Do You Need an Engineer for a Cost Segregation Study?

The legal answer: no requirement, but a stated preference

Nothing in the Internal Revenue Code or Treasury regulations requires a licensed engineer to sign off on a cost segregation study. That is the literal legal answer. But the IRS's Audit Techniques Guide (Pub 5653), written for its own examiners, ranks methodologies by reliability, and engineering-based approaches, whether from actual cost records or from a detailed cost estimate, sit at the top. Cost-estimating or survey-based approaches that apply default percentages by property type sit lower. The guide does not ban the lower tiers. It tells examiners to weigh them differently.

That distinction is the entire practical answer to whether you need one. You are not legally required to hire an engineer. You are choosing between a method the IRS's own reviewers trust more and one they trust less, and that choice matters more as a property gets more complex.

What an engineer actually does that a form cannot

A questionnaire or software tool takes basic facts, square footage, property type, purchase price, and applies a database percentage built from similar properties. It never looks at your building. An engineer or trained reviewer does the opposite: reviews the specific drawings, photos, or site conditions and classifies each reclassifiable item individually. Carpet, most flooring, decorative lighting, cabinetry, appliances, window treatments, and certain electrical or plumbing serving specific equipment go into the 5-year bucket. Certain fixtures and furniture go into 7-year. Paving, fencing, landscaping, site utilities, and outdoor lighting go into the 15-year land improvement bucket. Everything else, the structural shell, stays at 39-year (commercial) or 27.5-year (residential), including the roof and central HVAC, which are structural components even though owners commonly assume otherwise.

That specificity is exactly what a defensible study is built on. If a classification is later questioned, the answer is not "the database says restaurants typically run this way," it is "this specific piece of equipment, in this specific kitchen, meets this specific test."

Where the difference matters most

Complexity is the variable that decides how much engineering actually buys you. A simple, low-complexity property, standard residential rental with typical finishes, has less variance from the database average, so a percentage-based estimate can land in a reasonable range. A restaurant with kitchen equipment, a medical building with specialized systems, a hotel, or a property with unusual site work has enough building-to-building variance that a default percentage stops representing the actual property, and the gap between what a form assumes and what the building actually contains grows.

See what the price difference between engineered and software studies actually buys for the full breakdown of that tradeoff, and where DIY approaches are genuinely reasonable for the honest floor on how simple a property has to be before self-serve makes sense.

A concrete example of what engineering catches

A real quoted Free-Standing Restaurant study shows the gap plainly. On a $2,804,440 building basis, the engineered study identified $599,678 in first-year deductions, a 66.6:1 ratio against its $9,000 fee. A generic restaurant percentage from a database would apply the same reclassification rate to every restaurant regardless of its actual kitchen layout, whether it has a walk-in cooler or two, how much of the electrical serves cooking equipment specifically versus general lighting, and how the site was built out. The engineered number came from someone looking at this restaurant's actual equipment and wiring, not the category average for restaurants generally.

A Mid-Rise Office study, by contrast, on $2,971,345 in basis, produced $479,220 in first-year deductions, a 39.9:1 ratio. Both are real, both are engineered, and both land at different percentages of basis because the buildings themselves are different, exactly the variation a database percentage is built to average away rather than capture.

How to check whether a provider actually uses an engineer

Since there is no legal requirement, the only way to know whether a study is engineering-based is to ask directly. Who classifies the components, by name or role, not just "our team." Do they review the specific building, drawings, photos, or a site visit, or do they work from square footage and property type alone. Does the deliverable show individual components with their own cost allocation, or one summary percentage. A vague or evasive answer to any of these is worth treating the same way as the other red flags worth watching for when vetting a cost segregation firm.

How Basis handles this without site visits on STRs

For commercial properties, engineering-based classification generally means a site review or a detailed review of drawings and photos. For short-term rentals, the same component-level classification happens without a site visit and without owner homework: the listing photos already on Airbnb or VRBO feed the classification directly. It is still property-specific engineering, just built around the documentation an STR owner already has rather than requiring a new site visit.

Both Basis tiers, budget and full engineered studies, deliver the same 70-page engineered report format, aligned to the IRS's Audit Techniques Guide. The tiers differ in scope, not in whether an engineer touches the classification.

Getting a property-specific number instead of a database guess

A free Preliminary Benefit Estimate at /qualify models a building's likely first-year acceleration using an engineering-based approach before any commitment, so the projected number reflects the specific property rather than a category average.

Whether engineering-based classification is worth the added scope for your particular property is a question you can weigh against the estimate itself, and against your CPA's read on how much documentation your situation calls for. What the estimate shows is the number the mechanics produce once someone actually looks at the building, the restaurant's kitchen, the office's finishes, the rental's listing photos, rather than a category average.

Frequently asked questions

Is it illegal to skip an engineer for cost segregation?

No. There is no statute or regulation requiring an engineer. The IRS's Audit Techniques Guide simply ranks engineering-based methods as more reliable than cost-estimating or questionnaire approaches, particularly for complex properties.

What kind of engineer does a cost segregation study?

Typically someone with construction, engineering, or cost estimating expertise who can review a building's actual components, structural drawings, electrical and mechanical systems, and site improvements, and classify each one against IRS class-life rules.

Do short-term rental studies require a site visit from an engineer?

No. STR studies use the property's existing listing photos from Airbnb or VRBO to classify components, so there is no site visit and no additional owner homework required, even though the classification work is still property-specific rather than a database average.

Does hiring an engineer guarantee my study won't be questioned?

No study guarantees zero examiner questions. Engineering-based classification means the documentation is stronger and more specific to the building, which is what a review process actually evaluates if questions do come up.