Expensive land shrinks the prize without changing the percentages. Depreciation applies only to improvements; land is excluded under IRC section 167 and IRS Pub 946. In a coastal or urban market where 60% of the purchase price is land, only 40% of what you paid enters the depreciation system, and a cost segregation study reclassifies slices of that smaller number. The study percentages hold up fine; the base they multiply against is what collapses.
How land allocation actually works
Purchase contracts rarely split the price between land and improvements, so the allocation comes from evidence: the county assessor's land-to-improvement ratio, a qualified appraisal, or a replacement-cost analysis. The IRS Cost Segregation Audit Techniques Guide expects the allocation to be supportable, and the assessor ratio is the default examiners check first. Claiming 20% land when the assessor says 55% is an invitation to an adjustment unless an appraisal backs you up.
One nuance cuts in your favor: land improvements are not land. Paving, fencing, landscaping, site utilities, retaining walls, and pools are 15-year MACRS property even though the ground beneath them is not depreciable. High-land markets still have driveways, decks, and site work, and a study captures them.
Side-by-side example
Two $1,500,000 purchases in 2026, both eligible for 100% bonus depreciation under IRC 168(k).
- Coastal duplex, 60% land. Depreciable basis is $600,000. A study reclassifies 25%, or $150,000, into short-life classes. At a 35% rate, the first-year deferral is $52,500.
- Midwest flex building, 15% land. Basis is $1,275,000. The same 25% reclassification is $318,750, deferring $111,563.
Same price, same study quality, same percentages, and a 2.1x difference in benefit driven entirely by land. The coastal deal still produces $52,500 against a study fee measured in thousands, so it clears, just with less margin.
When high land genuinely kills the deal
- Small properties in high-land markets, where basis after the land carve-out drops under roughly $200,000 and the fixed study fee eats the benefit.
- Short planned holds, where Section 1245 recapture unwinds a benefit that was modest to begin with.
- Losses you cannot use, with no REPS, no short-term rental treatment under Reg 1.469-1T(e)(3), and no passive income; the deferral arrives late and discounts the case for a paid study.
The move is to screen on depreciable basis before anyone quotes you a study. Amadae's cost segregation screen starts from your purchase price and a supportable land allocation, then produces a screening estimate of reclassification and first-year benefit. It is not an engineering study, and that is the point: in a high-land market it tells you cheaply whether the smaller base still pays.