Massachusetts
Massachusetts abolished nine of its fourteen county governments and kept electing the county officers anyway. Underneath that, 351 cities and towns cover every square foot of the state with no unincorporated land at all — so the town is not the bottom of the ladder here, it is the whole ladder.
What does county mean in Massachusetts?
State deltas
Every state has hair. Here is Massachusetts's.
The rules that make this state different from the last one, what each breaks if you assume the national default, and how the engine resolves it.
Five of fourteen counties retain a county government. The other nine were dissolved into the Commonwealth — and yet a Clerk of Courts was elected in thirteen counties in 2024, seven of which have no county government to belong to. The office outlived the government. Any model that hangs elected offices off a parent jurisdiction has nowhere to attach seven of those thirteen, and any model that infers the existence of a county government from the presence of a county-wide elected office will invent nine that do not exist.
An office's electoral geography is stored separately from its governing body, so a seat can be elected county-wide while belonging to the state, and a county can exist as geography with no government attached.
351 cities and towns tile the entire Commonwealth with no gaps. There is no territory that belongs only to a county, and therefore no residual role for a county to play. Pipelines built on the national default — municipalities as islands inside a county that administers the space between them — have the containment backwards here: the town is the general-purpose government and the county is the remainder, not the reverse. Address-to-jurisdiction resolution that falls back to a county when no municipality matches will never fire, and if it does the match was wrong.
Municipal coverage is a property of the state rather than assumed, and jurisdiction resolution is driven by the tier that actually blankets the territory.
The 2024 cycle ran a Democratic primary, a Republican primary and a Libertarian primary — 319 contests each, on the same day, as three separate ballots. Most of this corpus has two, and a schema with a boolean for which major-party ballot a voter took cannot represent a third. The Libertarian primary is not a formality either: it appears against the same offices at the same stage, so a contest joined on office and year without also joining on party stage will collapse three distinct races into one.
The nominating stage is stored as the party that ran it rather than as a two-valued flag, and contests join on the stage as well as the office.
Every return carries a Blanks column, so ballots cast and votes cast are separate published numbers — 3,512,930 against 3,473,668 for president in 2024, a gap of 39,262. Elsewhere in this corpus the gap has to be reconstructed, if it can be recovered at all. A consumer that sums candidate columns to get turnout is short by up to 4.29% depending on the town, and one that treats Blanks as a candidate row will invent a contender who placed third in some municipalities.
Ballots cast and votes cast are separate fields on a contest, and non-votes are stored as a disposition rather than as a candidate.
Massachusetts labels a candidate or voter with no party enrollment as Unenrolled, and it appears directly in the published returns as a party value — four of the eight presidential tickets in 2024 carried it. It is not a party, and it is not the same token as Oregon's Non-Affiliated, Alaska's Undeclared or Alaska's Nonpartisan, each of which is a distinct concept in its own state. Normalising all of them to one value discards the state's own vocabulary; treating Unenrolled as a party name creates an organisation that has never existed.
Affiliation is stored as the state's own token with a flag for whether it denotes a party at all, and cross-state comparison happens on the flag rather than on the label.
The Register of Deeds is a county-scale elected office whose geography is the registry district, and the two do not coincide. Berkshire County alone contains three — Northern, Middle and Southern — each electing its own register on its own staggered clock, so in 2024 only four registry contests appeared statewide. Joining these seats to counties one-to-one splits or merges the wrong way in both directions, and a county-keyed office table cannot hold three registers for one county.
Each office carries its own district geography as a first-class key, and no office is assumed to share the boundary of the tier it is named after.
The Governor's Council is elected from eight districts and confirms judicial appointments, pardons and warrants. It is a statewide elected body that is neither the legislature nor a constitutional officer, and almost nothing else in this corpus is shaped like it. A tier model with federal, state, legislature, judiciary and local has no slot for an elected council whose function is to confirm the judiciary, and it is not the judiciary itself.
The office inventory carries bodies rather than assuming a fixed set of tiers, so an elected council can exist at the state level without being the legislature or the courts.
Grouping the 2022 Census government-units list on its own STATE column puts four New England towns in the wrong state — Connecticut's Wolcott lands in Alabama, Maine's Lincoln Plantation and Vermont's Bloomfield land in New Hampshire, and Vermont's Sudbury lands in Massachusetts carrying a Massachusetts mailing address copied from the other Sudbury. Massachusetts appears to have 352 subcounty units and actually has 351. All four rows carry the correct FIPS_STATE, so the defect is in the text field, not the identifier.
Census units are grouped on FIPS_STATE rather than the STATE text field, and any count derived from the unit list is reconciled against the aggregate table before use.
Federal to special district
What Massachusetts actually elects.
Nine House seats, a legislature of two hundred, an elected council that confirms the judiciary — and beneath it 351 towns and no unincorporated ground.
Congressional map
Where those seats are.
The 9 congressional districts in force for the 2026 cycle — not necessarily the map this delegation was elected under. Massachusetts has not redrawn since the 119th Congress was seated.
| District | Population | Median income |
|---|---|---|
| MA-01 | 767,012 | $72,540 |
| MA-02 | 798,884 | $95,495 |
| MA-03 | 767,022 | $95,825 |
| MA-04 | 783,039 | $109,055 |
| MA-05 | 776,923 | $122,808 |
| MA-06 | 785,015 | $112,489 |
| MA-07 | 744,098 | $88,518 |
| MA-08 | 791,775 | $114,053 |
| MA-09 | 787,631 | $95,667 |
Municipality ledger
Three hundred fifty-one cities and towns, and a blank-ballot column the rest of the country does not publish.
Figures below are the 2024 presidential return published by city and town — the state's own download, one row per municipality. All 351 rows sum to 3,512,930 votes cast and 39,262 blanks, 1.12% statewide. The second column is the share of ballots cast that recorded no vote for president, which ranges from zero in Monroe to 4.29% in Gosnold. Thirteen of 351 shown.
| Municipality | Feed vehicle | Votes cast · 2024 president | Blank share | Status | Last read |
|---|---|---|---|---|---|
| Boston · largest | County HTML | 265,413 | 0.99% | Verified | 1 hr |
| Worcester | County HTML | 66,186 | 0.99% | Verified | 1 hr |
| Springfield | County HTML | 51,370 | 0.92% | Verified | 1 hr |
| Cambridge | County HTML | 50,661 | 1.06% | Verified | 1 hr |
| Newton | County HTML | 47,343 | 1.35% | Verified | 1 hr |
| Brookline · a town, not a city | County HTML | 28,586 | 1.19% | Verified | 1 hr |
| Leominster · high blanks | County HTML | 21,051 | 2.61% | Verified | 1 hr |
| Amherst | County HTML | 10,743 | 0.74% | Verified | 1 hr |
| Provincetown · lowest blanks shown | County HTML | 2,671 | 0.34% | Verified | 1 hr |
| New Marlborough | County HTML | 1,002 | 3.39% | Verified | 1 hr |
| Mount Washington | County HTML | 125 | 3.20% | Verified | 1 hr |
| Gosnold · highest blank share | County HTML | 70 | 4.29% | Verified | 1 hr |
| Monroe · smallest · zero blanks | County HTML | 66 | 0.00% | Verified | 1 hr |