Market Intelligence · September 2026
The Palm Beach market, measured carefully.
A median can describe direction. It cannot explain a landmark estate, an ocean-to-lake parcel, a teardown acquired for land, or a transaction that never appeared in the MLS.
Areas 5001–5003 · Data through September 20, 2026The current evidence
Three figures that frame the market
Palm Beach single-family and attached homes sold for a median of $3,167 per living square foot in 2026 year-to-date, 159% above 2019 and 1.2% below the 2022 median.
Since 2024, 48% of concluded listing attempts in these three areas ended without a sale: 195 expired, canceled or withdrawn attempts against 211 closed attempts.
The trailing-year median sold prices were The North End $10.43M; Midtown & In-Town $11.85M; The Estate Section $25.15M. These are different housing markets within one town.
Historical direction
The repricing held.
The 2023–2024 correction was real. The floor held at more than twice the pre-pandemic level, and 2025–2026 returned to near the 2022 peak.
| Year | Closed sales | Median sold price | Median sold $/living SF |
|---|---|---|---|
| 2019 | 79 | $4,900,000 (n=79) | $1,225 (n=79) |
| 2020 | 180 | $5,325,000 (n=180) | $1,340 (n=180) |
| 2021 | 143 | $7,630,000 (n=143) | $1,928 (n=143) |
| 2022 | 62 | $12,706,288 (n=62) | $3,204 (n=62) |
| 2023 | 42 | $8,855,000 (n=42) | $2,764 (n=42) |
| 2024 | 51 | $10,750,000 (n=51) | $2,634 (n=51) |
| 2025 | 76 | $12,062,500 (n=76) | $2,953 (n=75) |
| 2026 YTD | 82 | $12,250,000 (n=82) | $3,167 (n=82) |
Partial-year figures are marked. Annual medians reflect modest samples and the mix of what sold.
The island in three sections
One geography. Three housing markets.
| Area | Sales | Median sold price | Median $/living SF | Active listings | Months of supply |
|---|---|---|---|---|---|
| The North End | 44 | $10,425,000 (n=44) | $2,848 (n=44) | 25 | 6.8 |
| Midtown & In-Town | 38 | $11,850,000 (n=38) | $3,251 (n=38) | 7 | 2.2 |
| The Estate Section | 19 | $25,150,000 (n=19) | $3,937 (n=19) | 13 | 8.2 |
The skeptic’s ledger
This is not a market in which anything sells at any price.
Since January 1, 2024, these three Palm Beach areas recorded 211 closed listing attempts and 195 expired, canceled or withdrawn attempts: 48% of concluded attempts ended without a sale. The closed attempts represent 209 distinct closings after duplicate transaction records are removed. 90 of 209 eligible closings (43%) had a final asking price below the original asking price. Median time from market entry to contract was 104 (n=205) days.
The listing-attempt percentage describes outcomes in the record; it is not a property’s probability of selling.
Each price band is its own market.
The mix of land, frontage, size and condition changes across tiers. These are observed medians, not an isolated premium for any one feature.
| Sold-price tier | Closed sales | Median sold $/living SF |
|---|---|---|
| Under $5M | 17 | $1,999 (n=17) |
| $5M–under $10M | 110 | $2,488 (n=110) |
| $10M–under $15M | 79 | $3,062 (n=79) |
| $15M–under $20M | 41 | $3,461 (n=40) |
| $20M–under $25M | 17 | $3,742 (n=17) |
| $25M and above | 49 | $4,930 (n=49) |
Housing mix
| Property type | Closed sales | Median sold price | Median sold $/living SF |
|---|---|---|---|
| Single-family detached | 89 | $12,689,525 (n=89) | $3,201 (n=89) |
| Townhouses and villas | 11 | $9,000,000 (n=11) | $2,514 (n=11) |
| Conflicting MLS property-type records | 1 | Limited sample (n=1) | Limited sample (n=1) |
Property-specific context
Use the market to frame the property—not replace it.
A private discussion applies the correct area, tier, condition, and transaction evidence to the actual brief.
Periods, formulas and data limits
Single-family homes, townhouses and villas · North End, Midtown & In-Town and Estate Section · Condos and co-ops excluded. Source groups correspond to MLS areas 5001–5003. Records must identify Palm Beach and ZIP 33480. These are the named development groups, not all transactions in ZIP 33480.
Medians require at least 5 eligible observations; percentage comparisons require at least 10 in each comparison group. Smaller samples show “Limited sample” and the count. n is the eligible count for the particular metric. Sold $/living SF is each closing price divided by its recorded living area, then the median of those ratios. Counts remain visible when medians are suppressed.
Closed sales are deduplicated by parcel number, closing date and closing price; listing key is used when parcel number is absent. When duplicate closing records disagree on detached versus attached housing type, the sale remains in the overall total and appears in a separate conflicting-type row. The ledger counts listing attempts separately: a property can expire and later sell. It is not a count of unique owners or unsuccessful properties. Failed-attempt dates use the recorded status-change date. Price reductions compare final asking price with original asking price; they do not reconstruct every price change.
Market time runs from OnMarketDate (or ListingContractDate if absent) to PurchaseContractDate; missing, negative or over-ten-year intervals are excluded. Months of supply = active listings ÷ (trailing-year closings ÷ 12), with at least 5 closings. Only Active status counts as available inventory; pending and active-under-contract listings are excluded.
Annual rows before 2026 are full calendar years; 2026 is partial through 2026-09-20. Changes in medians reflect sales mix, not repeat-sales appreciation. Waterfront groups follow explicit MLS flags and features; ocean access or beach access alone does not establish frontage. A record coded for both oceanfront and lakefront does not verify an ocean-to-lake parcel. Tax parcel counts are deduplicated by property-control number and include vacant parcels; they are not a count of homes for sale.