or an estimate fitted to a total?
• How was cost attributed between buildings placed in service in different tax years, and how was shared site cost allocated?
• Which year does the projection assume each building opens in, and what moves if construction runs long?
• Is the study fee an expense of the fund, and was it inside the budget you were shown?

A sponsor can answer those. What the schedule then means for your own return is a question for your accountant.

The study does not make the building work

A cost segregation study is an accounting exercise performed on a finished asset. It changes when deductions are taken. It does not change whether the asset should have been built.

Our pro formas use untrended rents, and if a deal does not make sense at today’s rents it is not good enough for us to build — why an untrended rent can be checked against the leases being signed nearby covers what that does during lease-up. That test comes before any tax analysis.

Private real estate development is illiquid and speculative. It carries construction delay and cost-overrun risk, lease-up risk, interest-rate and refinancing risk, and the risk of losing some or all of the capital invested. No return is guaranteed. Tax law changes, and this treatment changes with it. A cost segregation study is a position taken on a return: it can be examined, and it can be adjusted. None of this is tax advice about your own situation — the study belongs to the partnership, but the return it lands on is yours.