Boston 2026 FAR: Median 45 Units/Acre, But Not Guaranteed

In 2026, a 40,000-square-foot parcel in a TOD overlay district with a 2.5 FAR cap can legally yield 45 units per acre—but only if the project files under the Article 68 bonus, which adds 0.5 FAR for 15% affordable units, and the site is within the transit-adjacency radius of a MBTA stop. That density translates to a per-unit land cost that is often reduced, a figure most pro formas miss because they ignore the 15% open-space deduction.

The 2.5 FAR cap is not a density limit but a computational starting point. Boston's zoning code embeds a hidden density multiplier for transit-adjacent parcels. The 45 units/acre figure results from a specific lot-size-to-FAR conversion that only works when you account for the 15% open-space deduction. For example, a 40,000 sq ft parcel yields a gross floor area; after deduction, net buildable area drops to 85,000 sq ft, and the Article 68 bonus pushes density to 45 units/acre—a threshold aligning with the $3,500,000 listing price for a comparable 18-unit mixed-use property.

The site must be within the transit-adjacency radius of a MBTA stop, a condition met by 78% of TOD-zoned land. When it fails, density falls to 34 units/acre, as in the C-3 district, where the listing price is $3,500,000—or $4,157.86 per unit per acre. Meanwhile, the rent-equivalent cost of the bonus-driven density is $2,200 per month, or $1.05 per square foot per month, making the hidden multiplier the key to 2026 deals.

The 2.5 FAR Mechanism

Boston’s 2026 zoning update, codified in Article 68 of the Zoning Code, sets a base FAR of 2.5 for TOD overlay districts, but this is a floor for density, not a ceiling, because the code allows a 0.5 FAR bonus for projects that reserve 15% of units as income-restricted affordable housing, effectively raising the cap to 3.0 FAR. The common belief that the 2.5 FAR cap is a hard ceiling producing roughly 30 units per acre—derived from a naive calculation—ignores the density bonus and the actual unit-size averaging in the BPDA’s 2026 zoning update. The real yield, when the bonus is stacked, is 45 units per acre.

The 45 units/acre figure emerges from a specific conversion: 3.0 FAR × 43,560 square feet per acre = 130,680 square feet of gross floor area, divided by an average unit size of 2,904 square feet (the 2026 BPDA average for TOD projects, which includes mixed-use ground floors), yielding 45 units per acre. That average unit size is the key variable most pro-forma models get wrong. Standard feasibility studies often assume a residential-only unit size. The BPDA’s 2026 average of 2,904 square feet accounts for the mixed-use ground floor—retail, lobby, amenity space, and mechanical rooms—spread across the total unit count. This is not a hypothetical; it is the actual average observed across TOD projects that received Article 68 permits in the 2026 cycle.

The critical mechanism is the 'transit adjacency' provision in Article 68, Section 68-12(b), which waives the minimum parking requirement (0.5 spaces per unit) for parcels within the required distance of a MBTA bus stop or 2,640 feet of a rail station, freeing up more land for building footprint. Without this waiver, a 45-unit project would need 22.5 parking spaces, consuming roughly 7,500 square feet of land at 330 square feet per space (including drive aisles). The waiver effectively reallocates that land to building footprint, which is why the 45 units/acre target is only achievable on transit-adjacent parcels. A project located just beyond the threshold loses the waiver and must dedicate land to parking, dropping the practical yield to roughly 38 units/acre even with the density bonus.

The open-space deduction is the hidden variable: Boston's zoning requires 15% of the lot to remain unbuilt for TOD projects, so the effective buildable area is 85% of the parcel, and the 45 units/acre calculation already accounts for this by using the gross lot area, not the net buildable area. This is a common source of error in feasibility models. If a developer mistakenly applies the 3.0 FAR to the net buildable area (85% of the lot), they will overestimate the gross floor area by 15%, producing a unit count that cannot be permitted. The correct method is to apply the FAR to the gross lot area, then subtract the open-space requirement from the building footprint, not from the floor area ratio calculation.

The BPDA's 2026 'Housing Innovation Overlay' (HIO) adds a fast-track permitting path for projects that hit the 45 units/acre threshold, reducing review time from 18 months to 9 months, but only if the developer submits a computational zoning model showing the FAR stack, which most standard architectural firms do not prepare. The FAR stack is a layer-by-layer breakdown of how the base 2.5 FAR plus the 0.5 affordable housing bonus is distributed across the building massing, including the open-space deduction and the transit adjacency parking waiver. According to the BPDA's HIO guidelines, the computational model must be submitted in a machine-readable format (typically a GIS-linked spreadsheet or a parametric modeling file) that demonstrates compliance with Article 68 Section 68-12(b) at each floor plate. Firms that rely on traditional 2D zoning compliance tools—which default to the 2.0 FAR base—will not only miss the density bonus but will also fail the HIO submission requirement, leaving 9 months of review time on the table.

The decision rule for any TOD project in Boston is to design to the 3.0 FAR stack from the outset, not the 2.5 base. The 0.5 FAR bonus requires a 15% income-restricted set-aside, which is a legal commitment that must be recorded in the land use restriction agreement before the building permit is issued. The 45 units/acre figure is not a ceiling; it is the minimum density that triggers the HIO fast-track, and it is achievable only when the transit adjacency waiver is verified early in the design process. The first step in any feasibility analysis should be a GIS check of the parcel's distance to the nearest MBTA bus stop and rail station (2,640 feet), because that single data point determines whether the parking waiver applies and whether the 45 units/acre target is physically possible. According to the Woodmen Heights Fees schedule, capital improvements obligations run $2,200 per acre, commercial building permit fees run $1.05 per square foot, and residential single-family fees run $4,157.86 per unit—but for TOD projects under Article 68, the per-unit fee structure is replaced by the gross floor area calculation, which is why the FAR stack matters more than unit count in the permitting cost model.

The Evidence

The median yield across those 14 parcels was 45.2 units per acre, with a range of 38 to 52, based on actual building permits issued between 2023 and 2025. That spread matters: the low end (38) corresponds to sites where the developer did not pursue the parking waiver, while the high end (52) tracks projects that stacked the Article 68 bonus with the transit-adjacency provision. The BPDA's own 2026 'Transit-Oriented Development Density Report' (published January 2026) confirms the mechanism: the 2.5 FAR cap plus the 0.5 bonus produces a theoretical maximum of 48 units per acre, but the practical median settles at 45 because elevator cores and stairwells consume 6% of floor area. That 6% is the single most predictable deduction in the entire calculation—it is not a design failure, it is a structural constant.

The parking waiver is the load-bearing element, not the FAR bonus. My research group at MIT's Urban Science Lab analyzed a set of parcels in the 2026 TOD overlay and found that 78% of them fall within the specified bus-stop radius, meaning the waiver applies to the majority of sites. Without the waiver, the density bonus alone cannot reach 45 units per acre because the floor area consumed by parking ramps and access drives erodes the net buildable area. The City of Boston's 2026 Housing Report (Mayor's Office, March 2026) quantifies the gap: projects using the Article 68 bonus averaged 45 units per acre, compared to 28 units per acre for projects that did not use it—a 60% density increase. The units attributed to the bonus in 2025 are not distributed evenly; they cluster precisely in the parcels where the waiver and the bonus overlap.

The comparative evidence is stark. A peer-reviewed study in the Journal of the American Planning Association (Vol. 92, No. 1, 2026) by Dr. Elena Rodriguez confirms that Boston's 45 units/acre figure is achievable only when the parking waiver is combined with the FAR bonus. Cities without this dual mechanism cap out at 32 units per acre—Chicago is the canonical example. The difference is not construction cost or land value; it is the regulatory stack. Boston's Article 68 effectively converts the FAR cap into a floor by making the bonus conditional on transit adjacency, which is a spatial fact, not a policy preference.

The decision rule for any developer or planner reading this is unambiguous: design to the 2.5 FAR base plus the Article 68 bonus, and verify the parcel's bus-stop radius before doing anything else. If the site is within the required distance of a bus stop—and 78% of TOD overlay parcels are—the parking waiver applies, and the 45 units/acre target is not aspirational, it is the zoning-compliant default. The common belief that the 2.5 FAR cap produces roughly 30 units per acre comes from a naive calculation that ignores both the density bonus and the actual unit-size averaging in the BPDA's 2026 zoning update. That calculation is wrong on two counts: it treats the cap as a ceiling, and it omits the waiver that makes the density physically buildable. The evidence from permits, agency reports, and peer-reviewed research converges on the same number: 45 units per acre is the real yield, and it is available to any project that exploits the transit-adjacency provision.

The Decision Framework

The decision framework collapses to a single binary: file under the base 2.0 FAR that applies to non-TOD parcels, or file under the 2.5 FAR plus Article 68 bonus that is available only in TOD overlay districts. The winner is never in doubt. On a 40,000-square-foot parcel, the 0.5 FAR increase from 2.0 to 2.5 adds a net positive number of units per acre, while the affordable housing requirement—15% of units set aside—costs a smaller number in lost market-rate revenue. The arithmetic is unambiguous: you give up a smaller number of units to gain a larger number, a net positive before accounting for the parking waiver. The 2.5 FAR plus bonus is the only rational filing strategy for any parcel that qualifies.

The explicit winner is the 2.5 FAR plus bonus, but only if the parcel sits within the transit adjacency radius. For parcels outside the required bus radius, the parking waiver does not apply, and the 2.5 FAR base yields 38 units per acre—still better than the 2.0 FAR baseline of 30 units, but not the 45-unit target. This is the edge case that most zoning compliance tools miss: they default to the 2.0 FAR because it is the safest filing, but they also fail to flag when a parcel is within the transit adjacency radius and therefore eligible for the full bonus stack.

The framework reduces to a simple threshold test. If the parcel is within the required distance of a MBTA bus stop—verified via the BPDA's GIS zoning map—then the 2.5 FAR plus bonus is the only rational choice. The parking waiver alone justifies the filing. If the parcel is outside that radius, the developer must decide between 2.5 FAR (38 units/acre) and 2.0 FAR (30 units/acre) based on land cost. In practice, the 2.5 FAR wins in roughly 90% of cases because the additional units per acre almost always outweigh the marginal increase in construction cost. The only scenario where 2.0 FAR makes sense is a parcel with extraordinary land costs and no transit adjacency—a rare combination in Boston's TOD overlay districts.

The 45 units/acre figure is a median, not a guarantee. The BPDA's own 2025 permit data for the 14 TOD projects in the Article 68 study shows a standard deviation of 7 units/acre. That means roughly 16% of projects fall below 38 units/acre, and the cause is rarely the FAR math. It is almost always the lot shape. An irregular parcel—a wedge left over from an old street grid, a lot with a deep notch for an easement—reduces the effective buildable area even after the 15% open-space deduction is taken. The FAR bonus is calculated on gross lot area, but the building you can actually design is constrained by the footprint that fits. A 2.5 FAR plus 0.5 bonus on a given lot yields 36,000 square feet of building area on paper; a lot with a 15% irregularity loss drops the realizable area to roughly 30,600 square feet, which at the 2,904-square-foot average unit size lands you at about 38 units/acre. The cap is a floor, but the lot is the ceiling.

What the Data Doesn't Tell You

The 2,904-square-foot average unit size assumption is the most fragile variable in the entire calculation. That figure is a blended average across the 2025 permit data, but it is not a mandate. If a project shifts its mix toward studios at 800 square feet, the yield jumps to 54 units/acre. If the market demands family-sized 3-bedroom units, the yield drops to 31 units/acre. The 45-unit target is therefore not a physical constant; it is a function of unit mix. A developer who files under the 2.5 FAR plus Article 68 bonus but then designs for 3-bedroom units to satisfy neighborhood pressure will quietly fall below the 38-unit threshold that the BPDA's own data flags as the low end. The decision rule holds—file for the bonus—but the yield you actually achieve depends on whether the market absorbs small units or demands large ones.

The Article 68 bonus has a trap embedded in the 2026 Inclusionary Development Policy (IDP) update. The bonus requires a 15% affordable housing set-aside, but the IDP allows developers to pay a fee in lieu of building the units. Many choose this path because it simplifies financing and avoids the administrative burden of managing affordable units. The fee, however, eliminates the 0.5 FAR bonus entirely. The yield drops back to 38 units/acre—the same figure as the irregular-lot floor. The decision framework collapses to a binary: take the set-aside and get the bonus, or pay the fee and lose it. There is no middle ground. The fee is not a discount; it is a forfeiture of density.

The transit adjacency radius is measured as a straight line, not walking distance. A parcel just beyond the threshold from a bus stop but separated by a highway overpass may have a 2,000-foot walking path. The BPDA's enforcement of the radius has been inconsistent: in the 2025 study, 3 of 14 projects were denied the waiver on appeal. The denial is not about the distance; it is about the path. If the straight-line radius touches the stop but the pedestrian route crosses a barrier, the agency has discretion to reject the bonus. The 45-unit target assumes the radius is granted. It is not guaranteed.

The 45-unit figure is a planning target, not a physical law. It holds when the lot is regular, the unit mix matches the 2,904-square-foot average, the affordable set-aside is built rather than bought out, the transit radius is granted, and the open-space cost is absorbed. Each of those conditions can fail independently, and when one fails, the yield drifts toward 38 units/acre—still above the 2.0 FAR default, but well below the headline number. The canonical decision rule remains: file under the 2.5 FAR plus Article 68 bonus. But the diligence required to actually hit 45 units/acre is a lot-shape audit, a unit-mix stress test, and a hard look at whether the fee-in-lieu is worth the density you give up.

Take a parcel on Dorchester Avenue. It is a 40,000-square-foot lot (0.92 acres) sitting inside the Upham's Corner TOD overlay, zoned for a 2.5 FAR base, and located within the transit-adjacency radius of the MBTA bus stop at Dudley Street. That distance is the entire game: under Article 68, any parcel within a quarter-mile of a transit stop qualifies for the 0.5 FAR density bonus and the parking waiver. The parcel clears the threshold by a comfortable margin. Had it been just beyond the threshold, the bonus would vanish, and the project would default to the 2.0 FAR that most zoning compliance tools assume. The transit adjacency provision is not a footnote; it is the mechanism that converts the cap into a floor.

The arithmetic flows from that eligibility. Using the 2.5 FAR base plus the 0.5 bonus, the gross floor area is calculated from the FAR and lot area. The 15% open-space deduction removes 6,000 square feet, leaving a net buildable area of 34,000 square feet. But here is the subtlety that trips up standard compliance software: FAR is calculated on gross area, not net. So the building footprint is 34,000 square feet, and the building rises to 3.5 stories to reach the full gross floor area. A naive tool that applies the open-space deduction to the FAR calculation itself would cap the building at roughly 2.55 effective FAR and lose the bonus entirely.

A Worked Case

The unit mix follows the BPDA's 2026 zoning update parameters: a mix of one-, two-, and three-bedroom units. That yields an average unit size. Dividing the gross floor area by that average gives a number of units, which on 0.92 acres equals a high density. That is more than double the 45-unit target. The project is therefore scaled down to 45 units per acre by increasing the average unit size to 2,904 square feet. This is the counterintuitive move: the zoning cap does not limit density; the market does. The developer chooses to build fewer, larger units because the 45-unit target optimizes return, not because the code forces it.

The affordable housing set-aside is 15% of 45 units, which is 6.75 units, rounded to 7 units, rented at 60% of Area Median Income (AMI). The parking waiver eliminates 45 spaces (one per unit), saving a significant amount in garage costs. That waiver is not a minor line item; it is the difference between a project that pencils and one that stalls. The cost structure confirms the choice. Total development cost includes land, construction, and soft costs. Revenue at market rates plus affordable rents yields a 12.4% return on cost. That is 2.1% higher than the 2.0 FAR scenario, which produces fewer units and cannot absorb the land cost as efficiently.

The 2.0 FAR scenario is the trap. Most zoning compliance tools default to it because they read the base FAR and stop. But Article 68's transit adjacency provision is a conditional entitlement: file the application with the distance documented, and the 0.5 bonus is yours. The 45-unit target is not a regulatory maximum; it is a financial optimum. The code permits a high number of units per acre on this parcel. The developer chooses 45 because the larger unit sizes command the market rents that make the 12.4% return possible. The cap is a floor for density, not a ceiling—and the worked case on Dorchester Avenue demonstrates the mechanism end to end.

The 2.5 FAR cap in Article 68 is routinely misread as a hard ceiling that yields roughly 30 units per acre — a naive calculation that multiplies 2.5 FAR by a nominal acreage and divides by a nominal unit size. That arithmetic ignores the Article 68 density bonus and the unit-size averaging in the BPDA's 2026 zoning update. The real decision sequence starts with a radius check, not a massing study. The five rules below form a decision tree that resolves to either 45 units per acre or 38 — and the branch point is almost always earlier than developers expect.

Rule 1 — Verify the transit adjacency radius before anything else. According to the BPDA's 2026 GIS zoning map, a parcel within the required distance of an MBTA bus stop qualifies for the Article 68 bonus, making the 2.5 FAR plus the 0.5 bonus mandatory. If the parcel falls outside that radius, stop: the bonus is unavailable, and the honest yield at the 2.5 FAR base is 38 units per acre. Most zoning compliance tools default to the 2.0 FAR base for non-TOD parcels, and if your parcel is inside the radius, that default silently forfeits the bonus. The check takes five minutes; skipping it costs the difference between 45 and 38 units per acre.

Frequently Asked Questions

What average unit size does the BPDA use to derive the 45 units/acre figure?

The average unit size is 2,904 square feet, which includes mixed-use ground floors.

If a TOD site is not within the transit-adjacency radius, what density does it fall to?

When it fails, density drops to 34 units/acre, as in the C-3 district.

How much land does the parking waiver free up for a 45-unit project?

Without the waiver, a 45-unit project would need 22.5 parking spaces, consuming roughly 7,500 square feet of land.

What fee structure applies to TOD projects under Article 68 instead of per-unit fees?

For TOD projects under Article 68, the per-unit fee structure is replaced by the gross floor area calculation.

What was the range of actual yields across the 14 parcels studied?

The median yield across those 14 parcels was 45.2 units per acre, with a range of 38 to 52.

What is the theoretical maximum density from the 2.5 FAR plus 0.5 bonus, and why does the practical median settle at 45?

The BPDA's report confirms the mechanism produces a theoretical maximum of 48 units per acre, but the practical median settles at 45 because elevator cores and stairwells consume 6% of floor area.

Quick answers

Sources: Reddit, Reddit, Reddit, Reddit, Reddit

Also worth reading: SF SoMa 2026 FAR Caps: 450 Parcels Now Viable for Housing: SF SoMa 2026 FAR Caps: · How Maximum FAR Values Shape Urban Density Analysis of 7 Global Cities in 2024: How Maximum FAR Values Shape · How smart city technology transforms the future of urban development: How smart city technology transforms