The Mainland

Median prices swing with which homes sold in a given week — a few beachfront closings can look like a boom. This index avoids that by comparing each property only to itself: every parcel that sold twice contributes the price change between its own two sales, and a regression (the Bailey repeat-sales method, the basis of the Case-Shiller index) combines those changes into one index. The shaded band shows ±1 standard error — where the true index most likely sits. Standard errors account for the same parcel appearing in multiple pairs, and long-held pairs are down-weighted when the data call for it.

From 45,896 priced deed sales (standard deeds of $25,000+), 7,302 parcels recorded more than one sale date, giving 17,605 sale pairs. Excluded: pairs held under a year (10,262 — re-recordings, corrections, and flips), implausible price ratios (408 — mostly new construction), and pairs whose deed addresses conflict (2,701 — different condo units or streets sharing one block/lot, which would compare two different homes). That leaves 4,234 clean pairs county-wide. Towns need 100+ clean pairs for their own quarterly index; smaller towns appear in the menu but fall back to the county-wide line.

Repeat-sales indexes track price change for resold homes; they say nothing about the level of prices (see town dashboards for medians) and can understate markets dominated by new construction.