When a short-term rental loss is examined and material participation is the issue, the request that arrives is narrow and predictable: show the hours, show who else worked on the property, and show something outside the log that corroborates it. Requests for records are made in writing, usually as an Information Document Request with a response date, and they tend to arrive in stages rather than all at once. The useful thing about that predictability is that you can assemble the whole answer before anyone asks for it.
Nothing below describes how a particular examination will go or what any examiner will accept. It describes the kind of records the regulations put at the centre of the question, so you can keep them while they are still easy to keep.
Why the hours are what gets asked about
A short-term rental that falls outside the definition of a rental activity in Treas. Reg. §1.469-1T(e)(3) is not automatically non-passive. It still has to clear a material participation test in Treas. Reg. §1.469-5T, and the loss is reported through Form 8582 either way. So the file splits into two questions: whether the average stay puts the property outside the rental definition, and whether the owner participated materially. The first turns on booking data. The second turns on a record of hours that someone other than the owner can read.
Treas. Reg. §1.469-5T(f)(4) is the paragraph that shapes every request. It says the extent of participation may be established by any reasonable means, and that contemporaneous daily time reports or logs are not required where participation can be shown by other reasonable means — appointment books, calendars, or narrative summaries among them. It also says that reasonable means does not include a post-event ballpark estimate. That is the whole tension: the rule is flexible about format and unforgiving about invention.
The usual sequence
Requests in this area generally work from the general to the specific, and each stage is shaped by what the last one produced.
- The log. A record of dates, the property, what was done, and the time spent, for the year or years under examination.
- The basis for the log. When it was written, and from what. A log produced as the examination opens invites a question a log kept weekly does not.
- A sample of entries, corroborated. Rarely all of them. A handful of dates, with whatever exists outside the log that puts you at the property doing that work.
- Everyone else's hours. Cleaners, co-hosts, managers, handymen, family.
- The average-stay evidence. The booking list the 7-day figure comes from.
- Personal-use days. Counted under a different code section, asked about in the same breath.
A package assembled in that order answers most of it in one pass.
The log itself, in a readable format
A log answers the request when a stranger can read a line and know what happened. Date, property, task described in its own words, start and end time, total hours, and who did the work. A spreadsheet does this if it is disciplined. A screenshot of a phone's notes app usually does not, because the dates belong to the file rather than to the entries.
Two things make a log easier to hand over than to explain. The first is a total per property alongside the test being claimed, so the arithmetic is visible rather than implied. The second is any evidence of when each entry was created. If your records live in STR Tracker, the detailed PDF report prints every entry in the date range with its notes and its attached photos and receipts inline, so the log and its corroboration arrive as one document instead of a folder of files someone has to reconcile.
Corroboration for a sample of entries
Corroboration is the part owners under-build, because the log feels like the work. It is the independent trace: a hardware-store receipt timestamped inside the hours you logged, a guest message thread, a photograph of the repair, a cleaner's invoice, a toll or mileage record, a permit, a delivery confirmation.
The point is not to corroborate everything. It is that a sample drawn at random should land on entries that have something behind them. Pick five dates from your own log right now and see what you could produce. The gaps you find are the habit to change, and changing it costs nothing in October that it would cost in an examination. Our guide on documentation to keep for a short-term rental position goes through the categories in more depth.
Other participants' hours
The 100-hour test in Treas. Reg. §1.469-5T(a)(3) compares your participation to that of every other individual, so the hours worked by other people are not a side note. They are half of that test, and they are the half owners usually cannot produce.
Cleaners and managers are the common blind spot. A cleaning company that turns your property forty times a year is a number you can reconstruct from invoices if you kept them, and cannot if you did not. Ask for hours on the invoice, or record the turn count and the typical duration as you go. Where a manager runs the property, the comparison is often the deciding fact, and "I do not know what they spend" is not a neutral answer.
What a weak answer looks like
Patterns that create follow-up questions rather than closing them:
- Round numbers everywhere. Every entry at 2.0 or 3.0 hours reads as estimated, because real work is not shaped that way.
- Hours that do not fit the calendar. Logged turnover time on a date with no checkout, or a repair during a week the property was occupied.
- Durations out of proportion to the task. Six hours to replace a shower head.
- Investor-type work counted as participation — reviewing statements, studying the market, researching the next purchase. IRS Publication 925 and §1.469-5T(f)(2)(ii) treat that work as a different category.
- One log covering several properties with no way to split it, when the election to group has not been made.
- A log created in a single sitting after the notice arrived, with no earlier source behind it.
None of these decides anything by itself. Each one shifts the conversation from your hours to your credibility, which is a worse conversation to be having.
Assembling the package once
The work is the same whether you do it in January or under a response date, and it is cheaper in January. Build one bundle per property per year: the log with per-property totals and the test claimed, the corroboration sample, other participants' hours with their source, the booking list behind the average stay, and the personal-use day count. Our post on what to send your CPA covers the same bundle from the adviser's side, and the free log template gives you the columns if you are starting from nothing.
Then keep it. Records supporting a loss can matter well past the year they were created, particularly where losses carry forward, so the folder that closes a year is worth more than the effort of closing it properly.
What to write down, starting today
For each entry: the date, the property, what you actually did in plain words, the clock times, and the person who did it. Separately, once a month: the hours other people worked, where that number came from, and anything that corroborates a few of your own entries. Ask your CPA or EA which test you are aiming at for each property before the year closes, because the record you need depends on the answer.