The safe harbor that isn't: Braintree and the 40B arithmetic

Braintree, Massachusetts lost a housing case by 26 basis points. The town's zoning board said 1.65 percent. The state's Housing Appeals Committee said 1.39. The statutory line sits at 1.5, and on June 18, 2025, the Massachusetts Appeals Court decided whose arithmetic controls, in a decision that vacated the town's denial and put a recount question in front of every Massachusetts municipality running the same defense.

The disputed input was 244 acres nobody can build on.

Some machinery first, because Chapter 40B only makes sense as a machine. M.G.L. c. 40B §§ 20-23 lets an affordable-housing developer ask the local zoning board for one consolidated permit in place of the usual stack of separate local approvals, and it gives the state's Housing Appeals Committee the power to review a denial de novo and order the permit issued anyway. The counterweight is a set of safe harbors in regulation 760 CMR 56.03: a municipality that already carries enough affordable housing, by one of several statutory measures, is presumed consistent with local needs, and its denial is shielded from HAC reversal. The measure Braintree invoked is the General Land Area Minimum, GLAM, which asks whether low- or moderate-income housing sits on at least 1.5 percent of the town's land zoned for residential, commercial, or industrial use.

Numerator, denominator, threshold. The whole fight lives in those three numbers, and the Braintree record shows how a town can lose on the one it thought was settled.

How did 244 acres flip the math?

When 383 Washington Street, LLC filed for its consolidated 40B permit on two Braintree parcels, the zoning board denied it and claimed GLAM protection: by the town's count, 1.65 percent of its qualifying land carried affordable housing, clearing the 1.5 percent bar. The count depended on one exclusion. Braintree left a 244-acre conservation parcel out of the denominator, reasoning that land inside a residential district that can never legally be developed should not count as residentially zoned land.

HAC put the parcel back in. A bigger denominator shrinks the share, and Braintree's 1.65 became 1.39, below the threshold. No safe harbor, de novo review, denial overturned, permit ordered.

The town appealed and lost twice. The Superior Court held that "zone" in the regulation refers to the municipality's applicable zoning map, not to some buildable sub-region within a district, and the Appeals Court unanimously affirmed on June 18, 2025, in the decision reported at 2025 WL 1699014. Conservation land inside a residential zoning district is residentially zoned land, full stop, whatever its development prospects.

Land you cannot build on still counts as land you zoned.

Twenty-six basis points may be the smallest margin I have seen decide an entitlement fight, and it is worth noticing what did the deciding. Braintree's politics never mattered. Neither did the hearing testimony or the abutters. The case turned on a definitional reading of one undefined word in a state regulation, which is why the file's summary line is the right one: the math, not the politics, is what decides Massachusetts 40B cases.

What does the ruling do to every other Massachusetts town?

It orders a recount, effectively, and it strips the varnish off a standard defensive play. The ZBA posture in these cases is standard: invoke the statistical threshold to shield the denial from HAC's de novo review, and force the applicant to litigate the math. That play works when the math holds. After Braintree, any municipality with meaningful conservation acreage inside residential, commercial, or industrial zoning districts now has to include that land in its GLAM denominator, which pushes the affordable-housing share down, and in some towns will push it below 1.5 percent. A jurisdiction that has been denying 40B applications behind a pre-2025 calculation may discover its shield was never real, and per the record, any denial that leaned on the old arithmetic is now vulnerable. The regulatory drafting gap that invited all this, the word "zone" left undefined by the state housing agency that wrote 760 CMR 56.03, is also the reason the ruling could be superseded: the current housing office could amend the regulation, and the Supreme Judicial Court could someday revisit the question. Until either happens, the Appeals Court's reading controls statewide.

Step back and the structural point comes into focus. HAC's authority to redo a town's safe-harbor math de novo, and to have its calculation control through two levels of court review, is what makes Chapter 40B function as a state-level preemption tool rather than a suggestion. Braintree did not lose to the developer. It lost to the state's arithmetic, applied by a tribunal the statute built for exactly this purpose. Watch the regulator too: the housing office, now the Executive Office of Housing and Livable Communities, wrote the rule whose undefined word produced this litigation, and the cleaner path forward is a rulemaking that settles the definition rather than a re-run of the fight in every town. EOHLC's docket belongs on the same monitoring list as the town's zoning agenda.

Massachusetts is not alone in running housing entitlement on statutory arithmetic. California plays the same game under a different name: a Los Angeles Superior Court judge ruled on August 12, 2025 that Beverly Hills had to process a Builder's Remedy project the city had refused to process, in the Linden Drive case. Different statute, same grammar. When the numbers trip, local discretion goes quiet, and the fight moves from the hearing room to the spreadsheet.

What should a 40B pro forma budget for?

Time, first. Braintree's applicant ran the full chain, zoning board to HAC to Superior Court to Appeals Court, and the record puts the fight at more than three years. The multifamily developer modeling a contested 40B filing as a one-cycle permit process is modeling a different statute. Model it as litigation with a construction project attached: carry reserves for the appeal chain, plus an underwriting case where the permit arrives on HAC-ordered terms rather than negotiated ones.

Then spend the diligence money where this case says it pays. Before filing, recompute the target town's GLAM denominator yourself: pull the zoning map and add back the acreage of every conservation parcel sitting inside residential, commercial, and industrial districts. If the recount drops the town below 1.5 percent, its loudest defense is already dead, and the entire negotiation changes before a single hearing. That is a spreadsheet exercise. Braintree's version of the same answer took two appellate courts.

My bet: within 18 months, a post-383-Washington denominator recount becomes a standard pre-filing exhibit in contested Massachusetts 40B applications, and at least one town that has been denying behind the GLAM shield discovers at its own hearing that it no longer clears 1.5 percent. If the state housing office amends 760 CMR 56.03 to redefine "zone" before that plays out, this note ages fast and I will update it. Either way, run the recount before you budget the fight. The margin you are looking for is 26 basis points wide.

This analysis is a source-cited research summary drawn from public records, not legal advice. It can contain errors and should be verified independently before any investment decision.