How to Set Rent for a Rental: Comps, FMR, Vacancy

The listing sat for forty-one days at $1,395. I dropped the rent to $1,325 on a Tuesday and had a signed lease nine days later — same photos, same house, same week of the month. Giving up seventy dollars costs me $840 over a twelve-month lease. The forty-one vacant days had already cost about $1,880.

I still don't know whether the price was the reason. Somebody may have been about to call at $1,395. What I do know is the arithmetic, and that is the only part of this knowable in advance.

Two rented houses, no licence, no manager. Everything below was checked on 22 August 2026, and the HUD figures are dated — read the section on that before you use them.

Nobody has three comparable houses, and appraisers know it

The three-comparables idea comes from sales appraisal, where an underwriter needs a defensible number and there are recorded transactions to work from. Rental pricing has neither condition. No public record exists of what the house down the street actually rented for. There is only what somebody asked.

That distinction is the whole problem. An asking rent is an offer. The distribution of asking rents in your area is a distribution of offers, and it is wrong in a specific direction: overpriced units stay in it for months, while underpriced ones vanish in four days. Scroll a rental site on any given afternoon and you are looking at a sample biased toward the prices nobody accepted.

The first time I priced a unit I averaged nine asking rents and felt rigorous about it. I never wrote down which of the nine were still listed three weeks later. Six of them were. I had carefully computed the average of a set of prices that were, mostly, not working.

Fifteen asking rents, and the date each one disappears

Build the sample yourself, over about three weeks, before the sign goes up. Fifteen entries is enough for a single house. You are not doing statistics. You are watching a small market move.

Define the box first, and write it down so you don't quietly widen it when the results are inconvenient:

• Same bedroom count. This dominates everything else in one-to-four-unit housing.
• Within about 1.5 miles, or the same school attendance area, whichever is tighter.
• Within roughly 25 percent of your square footage.
• Listed within the last 60 days.

Then record, for each one: asking rent, beds and baths, square footage, which utilities the tenant pays, pet policy, the date you first saw it, and the date it stopped appearing. Check the list twice a week. That last column is the one everybody skips and the only one that turns offers into evidence.

Where to look: the big rental portals, Craigslist and Facebook Marketplace in smaller markets, the sites of local property management companies, and the listing board your regional housing authority points voucher holders to. Managers price to lease quickly, because vacancy is their cost too — their numbers are worth more than a private owner's hopeful posting.

Two cautions. Price per square foot is unreliable at the small end — a 900-square-foot two-bedroom and a 1,250-square-foot one often rent within fifty dollars of each other. And what you paid for the house has nothing to do with what it rents for.

Fair Market Rent is not market rent, and it is not your rent either

HUD publishes a rent estimate for every metropolitan area and non-metropolitan county in the country, free, with documentation. Useful as a check on your own sample. Also misread constantly, in two specific ways.

It is a 40th percentile figure. 24 CFR 888.113 sets FMRs at "the dollar amount below which the rent for 40 percent of standard quality rental housing units fall," drawn from units occupied by recent movers, with public housing and substandard units excluded. That is deliberately below the middle. HUD also publishes 50th percentile rent estimates for the same areas, and comparing the two shows the size of the gap:

Computed from two HUD spreadsheets, both free: FY26_FMRs_revised.xlsx, column fmr_2, on the FMR data page, and FY2026_FMR_50_county.xlsx, column rent_50_2, on the 50th percentile page. The revision effective 21 May 2026 moved seven areas; none of these six is among them, so the original FY2026 file gives the same figures. Six to eight percent — enough to matter, not enough to justify ignoring the FMR.

It is a gross rent. The same regulation defines FMRs as "estimates of rent plus the cost of utilities, except telephone." If your tenant pays their own heat, electric, water and trash, HUD's number covers all of that and your contract rent does not. To compare like with like, get the utility allowance schedule from your county's public housing authority — 24 CFR 982.517 requires every PHA to keep one, priced category by category: space heating, air conditioning, cooking, water heating, water, sewer, trash collection, other electric, and a few more besides. Subtract the categories your tenant pays. Which is a good moment to be certain who pays for what in your own lease, since that allocation is also the one that decides who calls the plumber.

Dates matter here. FY2026 FMRs took effect 1 October 2025, and HUD published revised FY2026 FMRs for seven areas effective 21 May 2026. As of 22 August 2026 the FY2027 tables were not out. 42 U.S.C. 1437f(c)(1)(B) says FMRs "shall become effective no earlier than 30 days after the date of such publication"; HUD's practice is to have the new year's tables in force on 1 October. So if you are reading this after roughly 1 September 2026, go look for a newer file.

One metro number covering eighty-eight ZIP codes

The county-level FMR is a single figure for an entire metropolitan area, which is exactly the resolution you don't want when the question is one house on one street.

HUD's Small Area FMRs fix that, and the spread inside a single metro is the most useful thing on the whole site. The Dayton–Kettering–Beavercreek MSA has one two-bedroom FMR of $1,273. The FY2026 SAFMR file — fy2026_safmrs_revised.xlsx, two-bedroom column — carries 88 ZIP codes for that metro, running from $970 to $1,970. The top of the range is more than double the bottom. Fort Wayne runs $960 to $1,670 across 63 ZIPs. Grand Rapids, $1,300 to $2,300 across 50.

Look up your own ZIP before you take the metro figure seriously. If the SAFMR for your ZIP sits 40 percent above the metro FMR, the metro FMR was never describing your street.

What one percent of rent buys you, in days

Here is the piece that settles most pricing arguments, and it is one line of arithmetic.

A month of rent is roughly 30.4 days, so an empty unit costs the monthly rent divided by 30.4 every day. Cutting the rent costs you the amount of the cut every month for as long as the tenancy lasts. Set the two equal and you get the break-even:

Days of vacancy the cut must save = (cut ÷ rent) × months of tenancy × 30.4

A one percent cut over a twelve-month lease works out to 3.65 days. Call it 3.7 days of vacancy per 1 percent of rent, per year the tenant stays.

For a unit at $1,300 a month, where a vacant day costs $42.76:

It is linear, which is what makes it usable in your head: at $1,300, every $25 off the asking price has to buy a week — for each year the tenant stays.

Run my forty-one days through it. Seventy dollars off $1,395 is 5.0 percent, so the break-even over a twelve-month lease is about 18 days of vacancy avoided. If the cut really did save 29 days — if the lease that arrived on day 50 would otherwise have arrived on day 79 — it won by roughly $490. But if that tenant renews and stays two years, the break-even doubles to 37 days, the cut never earns it back, and I am down about $350. The same decision is right or wrong depending on something I could not know when I made it.

That asymmetry is the practical part. A price cut is a permanent concession bought with a one-time saving. It pays off best where tenants turn over quickly and worst in exactly the situation most small landlords say they want — a good tenant who stays five years. For context, the Census Bureau put the rental vacancy rate at 7.3 percent nationally in the second quarter of 2026, 9.5 percent in the South and 5.3 percent in the West — release CB26-116, 28 July 2026. Regional averages will not price your house. But a market at 9.5 percent vacancy punishes an optimistic price far harder than one at 5.3.

A month free and fifty dollars off are different instruments

One month free on a twelve-month lease is worth $108 a month on a $1,300 unit — a shade more than a $100 price cut, and cheaper than most landlords assume when they hear "free month." But the reason to prefer one over the other usually isn't the arithmetic.

A concession keeps the headline number intact. That matters in three places: the renewal conversation starts from $1,300 rather than $1,192; your next listing has a defensible asking price; and anything in your paperwork keyed to "one month's rent" — including the security deposit cap in the states that write it that way — is computed off the higher figure.

Some states now regulate how you write it down. California's Civil Code § 1947.12 requires that "rent discounts, incentives, concessions, or credits" be separately listed and identified in the lease, and excludes them when determining the lowest gross rental rate that future increases are measured against. Which is the outcome you wanted anyway — but only if the concession is documented as a concession instead of buried as a lower rent for one month. That is a drafting question, and drafting questions are where clauses quietly fail.

The number you picked may not be the number you can keep

Setting the initial rent on a genuinely new tenancy is usually unrestricted. California says so in as many words: § 1947.12(b) lets an owner establish the initial rental rate free of the cap where no tenant from the prior tenancy remains in lawful possession.

The increases afterward are a different matter, and that is what turns a pricing decision into a multi-year one.

• Oregon caps the annual increase at the lesser of 10 percent, or 7 percent plus CPI. ORS 90.324 makes the Department of Administrative Services publish next year's figure by 30 September; for 2026 it is 9.5 percent, down from 10.0 percent in 2025. ORS 90.323 also bars any increase during the first year and requires 90 days' written notice after that. A unit whose first certificate of occupancy issued less than 15 years before the notice is out from under the percentage cap — the cap only, not the notice period or the first-year bar.
• The workaround is closed there too. ORS 90.323(4) says a landlord who terminates a tenancy without cause in the first year may not charge the next tenancy more than the cap would have allowed. The initial rent is only free when the vacancy is.
• California caps at 5 percent plus the local change in the cost of living, or 10 percent, whichever is lower, with exemptions for newer construction and some single-family homes.
• Your city may have an ordinance on top, and many states preempt local rent regulation outright. Search your own legislature's site for "rent" and "increase" before assuming either way.

None of this is legal advice and I am not a lawyer. It is the list of places where the number you chose stops being yours.

Three weeks before the sign goes up

The sequence that costs nothing but calendar time:

1. Week one. Define the comp box in writing. Start the sheet: fifteen listings, all seven columns, including who pays utilities and the date first seen.
2. Week one, same afternoon. Look up your county FMR and your ZIP's SAFMR at huduser.gov, pull the 50th percentile figure for the same area, and get the utility allowance schedule from your county housing authority. Convert HUD's gross number into a contract-rent equivalent for your utility split.
3. Weeks two and three. Re-check the sheet twice a week and fill in the disappearance dates. Treat anything still listed after 30 days as a ceiling, not a comparable — the market has already declined that price.
4. Before you list. Take the median of the entries that left the market inside three weeks, and check it against the SAFMR-adjusted HUD figure. If the two disagree by more than about 15 percent, suspect your comp box before you suspect the data.
5. Write the fallback down now. Pick the date you will reprice and the size of the step, using the days-per-percent table, while you are still calm about it. Mine is day 21 and one step of 3 percent. Deciding it in week six, staring at an empty house, is how a $50 problem turns into a $200 one.