Vacation and Second Home Statistics (2026)

The Census Bureau does not publish a category called second homes. It publishes two categories that between them capture most of them.

The first is vacant units held for seasonal use, which was 3,425,000 units in the second quarter of 2026, or 2.3% of the housing stock. The second is units held off market for occasional use, at 1,964,000 units or 1.3%.

Together that is roughly 5.4 million homes out of 149,454,000, or about one in 28. Neither category is defined by who owns the property, so both include homes that are rented out part of the year.

For comparison, there were 86,985,000 owner-occupied units and 46,827,000 renter-occupied units in the same quarter.

The entire vacant stock was 15,643,000 units, or 10.5% of housing, of which seasonal units are roughly a fifth.

The homeownership rate was 65.0%, virtually unchanged from a year earlier, so the second-home stock is not moving in step with primary ownership.

Seasonal units fell by 18,000 over the year and occasional-use units fell by 44,000, while the overall housing stock grew by 1,432,000 units.

That is a small decline in absolute terms and it runs against the direction of the stock as a whole, so the second-home share of housing edged down.

Both changes sit inside their margins of error, which the Census publishes alongside the estimates. The honest reading is that this category is flat to slightly down rather than collapsing or booming.

The single most valuable provision for a second-home owner is what the IRS calls minimal rental use. If you use a dwelling unit as a residence and rent it for fewer than 15 days in the year, you do not report any of the rental income and you do not deduct any rental expenses.

That is genuinely tax-free income, and it is the reason letting a house during a local event for a week or two is treated differently from running a rental.

The threshold is a cliff rather than a slope. Renting for 15 days rather than 14 moves the whole year into the reporting regime.

You are treated as using a dwelling as a residence if your personal use exceeds the greater of 14 days or 10% of the total days it is rented to others at a fair rental price.

A day of personal use includes use by you, by anyone with an interest in the property, by a family member unless it is their main home at a fair rental price, and by anyone paying less than a fair rental price.

It also includes days used under an arrangement that lets you use somebody else's property in exchange. That clause catches informal swaps that owners rarely think of as personal use.

Once a property is used for both rental and personal purposes, total expenses must generally be divided between the two based on the number of days used for each purpose.

The IRS worked example is instructive: a cottage available 92 days, rented 85 of them, with 14 days of family use, produces a rental share of 85/99, or 86%, of the cottage expenses. Days available but not rented do not count as rental use.

A weekend the owner used while a tenant was paying for the month counted as rental use for the expense split, because a fair rental price was received, and as personal use for the residence test. The same days can be classified differently for the two purposes.

Deductible rental expenses cannot exceed the gross rental income limitation: gross rental income less the rental portion of mortgage interest, real estate taxes, casualty losses and rental expenses such as agent fees and advertising.

In practice that means a second home rented at a loss usually cannot generate a deduction against your other income in the way an investment property might.

Excess expenses can be carried forward to the next year, subject to the same limitation in that year, so they are deferred rather than lost.

Short-term letting platforms have made it straightforward to rent a second home for part of the year, which pushes more owners across the 15-day threshold without them treating it as a decision.

Crossing it converts a personal asset into something with an annual filing consequence, a day-counting obligation and an expense allocation to maintain.

The record-keeping is the real cost. Personal days, rental days, fair rental price and days available but unrented all have to be tracked, and none of it can be reconstructed convincingly a year later.

A second home carries a full set of costs whether or not anybody is in it: mortgage interest, property tax, insurance, utilities, maintenance and often association fees.

Unlike a primary residence, it does not save you rent. The return has to come from use value, from rental income, or from price appreciation.

Rental income that crosses the 15-day threshold arrives with the tax treatment above, which is why the honest way to model a second home is on the after-tax rental figure rather than the headline nightly rate.

They count units, not owners. One person owning three seasonal properties appears three times.

They say nothing about value, mortgage status or whether the unit is rented, because the Housing Vacancy Survey is a count of occupancy status rather than a survey of owners.

They also exclude second homes that are occupied by somebody at the time of the survey, so a property with a long-term tenant is counted as renter-occupied rather than as a second home.

The Housing Vacancy Survey is produced quarterly by the Census Bureau, benchmarked to the 2020 Census, with estimates published in thousands.

Every line carries a margin of error, which the Census publishes alongside it. For the seasonal estimate the margin is 216 thousand units, which is larger than the annual change.

That is why a single quarter's movement in this series should not be read as a trend, and why this page reports the level with its uncertainty rather than the change on its own.

All housing unit counts, shares and year-over-year changes are from the Census Bureau's Housing Vacancy Survey for the second quarter of 2026, Table 3.

All tax rules, thresholds and the worked expense-allocation example are from IRS Topic no. 415, Renting residential and vacation property, and Publication 527.

Nothing on this page is a projection or an estimate of our own.