Real Estate Investor · Austin, TX · Member since 2017 · 142 posts · 35 votes
Every cost seg starts by carving land out of the purchase price, since land can't be depreciated. What I've seen in practice is three ways firms do the split: the tax assessor's land/building ratio, an appraisal, or an engineering estimate built into the study itself.
The split matters because every dollar that lands on land is a dollar that never gets depreciated, so a study that leans heavily on the assessor's ratio can understate the building basis — or overstate it, depending on the county.
Curious what others have seen: which method did your study use, and did your CPA push back on the land number?
Ideally your cpa should be helping you come up with the land number and have a defensible method for it . Assessor and appraisal are the two most common ways to allocate land/building.
Ideally your cpa should be helping you come up with the land number and have a defensible method for it . Assessor and appraisal are the two most common ways to allocate land/building.
The part I'd add is why the number matters in dollars, and one way the assessor method goes wrong. Say you pay $400,000. If the assessor shows land at 30%, that's $120,000 of land and $280,000 of building, which depreciates at about $10,182 a year over 27.5 years. If an appraisal says land is 20%, the building is $320,000 and the depreciation is about $11,636. That's roughly $1,450 a year, every year, from one percentage.
The catch with the assessor ratio is that the county's total value is rarely your purchase price. The usual approach is to take the assessor's percentage and apply it to what you actually paid, not to use the county's dollar figures.
Whatever method gets used, I'd want it written down at closing and used the same way on the tax return, the study, and any later sale. The thing that causes trouble isn't a slightly high land number, it's three documents with three different numbers. Aaron's point is right that your CPA should be comfortable defending it.
Sometimes an appraisal doesn't come with a cost approach. You could always reach out to appraisers and see what they'd likely come in at for a land percentage. If they can beat the tax assessment by a worthy amount, it makes sense to pay for the appraisal.
One thing that gets lost here is that the stuff on the land isn't land. Driveway, fencing, patio, retaining walls, and landscaping are 15-year land improvements and can take bonus, so they shouldn't get lumped into the land number.
On the split itself, the assessor ratio is fine when it looks reasonable for the area. It gets shaky where the county puts most of the value on the lot, or barely values the land at all. That's when an appraisal with a separate land value is worth paying for. Whatever method you use, keep the backup in the study file, since that's what gets asked about on exam.
The assessor’s ratio is the default the IRS tends to fall back on in an audit, so it’s the easiest number to defend. That’s also why we push back when a study moves meaningfully away from it without backup. An engineering estimate can absolutely be the better number, but if it puts less on land than the assessor did, we want an appraisal or comparable land sales in the file to support it. Otherwise it’s the first thing an examiner resets.
The bigger miss we see isn’t the land split itself. It’s site improvements getting lumped in with land. Paving, sidewalks, fencing, landscaping, and drainage are 15-year property and eligible for bonus, and if they’re folded into the land number, those deductions are lost for good. Pulling those out is usually where a good study earns its fee over just using the assessor’s ratio.