Analysing a Short-Term Rental Deal in 2026

Underwriting is the part of this business that actually determines outcomes, and it is the part most buyers do least. We screen roughly a thousand deals a week and eliminate about 98%, and almost all of that elimination happens in the analysis rather than at the viewing. This is where it stood in 2026, which was the year the tax strategy is back at full strength and market selection decides everything.

What 2026 changed

2026 is the first full year with 100% bonus depreciation permanently restored under OBBBA. For a high earner buying a property that clears the seven-day average stay test and where they materially participate, the first-year deduction is back to where it was in 2021. What is not back to 2021 is the market: supply is deeper, regulation is tighter in the places that tightened, and buying badly is no longer covered by a rising tide.

The spread between markets is wider than at any point in this period. Arizona and Tennessee are workable and stable. California and much of Colorado are not, for reasons that have nothing to do with demand.

Buyers are underwriting to current financing costs rather than to a hoped-for future, which has made pricing more rational than it was in either the boom or the correction.

How it works

Underwriting is the part of this business that actually determines outcomes, and it is the part most buyers do least. We screen roughly a thousand deals a week and eliminate about 98%, and almost all of that elimination happens in the analysis rather than at the viewing.

1. Assemble twelve to twenty genuinely competitive listings: same submarket, same bedroom count, same amenity tier, and pull twelve months of booked nights and rates.
2. Model twelve individual monthly revenue figures rather than dividing an annual number by twelve, because seasonality is the whole shape of the risk.
3. Build the complete expense stack: debt service at the actual rate, property tax reassessed at your purchase price, insurance quoted for the address, management all-in, cleaning per turnover, utilities at rental-use levels, dues, and reserves.
4. Stress it: revenue at 75%, three lost peak weeks, insurance 40% higher, and for city properties the long-term rental floor.

What that meant in 2026 specifically

The spread between markets is wider than at any point in this period. Arizona and Tennessee are workable and stable. California and much of Colorado are not, for reasons that have nothing to do with demand.

The risk in 2026 is the same one that has been true throughout: buying on the tax benefit rather than on the property. A permanent 100% deduction makes a good purchase excellent and does not make a bad purchase acceptable.

2026 rewards market selection and basis. The tax side is as favourable as it has ever been, which means the differentiator is everything else.

Where it goes wrong

The failure modes are consistent across years, which is itself useful information: they are not caused by the market cycle, so a different year does not protect you from them.

• Treating a seller's proforma as data. They are routinely built by extrapolating peak-season rates across twelve months.
• Building a comparable set on bedroom count and distance alone, so the subject property is compared against inventory it does not actually compete with.
• Omitting the maintenance and capital expenditure reserves, which is the most common reason a projected return exceeds the realised one.

What a buyer should have done in 2026

2026 rewards market selection and basis. The tax side is as favourable as it has ever been, which means the differentiator is everything else.

The underwriting discipline does not change with the year. Twelve individual monthly revenue figures from a comparable set you assembled, a complete expense stack including reserves, and a stress test at 75% of projection that still covers debt service.

We screen roughly a thousand deals a week and eliminate about 98% of them. That ratio has held across every year on this site, through the boom, the correction and the stabilisation, because it is a function of how listings are selected rather than of the market.

What generalises, and what does not

Reading a year in isolation is the most common analytical error in this business. 2026 had its own conditions, and someone who learned the wrong lesson from it carried that lesson into a market that no longer rewarded it.

What generalises is the mechanics above. The definitions, the tests, the sequence and the failure modes are the same in every year on this site, which is why they are worth learning properly once rather than relearning each cycle.

What does not generalise is the environment: the cost of capital, the depth of supply, the bonus depreciation percentage, and the regulatory posture of a given jurisdiction. Those change, sometimes abruptly, and a model that treats them as fixed is a model that was only ever right about one year.

The practical consequence is to build the analysis so the environment is an input rather than an assumption. A property that only works at one interest rate, one occupancy level and one tax treatment is not an investment thesis, it is a bet that nothing moves.

Frequently asked questions