REPS Hours Reconstruction: How to Build a Defensible 750-Hour Log From Bank Data and Mileage

What REPS actually requires

Real Estate Professional Status is defined in IRC §469(c)(7). Two tests, both required, for the taxpayer (or a filing spouse):

• More than 750 hours of personal services performed during the year in real property trades or businesses in which the taxpayer materially participates, and
• More than half of the personal services the taxpayer performed in all trades or businesses during the year were performed in real property trades or businesses.

Both prongs live in IRS Publication 925. The 750-hour threshold is a floor, not a ceiling — hitting 751 is fine, but the more-than-half test means a full-time W-2 job at 2,000+ hours makes REPS effectively unreachable for that spouse. This is why REPS-claiming STR households almost always designate the non-W-2 spouse as the qualifier.

REPS is different from the "STR loophole" material participation regime. The STR loophole (average guest stay ≤ 7 days, plus meeting one of the seven material participation tests in Temp. Reg. §1.469-5T) recharacterizes an STR from rental to non-rental business under Reg. §1.469-1T(e)(3), which unlocks losses without needing REPS at all. REPS is the broader, harder-to-hit status that unlocks losses across your entire real estate portfolio — long-term rentals included. Many operators pursue both paths for different properties in the same year.

Why reconstruction from memory fails

Every CPA who has defended a REPS audit will tell you the same thing: a hours log created after the fact, from memory, in a fresh spreadsheet with timestamps that all fall inside the week the CPA asked for it — is the fastest way to lose. IRS Publication 925 explicitly notes that the "extent of an individual's participation in an activity may be established by any reasonable means. Contemporaneous daily time reports, logs, or similar documents are not required if the extent of such participation may be established by other reasonable means."

"Other reasonable means" is the escape hatch that makes reconstruction possible. But there is a hard line between two categories:

• Reconstruction from memory: a spreadsheet you build in April for last year, filled in from what you can remember, with round numbers like "8 hours" appearing over and over. Zero contemporaneous data behind it. Audit-fatal.
• Reconstruction from contemporaneous proxies: a log built from bank statements, mileage records, PMS message threads, calendar events, and vendor receipts — all of which were created in real time as the work happened, and all of which independently corroborate that you were doing rental work on a specific date. Defensible.

The distinction matters because audits look at metadata. A spreadsheet whose file-creation timestamp is three days before your CPA meeting, showing an even 800 hours split neatly across 12 months, reads as fabrication. A log whose every hour cites a bank transaction, a mileage entry, or a message thread with a real timestamp reads as evidence.

The 4-source framework

Contemporaneous proxies come from four data sources. Every hour you claim should tie back to at least one of them.

1 · Bank & credit card transactions

Every business expense on your card is evidence that some amount of your time was spent on rental work that day. A Home Depot charge means you drove there, walked the aisles, picked out material, drove back, and did the repair. A propane fill means you drove to the property, coordinated the delivery, and confirmed the tank. A TurnoverBnB or Turno payment means you scheduled a clean, communicated with the cleaner, and verified turnover.

The IRS accepts this reasoning if — and only if — your per-transaction time estimates are conservative and consistent. A $75 Home Depot repair receipt does not defensibly justify 8 hours. It justifies 1.5 to 2.5 hours (drive + purchase + install + cleanup). Consistency across the year matters more than any single estimate.

2 · Mileage log

If you kept a mileage log — even a phone app like MileIQ that recorded trips automatically — every business drive gives you two things: drive time at your metro's average speed, and task time at the destination. A 42-mile round trip at ~30 mph metro average is ~1.4 hours of drive alone; add the on-site task time and a single property visit routinely lands at 3–5 hours. Mileage logs also help the more-than-half test because they timestamp when you were physically doing real estate work vs. sitting at your W-2 desk.

No mileage log for 2026? Reconstruction is still possible if you have Google Maps Timeline / Apple Significant Locations / Life360 history — those are contemporaneous location data, and courts have accepted them as mileage evidence in prior tax cases. Export the history for the year and filter to visits at property addresses and known vendor addresses (Home Depot, Lowe's, propane suppliers, hardware stores).

3 · PMS message threads

Every reply you sent to a guest, every check-in issue you handled, every maintenance ticket you triaged inside Hospitable / Guesty / Hostaway / OwnerRez / Lodgify is timestamped and permanently logged. Export the message history for the year and count messages by day. Each substantive reply defensibly justifies 5–10 minutes; each check-in issue or complaint thread justifies 15–30 minutes; a booking dispute or damage claim justifies 45–90 minutes.

PMS message threads are the single strongest source of contemporaneous evidence in an STR REPS defense because they are timestamped, they are stored on a third-party server outside your control (so they cannot be back-dated), and they show substantive rental work — not just financial activity.

4 · Calendar events

Google Calendar, Apple Calendar, and Outlook all store creation timestamps. Blocked time for property walkthroughs, contractor meetings, closings, inspections, and cleaner briefings — if it was on your calendar in real time — is contemporaneous evidence of hours. Even a simple 90-minute block titled "walkthrough — Joy On The Lake" that was created weeks in advance is defensible time.

Credible per-vendor time attributions

The framework only works if your per-transaction estimates match industry norms. If you claim 8 hours for every $50 Amazon supply order, you are not being audit-defensible — you are inflating. Below is a conservative range CPAs have used successfully. Adjust to your own operation.

Data source	Typical time	Notes
Home Depot / Lowe's receipt < $100	1.0 – 2.0 hr	Drive + shop + return trip + install
Home Depot / Lowe's receipt $100 – $500	2.0 – 4.0 hr	Larger repair, longer install
Home Depot / Lowe's receipt > $500	4.0 – 8.0 hr	Major project — pair with mileage + calendar
Amazon supplies < $50	0.25 – 0.5 hr	Order + unbox + deploy on next visit
Propane / utility fill	0.5 – 1.0 hr	Drive + coordinate delivery + confirm tank
Cleaner platform payment (per turnover)	0.25 – 0.5 hr	Schedule + brief + verify + pay
Handyman / contractor invoice	1.0 – 2.0 hr	Scope + supervise + inspect + pay — not billed labor hours
PMS guest message (substantive reply)	5 – 10 min	Per non-automated response
PMS check-in issue / complaint thread	15 – 30 min	Full triage cycle
Booking dispute or damage claim	45 – 90 min	Photos + narrative + platform escalation
Property visit (mileage log entry)	drive time + 1.5 – 3 hr on-site	Non-cleaning walkthrough or maintenance
Closing / lender / insurance meeting	1.0 – 2.0 hr	Pair with calendar event

None of these numbers are IRS-published. They are the ranges practitioners use because they are conservative enough to survive audit scrutiny and specific enough to reconstruct meaningfully. The pattern the IRS is looking for is reasonableness — hours per task that a person doing that task would actually spend, not hours that happen to add up to exactly 751.

The de-duplication problem

This is the trap that catches operators building reconstruction logs manually. You cannot count the same hour twice from two data sources. If on 2026-04-15 you have:

• A Home Depot charge for $87
• A MileIQ mileage entry for a 38-mile round trip to that Home Depot
• A calendar block titled "repair kitchen faucet"

That is one event, not three. The defensible time for that day is something like 2.5 hours — drive + shop + install + cleanup — total, corroborated by three independent sources. Not 2.5 hours for the receipt plus 1.4 hours for the mileage plus 90 minutes for the calendar block.

The correct dedup approach is date-and-property based: group every proxy record by (date, property), then assign one time estimate per group based on the highest-fidelity source available (mileage > calendar > receipt > message). Then add corroborating sources as evidence for that group, not as additional hours. This is fiddly to do manually across 5 properties × 365 days. It's the primary reason spreadsheet-based reconstruction usually fails: not because the concept is wrong, but because manual dedup at scale is a nightmare.

What RentReel's Hours Reconstruction Wizard does

The Hours Reconstruction Wizard inside RentReel is built around this framework. You upload your data in whatever form you have it — a mileage export from MileIQ, a bank/credit-card CSV, a PMS message-history export, an .ics calendar file — and the wizard:

• Reads each transaction and applies a conservative default hours estimate based on the vendor category (Home Depot vs. Amazon vs. propane vs. cleaner platform, each with its own default range)
• Reads each mileage entry, computes drive time from distance ÷ metro average speed, and adds a default on-site task time
• Reads each PMS message and counts substantive vs. automated replies
• Cross-references every source by (date, property) and de-duplicates automatically — one event per group, highest-fidelity source wins, others attach as corroboration
• Outputs a per-week hours log with source citations for every hour claimed — a real IRS-facing document, not a spreadsheet
• Flags weeks where the reconstructed hours seem thin (a red-flag report) so you know where to add calendar or mileage evidence before you file

The output is a PDF workpaper your CPA can attach to your return workpapers, showing 52 weeks of reconstructed hours with every hour tied to at least one contemporaneous data source with a timestamp predating your tax filing.

The audit-defense story

Three specific things the IRS looks for in a REPS defense, and how a properly reconstructed log addresses each:

Contemporaneous data sources with timestamps predating filing

Bank statements, PMS message logs, mileage-app exports, and calendar-app exports all carry metadata that shows the underlying data was recorded in real time. The reconstruction log itself can be built in March; the data the log is built from must not be. Every source cited in the log needs to be independently produceable on request.

Per-hour source attribution

A log that says "Week of 2026-06-15 — 22 hours" is unfalsifiable and useless. A log that says "Week of 2026-06-15 — 22 hours: Home Depot repair 4hr [receipt #4471], property walkthrough Joy On The Lake 5hr [mileage entry MI-4482 + calendar Jun 16], guest triage 3hr [Hospitable thread 8931 + 8934], cleaner scheduling 2hr [Turno payments Jun 17-19], propane fill 1hr [Amerigas receipt Jun 18], insurance renewal call 2hr [calendar Jun 19], vendor payments and books 5hr [QuickBooks activity Jun 20]" is defensible because every line item is verifiable against an independent source.

Reasonable per-task time estimates

An auditor comparing your reconstructed hours against industry norms — hours per repair, minutes per guest message, hours per property visit — should find your estimates unremarkable. Estimates that consistently sit at the low-to-middle end of the range read as honest; estimates that consistently push the upper end read as aggressive. Reconstruction that lands at exactly 751 hours reads as too neat. Reconstruction that lands at 823 or 891 hours reads as real.

What to have ready before you start

If you're about to run a reconstruction — whether manually or through RentReel's wizard — pull these five things first:

1. Every business bank and credit-card statement for the tax year, ideally as CSV exports
2. Mileage log or location history — MileIQ export, Google Maps Timeline export, or Apple Significant Locations. Best-case: a real mileage app; fallback: passive location history from your phone
3. PMS message history export from every platform you use — Hospitable, Guesty, Hostaway, OwnerRez, Lodgify, or direct Airbnb / Vrbo inbox exports
4. Calendar .ics export for the year from every calendar you use for property work
5. A written policy note to yourself and your CPA describing the reconstruction methodology you used — data sources, per-vendor time estimates, dedup rules. This becomes part of the audit workpaper

Reconstruction is not shortcut work. Done properly it takes 4–8 hours per property per year. Done improperly (memory, round numbers, no source citations) it takes 30 minutes and fails on first contact with an IRS examiner. The difference is not effort — it's the discipline of tying every hour back to a piece of contemporaneous data that existed before you sat down to build the log.