Draft for review. This briefing is a plain-language rendering of a completed audit and has not yet had its second read. Do not rely on it until this notice is gone.

Where today's law sits

The old version: off the chart — Polling places moved openly to dilute the vote of named groups — Jim Crow precinct closures, with Massachusetts's 1894 literacy-test registrar discretion as the procedural cousin.

  1. Who decides? (L1) One vague test — public convenience or public health — with no criteria and no appeal.
  2. Measuring by map (L2) Polling places may go to non-adjacent precincts with no distance, transit or accessibility floor — worse than today's law.
  3. Where the line is drawn (L3) 20-day change window; the report may post as late as 3 days out — after mail-in ballots may have gone back.
  4. Who's left out entirely (L4) The report must cover race, national origin, disability, income and age — and omits language and sex.
  5. Automatic penalty, judgment call — or nothing? (L6) Nothing. No penalty, no election contest, no override, no officer liability anywhere in the bill.
  6. Half a protection (L7) A duty to evaluate and report, with nothing riding on what the report finds — a paper protection.
  7. Who does the paperwork? (L8) Electronic notice 'to the extent available'; no mailed notice, no bilingual notice, no sign at the old site.
One row for each drafting decision in this law. The dot shows where today's choice sits: toward the left is the rigged version — the worst law that would still look fair on paper — and toward the right is the most protective version. A solid dot is a choice the text makes; an open circle is a decision the text does not make at all. The old, openly discriminatory version sits off the chart; it is noted once for history and never used as the comparison.

The bill as drafted sits hard toward the rigged end on the four drafting decisions that decide who can actually get to the polls, its impact report is a paper protection that earns no credit toward the protective end, and it grants far more authority than the city asked for — narrowing it to the stated purpose is the largest single fix available.

Every one of these decisions was made by a person — some deliberately, some by copying an old form, some by not thinking it through. This audit does not sort out which. It shows what each decision does, because the ramifications are the same whether the choice was intentional, inherited, or sloppy. What changes after the audit is this: once you can see where a decision sits and what it costs, keeping it becomes a decision too.

What we found

Where the bill sits, and why. The bill as drafted lets the city council move any polling place, even to a precinct that does not touch the voter’s own, on 20 days’ notice, under a vague standard, with no consequence for skipping any duty in it. Those four drafting decisions — who decides (L1), measuring by map (L2), the consequence (L6), and half a protection (L7) — decide who can actually get to the polls, and all four sit close to the rigged version. The impact report the bill requires is a paper protection (Sub-rule 7a) and earns no credit toward the protective end.

The bill asks for far more than the city needs. The Board of Election Commissioners’ letter (Lyons, 26 May 2026) documents one intended use: consolidating Precincts 1A, 1B, 1C, 2A and 2B at the Liberal Club, 20 Star Street. The only legal obstacle is c. 54, § 24’s bar on polling places in buildings that serve alcohol. The bill instead grants standing authority to move any polling place, to non-adjacent precincts, on a 20-day clock. The structural problem lives in that excess scope. A one-section bill naming the Liberal Club for those five precincts — overriding c. 54, § 24 and c. 138 only on the alcohol point — removes most of the problem and gives the city exactly what it asked for.

The trouble spots. Who decides (L1): one vague test — “if it is determined that the public convenience or public health would be better served” — with no criteria and no appeal. Measuring by map (L2): “The city council may designate polling places in non-adjacent precincts,” with no distance, transit or accessibility floor; a step backward from today’s law. The consequence (L6): no penalty, election contest, override, or officer liability anywhere in the bill. Half a protection (L7): the council “shall evaluate and report on whether such change would have a disparate, adverse impact,” but nothing turns on what the report finds. And who’s left out (L4) sits inside that report duty: it covers “race, national origin, disability, income or age” — omitting language and sex. Language is the second-strongest finding in this audit. Two smaller spots: the report may post as late as three days before the change, after mail-in ballots may already be back (L3); and notice “by using an electronic means, to the extent available, such as via email or reverse 911 call” is undefined, reaches only addresses the city holds or landlines, and requires no mailed notice, bilingual notice, or sign at the old site (L8).

Who pays for it. The screen, run on public data down to 2020 Census blocks, shows the consolidation slightly shortens the average trip citywide (about 0.56 km) — but the average hides who gets worse. Precincts 1B and 2A improve; 1C and 2B get farther; 1A is roughly unchanged. Measured the standard way for disparate impact (the EEOC four-fifths test, adapted to distance), three groups are over-represented among voters whose trip gets longer: households without a car, 1.62 times; Portuguese speakers with limited English, 1.41 times; Black residents, 1.24 times. The first two sharpened at every pass of finer data; the proof is strong. The third appeared only at the finest resolution, on small numbers, and needs block-group confirmation before it carries weight. Two harms the screen cannot see: the polling place at the Edward F. Doolan Apartments serves that building’s senior residents, who would be sent 1.75 miles away — invisible when averaged across a 3,649-person precinct; and English-only ballots where 12 to 26 percent of residents have limited English.

The paper protection can be turned against the voters it names. A council finding of “no adverse impact” under this bill would be contradicted by public data on three counts. Once published, that finding becomes the record a court defers to — the city’s own defense exhibit against a later challenge under Voting Rights Act § 2, ADA Title II, Title VI, or state disparate-impact law. The provision is not just weak; it sets up a finding the council’s own bill cannot defend. A predicted risk when the audit was first written; the screen has since supplied the facts.

What’s missing, in order: a rule that no change takes effect if the assessment finds a disparate adverse impact unless mitigation is in place; a city-funded election-day shuttle from stops in Precincts 1C and 2B; bilingual Portuguese and Spanish notice and election-day help at the new site; a polling place kept at the Doolan Apartments for its residents; block-group confirmation of the Black-resident flag; and verification of Bristol County’s status under §§ 203 and 4(e) of the Voting Rights Act.

Fact-check. The screen is reproducible from open sources — MassGIS 2022 precinct boundaries, ACS 2024 five-year tables, 2020 Census blocks. The one genuinely protective step in the bill is adding “income” to the report’s list. Two questions are rated most urgent: the county’s § 203 status, a federal-compliance question independent of the consolidation; and whether Massachusetts deference doctrine would in fact turn the report into a shield — the factual half of that risk is demonstrated, the legal half still a prediction. The Doolan finding rests on strong inference; one call to the Election Commission would settle it. None of the protective options ties a benefit or burden to anyone’s race or protected status; they key to buildings, distance, transit, vehicle access and language spoken at home. No equal-protection flag.

Options for the drafter

Options to weigh, not instructions:

  1. Make the report bind — the central fix. Strongest: no change takes effect if the assessment finds a disparate adverse impact unless named mitigation is in place. Middle: an election contest under c. 56, a civil penalty, and voidability. Weakest but better than silence: require the report to list mitigation already adopted.
  2. Put a floor under the map (L2). Carless voters are concentrated in Precincts 1C and 2B. If “non-adjacent” survives, pair it with an accessibility and transit floor.
  3. Add language to the list (L4). Not optional in this city: bilingual Portuguese and Spanish notice and election-day assistance.
  4. Close “to the extent available” (L8). Mailed individual notice; a sign at the old site on election day; a city-funded shuttle.
  5. A provisional-ballot fallback (L6). The cheapest single consequence the bill could carry.
  6. Keep the Doolan Apartments as a polling place for its residents; consolidate everyone else from 1B at the Liberal Club.
  7. Check the county’s § 203 status. Not a drafting change — a compliance flag the screen raised. Advise the requestor and the delegation.
  8. Weigh the exposure the report creates. As drafted, the report-only duty may expose the city to a challenge it would not face if the duty were either removed or made binding.
  9. The largest single fix is the narrow bill. Drop “any polling place,” drop “non-adjacent precincts,” drop the general 20-day authority. Authorize the Liberal Club for the named precincts. The L1, L2, L6 and L7 findings shrink dramatically once scope matches purpose.

How to spot this pattern in a live bill

  1. Who decides: “may,” a multi-factor test (convenience, health, safety, fiscal), no listed criteria, no appeal.
  2. Measuring by map: the geographic limit removed outright — “any precinct or any location.”
  3. Where the line is drawn: the notice clock squeezed as short as it will go.
  4. Who’s left out: no list of protected groups, or one that omits the group most at stake locally — language in Fall River; language and disability in the transit-served cities of western Massachusetts.
  5. The consequence: an explicit immunity or no-review clause.
  6. Half a protection: “may consider… may publish” — every part of the equity duty is permissive. (This bill makes the report mandatory, but a mandatory report with nothing riding on it fails the same test.)
  7. The paperwork: website notice “deemed sufficient”; no individual notice.

Three or more of these together: treat the bill as the rigged version, whatever its stated purpose.

Sources

  • Draft Home Rule Petition, City of Fall River, received from the Office of the Secretary of the Commonwealth (authorizing polling-place changes notwithstanding G.L. c. 54, § 24)
  • Letter of Ryan Lyons, Chairman & Director, Fall River Board of Election Commissioners, to the Fall River legislative delegation, 26 May 2026
  • G.L. c. 54, § 24
  • G.L. c. 56 (election contests)
  • G.L. c. 138
  • Voting Rights Act, §§ 2, 4(e) and 203
  • Americans with Disabilities Act, Title II
  • Civil Rights Act of 1964, Title VI
  • U.S. Census Bureau, 2026 quinquennial § 203 determinations (Bristol County status pending verification)
  • American Community Survey 2024 five-year tables C16001, B02001, B03003, B19013, B25044, B01001 (via Census Reporter)
  • 2020 Decennial Census block-level population (TIGERweb 2020 MapServer, layer 10)
  • MassGIS Wards and Precincts 2022 shapefile
  • OpenStreetMap Nominatim (geocoding)
  • EEOC four-fifths adverse-impact test, adapted to distance-based access screening
  • SFFA — equal-protection stress test (no flag raised)
About this page. This is Part I of the audit — the briefing a legislator reads. Part II, the evidence appendix, holds the chapter scan, the decision inventory with quoted statutory text, the same law written four ways, the ratings, and the validation record. It is available to clients on request; two of its four versions are deliberately harmful teaching drafts and travel under a warning label. An analytic lens, not a determination of unlawfulness.
All five audits Request an audit