How it works

The 941 Index tracks single-family home prices in Sarasota and Manatee counties. It answers one question: what would the same house, in the same kind of place, sell for now compared with before? The 2016 average is set to 100, so a reading of 175 means that house costs 75% more than in 2016.

Which sales count

Recorded sales from both county property appraisers' public files, refreshed nightly.

• Qualified sales only. Each county's appraiser marks a sale qualified when it was an arm's-length sale at a market price. Distressed sales, family transfers, no-money deeds and multi-parcel deals are left out.
• Cash sales are included. A deed is recorded however the buyer paid, so cash buyers count. Indexes built from mortgage data leave them out.
• Single-family houses only. Condos behave differently and are left out. This index says nothing about condo prices.
• Resales only. Homes built in the sale year or the year before are left out. Builder prices carry incentives that don't track the resale market.

Holding the house constant

Each sale price is split into two parts: what the house is, and when it sold. Only the "when" moves the index. "What the house is" covers living area, lot size, age, bedrooms, bathrooms, pool, waterfront, golf frontage, stories, garage bays, CDD, the appraiser's quality grade, construction type and location. Assessed values and list prices are deliberately not used; they move with the market the index is trying to measure.

The model is refitted every week on the latest two years of sales. Sales far out of line with the rest are dropped in a second pass.

How location is handled

Each house is compared with others in its own neighborhood, as the county defines it. Where a neighborhood has fewer than eight sales in two years, the model falls back to the census tract, then the ZIP code. A month heavy with sales in expensive areas therefore doesn't lift the index.

Why it differs from other numbers

• Median prices and MLS statistics describe the homes that happened to sell. A month with more large or waterfront homes looks expensive even if nothing changed.
• Zillow's ZHVI estimates what all homes are worth, sold or not, from Zillow's own valuation model. The 941 Index uses only recorded sales. Quarterly changes in the two correlate 0.95.
• FHFA and Case-Shiller-style indexes use repeat sales: only houses that sold at least twice. FHFA's index is built from mortgages bought by Fannie Mae and Freddie Mac, so it leaves out cash and jumbo-loan purchases. Quarterly changes correlate 0.86 with FHFA's all-transactions index and 0.81 with its purchase-only index.

Those benchmarks agree with each other no better than that (0.78 to 0.85). On level, the 941 Index sits between them: FHFA higher, Zillow lower.

Why the decline from 2022 reads steeper

In the November 2026 backtest, the index peaked in May 2022 and stood 13.5% lower in August 2026. A repeat-sales index built from the same county records has the same peak month but a decline of 9.9%.

When the 941 method is run on only the houses that sold twice, the two methods agree within about 1% in every year tested. The gap comes from which houses are counted, not the method: houses that sold only once in the decade fell more than frequently traded ones, and a repeat-sales index cannot see them.

So the 941 Index describes every resale, and repeat-sales figures describe houses that trade often. When comparing, say which one you are quoting.

Lag and revisions

County records arrive after closing. Sarasota's arrive within about four weeks. Manatee's arrive in a batch six to eight weeks later, and 99% are in by week ten.

• Weeks stay provisional for ten weeks and are re-estimated as late sales arrive. On the chart, provisional months are dashed.
• After that they are frozen. Published history is never revised; the system refuses any run that would change a final month.
• The headline is the latest four weeks with enough sales, weighted by sales count. It can trail the calendar by a few weeks so it doesn't rest on Sarasota alone.

In a replay of 103 months, with each sale entering when it would actually have reached the county file, frozen months landed within 0.06% on average (0.32% at worst) of the value computed from complete records.

What the bands mean

The shaded range is the 95% sampling range: how far the number could move by chance with this many sales. A single week of 300 to 400 sales carries about 1% of noise, which is why the index is published monthly.

The bands do not cover modeling choices, revisions to provisional months, or the slow drift from chaining weekly estimates together (about 1.5% over the full history against a single fit of the whole period). Treat small moves as noise.

What features are worth

The model also estimates what each feature adds to price with everything else held equal: a pool, a garage bay, waterfront, more living area. Those estimates are updated nightly in a table on the home page.

Two cautions when reading it:

• The bedroom paradox. An extra bedroom can show as worth little, or even less than nothing. That is the value of one more bedroom at the same square footage, which means smaller rooms. Buyers pay for space; the value shows up under living area.
• Features travel together. Waterfront homes tend to be larger, better built and more likely to have a pool, so the model cannot separate each feature perfectly. The values are approximate market-wide averages, not an appraisal adjustment grid.

Known limits

• County files describe each house as it is today. An older sale may be credited with a pool or addition built later, which biases the index upward. Dated nightly snapshots remove this for new sales.
• The index prices a house of constant age. A house you own also gets older.

How to check our work

The inputs are public records. The index has been tested five ways:

• a week-by-week replay using only the sales available at the time;
• a repeat-sales index from the same sales (28,717 pairs), using no house characteristics;
• both methods on the same houses;
• comparison with Zillow and FHFA;
• leaving out waterfront sales in the six months after Ian, Helene and Milton, which moved no year by more than 0.07%. Inland flooding was not tested.

Figures on this page are from the backtest of November 12, 2026.

What it is not

• Not an appraisal or valuation of any property. Individual homes vary widely.
• Not a forecast. It measures what has happened.

Sources

Sarasota County Property Appraiser and Manatee County Property Appraiser public records. The index is computed by Nick Schlabach from those records and is not produced or endorsed by either office.