Groundwork

Recommendations

Ten ways the Bay Area could add more homes, ranked. We show our reasoning and say who pays the price for each one.

How to read the scores

Every option gets 1–5 on four dimensions: supply impact (how much housing it could unlock), evidence strength (how solid the research behind it is), public cost (5 = near-free to the public purse), and speed (how soon it could matter). Higher is always more attractive.

Ranks are our judgment, not a formula. Every score is contestable, and the sources are right there if you want to argue.

#1

Upzone near transit — and hold the line on implementation

Legalizing mid-rise housing near BART, Caltrain, and Muni (SB 79 plus local plans like SF's Family Zoning Plan) is the region's biggest durable supply lever. The best evidence anywhere in housing policy — Auckland's 2016 upzoning and Minneapolis 2040 — shows broad upzonings producing units at scale and holding rents down.

Supply impact
5/5
Evidence strength
4/5
Public cost
5/5
Speed
2/5
How it works, the tradeoffs, and where it stands

How it works

Zoning caps bind hardest where demand is highest. Raising the legal density ceiling near transit expands the set of parcels where redevelopment pencils; pairing it with ministerial approval (SB 423) removes the discretionary veto that historically neutralized paper capacity.

The honest tradeoffs

Neighborhood change is the honest cost: taller buildings, shadows, parking pressure, and redevelopment near stations, borne by incumbent residents. Research on displacement finds new market-rate construction reduces nearby rents and displacement on net, but localized effects are real, and the law's tenant protections matter. The other risk is quieter: upzonings that are too small or too gameable capitalize into land prices without units (Chicago's TOD upzoning produced zero new permits in five years). SB 79's fate turns on whether HCD holds the line on local exclusion ordinances — and note that Contra Costa's BART corridors are statutorily exempt.

Who bears it

Incumbent homeowners near stations; at the margin, tenants of redeveloped parcels.

Where it stands here

SB 79 effective July 1, 2026 in SF and Alameda counties (not Contra Costa or Marin); SF and Oakland pursuing exclusions and alternative plans of uncertain HCD fate; SF's Family Zoning Plan in effect but in litigation. First completions realistically 2029-2031.

Sources: Greenaway-McGrevy & Phillips, Auckland upzoning (JUE 2023) · Pew on Minneapolis 2040 · Freemark, Chicago TOD caveat (UAR 2020) · Holland & Knight SB 79 implementation tracker · Updated August 21, 2026

#2

Make approvals fast, certain, and ministerial

San Francisco needed 26.6 months to entitle a typical apartment building while Oakland needed 5.4 — same state laws, different local process. Speed and certainty reforms are cheap, measurable, and already delivering: SF's average permit approval fell from 605 to about 280 days in 18 months.

Supply impact
3/5
Evidence strength
4/5
Public cost
5/5
Speed
4/5
How it works, the tradeoffs, and where it stands

How it works

Time is carrying cost and uncertainty is a risk premium; both raise the feasibility bar before a shovel turns. Ministerial approval collapses hearings and appeal risk; process reform (PermitSF-style) attacks the post-entitlement backlog.

The honest tradeoffs

Ministerial means no public hearing and no discretionary design review for compliant projects — that loss of local voice is the explicit price, and it lands on incumbent neighbors. The labor standards attached to California's streamlining laws (prevailing wage on 10+ units) add roughly $80-95k per unit per Terner's LIHTC analysis, offsetting part of the gain. And the SF pipeline shows the ceiling: 7,200+ streamlined units approved, only a few hundred built — approvals cannot fix a financing freeze, they position the region for the next cycle.

Who bears it

Neighbors and design-review constituencies lose veto power; part of the surplus goes to construction workers via attached labor standards.

Where it stands here

SB 423 makes SF a 10% jurisdiction with a streamlined pipeline of 80+ projects; PermitSF cut approval times sharply though its software rollout lags; SF's remaining HCD-required action (fall 2026) would make conforming projects fully ministerial at entitlement.

Sources: CALES entitlement timelines (O'Neill, Biber et al.) · SF permit-timeline study (March 2026) · The Frisc on approvals vs. construction · Terner prevailing-wage cost analysis (2024) · Updated August 21, 2026

#3

Keep pushing ADUs — the proven fast lane

ADUs are California's cleanest housing-policy win: permits grew twenty-fold after state preemption, to 30,354 in 2024 — over a quarter of all homes permitted statewide. The Bay Area lags Southern California in uptake, which is exactly the opportunity.

Supply impact
3/5
Evidence strength
5/5
Public cost
5/5
Speed
5/5
How it works, the tradeoffs, and where it stands

How it works

State law made backyard and garage units ministerial, fee-exempt under 750 square feet, and immune to most local vetoes. Remaining acceleration is distribution and financing: pre-approved plan libraries, condo-ization (AB 1033) to unlock mortgage financing, and faster utility hookups.

The honest tradeoffs

Scale ceiling and slippage are the honest limits: these are small units; permits outrun completions (San Jose: ~3,000 permits, ~1,451 completions); and regional surveys find only about 45% of ADUs are rented to a paying tenant, with a fifth housing family for free. Cities may also be overcounting ADUs toward affordable RHNA targets without income restrictions. The upside nobody disputes: essentially zero displacement, since nothing is demolished.

Who bears it

Almost no one acutely — the cost is opportunity cost, borne diffusely by renters who still need apartment-scale supply.

Where it stands here

San Jose is the state model (pre-approved plans, first AB 1033 ADU condos); cities over 200k must now offer pre-approved plans; Berkeley moving toward AB 1033. Bay Area per-capita uptake still trails LA.

Sources: HCD ADU Handbook (2025 update) · CA YIMBY ADU reform retrospective · ABAG ADU affordability survey findings (2025) · Terner on ADU completion equity gaps · Updated August 21, 2026

#4

Cut, defer, and standardize development fees

California's impact fees average around $29,000 per unit against under $1,000 in Texas, and some Bay Area cities have charged several times that. Fees are a fixed cost that kills marginal projects — but the causal evidence on how many units fee cuts buy is thinner than the outrage.

Supply impact
3/5
Evidence strength
3/5
Public cost
3/5
Speed
3/5
How it works, the tradeoffs, and where it stands

How it works

Fees are due before revenue and mostly invariant to unit size, so they burden modest and multifamily units disproportionately. Cuts, deferral to occupancy (SB 937), and predictable estimates (AB 1820) lower the feasibility threshold and financing risk.

The honest tradeoffs

Fees fund real things — parks, roads, schools, affordable-housing funds — because Prop 13 constrains the alternatives; cuts shift costs to general funds or defer infrastructure, borne by existing residents. Incidence research complicates the story: fees partly capitalize into land prices, so cutting them partly enriches current landowners, and fees that visibly fund infrastructure can even grease approvals. The strongest honest claim is about the extreme outlier level of Bay Area fees, not a precise units-per-dollar elasticity.

Who bears it

City capital budgets and existing residents via infrastructure funding; partial windfalls to landowners.

Where it stands here

SF cut impact fees 33% for pipeline projects (through Nov 2026) and waived them for downtown conversions; San Jose waived construction taxes downtown; statewide, SB 937 defers fees to occupancy and AB 1483/AB 1820 force transparency. Fremont remains the high-fee cautionary tale.

Sources: RAND, The High Cost of Producing Multifamily Housing in CA (2025) · Terner Center, It All Adds Up (2018) · Terner/HCD impact fee study (2019) · Updated August 21, 2026

#5

CEQA infill exemptions — now implemented, watch the follow-through

AB 130 and SB 131 (June 2025) exempted most urban infill housing from CEQA outright. The mechanism is sound — removing 1-2 years of study and an unbounded litigation tail — but only 1-3% of projects were ever actually sued, so treat claims of transformation with patience until data arrives.

Supply impact
3/5
Evidence strength
2/5
Public cost
5/5
Speed
4/5
How it works, the tradeoffs, and where it stands

How it works

A statutory exemption removes environmental study costs, delay, and the litigation risk premium for qualifying infill projects; SB 131's rezoning exemption also de-risks the big upzonings (it is why SF's Family Zoning Plan needed no multi-year EIR).

The honest tradeoffs

Site-specific review and the public's main formal disclosure lever disappear for exempt projects — environmental-justice groups' core objection — though infill is the lowest-impact housing per unit and hazardous sites are carved out. Construction trades lost prevailing-wage leverage below 85 feet, a genuine transfer away from workers. Early permit surges (CA multifamily permits up ~105% year-over-year in Q1 2026) are confounded by rate cuts and other laws; no study isolates the CEQA effect yet.

Who bears it

Trades below 85 ft and neighbors/EJ communities who lose CEQA's disclosure-and-delay lever.

Where it stands here

In force since June 30, 2025; SF runs a dedicated exemption application and Bay Area land-use practices report active use; no public count of exemptions granted yet. A CalChamber ballot initiative (Nov 2026) would go further, keeping this contested.

Sources: Greenberg Traurig analysis · Rose Foundation, CEQA by the Numbers (litigation rates) · ABAG guidance (April 2026) · Updated August 21, 2026

#6

Attack the construction cost stack (modular, standards, labor)

Bay Area multifamily costs run roughly triple Texas levels, and RAND attributes most of the gap to policy choices rather than geography. It's the biggest theoretical lever in housing — and the one with the roughest real-world track record so far.

Supply impact
4/5
Evidence strength
2/5
Public cost
4/5
Speed
2/5
How it works, the tradeoffs, and where it stands

How it works

Factory-built construction promises 10-25% hard-cost savings and 20-50% faster schedules under the right conditions; standardizing approvals for factory-built housing, trimming prescriptive requirements, and growing construction labor supply attack the same cost stack from other angles.

The honest tradeoffs

The industrial record is sobering: Katerra burned over $2 billion before its 2021 bankruptcy and Veev liquidated in 2023 — factory economics need steady order pipelines that housing cycles don't provide, and even Vallejo's pioneering union factory has wobbled. Meanwhile 2025-26 immigration enforcement is shrinking construction crews and raising wages (about 60% of California construction workers are immigrants), actively eroding policy-won savings. On-site trades also bear work shifting to factories unless the factories are union.

Who bears it

Existing taxpayers (if fee cuts fund it), on-site trades, and investors carrying factory pipeline risk.

Where it stands here

Harbinger (ex-Factory OS, Vallejo) has delivered 4,000+ union-built modular homes; a 2026 Wicks package would streamline state factory-built approvals; RAND/Terner cost findings are actively driving SF and San Jose fee debates.

Sources: RAND cost decomposition (2025) · KQED on factory-built housing and the Wicks package · AGC workforce survey (Aug 2025) · Updated August 21, 2026

#7

Legalize single-stair buildings up to six stories

Requiring two staircases on 4-6 story buildings makes small-lot midrise infeasible and biases what does get built toward small units. The safety record of modern single-stair buildings is strong (zero exit-related fire deaths in NYC and Seattle over 12 years) — but California's reform died in committee this spring.

Supply impact
2/5
Evidence strength
3/5
Public cost
5/5
Speed
1/5
How it works, the tradeoffs, and where it stands

How it works

A second stair consumes 7.5-12% of construction cost and forces double-loaded corridors; allowing one stair with sprinklers and unit caps recovers floor area, enables family-sized cross-ventilated units, and unlocks narrow infill lots that cannot fit a compliant midrise at all.

The honest tradeoffs

Fire services bear perceived operational risk (egress redundancy, attack path, mobility-impaired evacuation) and residents any residual safety risk; the observational evidence says that risk is small in sprinklered modern buildings, but California data would be new construction under new rules, and the State Fire Marshal recommends capping at four stories pending further study. Supply magnitude is honestly modest: hundreds to low thousands of Bay Area units over a decade, concentrated on small infill lots.

Who bears it

Fire services (perceived risk) and residents (small residual risk) in exchange for cheaper family-scale infill.

Where it stands here

No Bay Area city has adopted it. AB 2252 (Lee) got a no-vote hearing in April 2026 and is dead for the session; the new federal ROAD Act directs HUD to issue single-stair guidance up to six stories, which may reopen the state debate.

Sources: Pew safety record study (Feb 2025) · CalMatters on the State Fire Marshal report · AB 835 (fire marshal study mandate) · Updated August 21, 2026

#8

Scale up public money for below-market homes — costly but irreplaceable

Subsidized housing is the only thing that produces homes for incomes the market never serves, and an $11.25 billion state bond is headed to the November 2026 ballot. But at Bay Area costs — up to $1 million per unit in SF — each public dollar buys few homes, and credible research finds each subsidized unit adds well under one net unit to total supply.

Supply impact
3/5
Evidence strength
3/5
Public cost
1/5
Speed
2/5
How it works, the tradeoffs, and where it stands

How it works

Deed-restricted rents can't service construction debt at Bay Area costs, so subsidy fills the gap through layered capital stacks (federal and state tax credits, bonds, soft loans). More public dollars close more stacks and build units targeted at low incomes.

The honest tradeoffs

Taxpayers carry roughly 30 years of debt service, and the crowd-out literature (Sinai-Waldfogel; Eriksen-Rosenthal) finds each subsidized unit yields 0.33-0.5 net new units, with local crowd-out sometimes near 100% — though it is theorized to be smallest exactly where private supply is most constrained, like here. Benefits concentrate on lottery-winning households rather than lowering market rents broadly, and labor and design standards attached to public money inflate the costs being subsidized. RAND puts CA subsidized hard costs at $511 per square foot — 4.4x Texas market-rate.

Who bears it

State taxpayers via decades of debt service; renters outside the lotteries see little direct benefit.

Where it stands here

The $11.25B Veterans and Affordable Housing Bond Act (AB 736/SB 417) is on the November 2026 ballot, claiming ~35,000 new deeply affordable homes statewide. The regional BAHFA bond remains dead after its 2024 withdrawal. State LIHTC held at $500M/yr — extended, not expanded. Terner's SB 555 social housing study is due to the Legislature December 2026.

Sources: Governor's announcement of the 2026 bond · Eriksen & Rosenthal on LIHTC crowd-out · RAND cost findings (2025) · HCD SB 555 social housing study · Updated August 21, 2026

#9

Office-to-residential conversion — a bet, not a track record

San Francisco has assembled the nation's most aggressive conversion incentive stack — and has completed zero major office-to-residential conversions in over four years. The constraint is arithmetic (deep floor plates, $400-700+ per square foot), not law.

Supply impact
2/5
Evidence strength
2/5
Public cost
2/5
Speed
2/5
How it works, the tradeoffs, and where it stands

How it works

Convert devalued, ~21%-vacant downtown office stock into housing by removing code barriers and closing the feasibility gap with tax and fee waivers plus tax-increment financing — supply without land acquisition or displacement.

The honest tradeoffs

The city's Downtown Revitalization Financing District pledges future property-tax increment — estimates run $610M to $1.2B over 30 years, roughly $100,000 per unit — revenue the general fund never sees, with a real risk of subsidizing conversions that would eventually have penciled anyway. Both flagship SF projects died in 2025 (988 Market abandoned; 785 Market reverted to office after starting construction). The district's own projection is ~4,400 units — a third of the theoretical ceiling.

Who bears it

Future SF general-fund taxpayers via 30 years of diverted tax increment — or downtown itself if the units never materialize.

Where it stands here

Prop C transfer-tax waiver, impact/inclusionary fee waivers, adaptive-reuse code path, and the Feb 2026 financing district are all live in SF; first subsidized completions realistically 2028-2030. Oakland and San Jose activity negligible.

Sources: SPUR/ULI feasibility study (2023) · SF Mayor's office on the financing district (Feb 2026) · Planetizen on zero completions (Feb 2026) · Updated August 21, 2026

#10

Vacancy taxes and land-value taxation — revenue tools, not supply tools

We include these because they come up constantly — and the evidence says they barely move supply. Vancouver's well-studied empty homes tax cut vacancies ~21% but had no measurable effect on rents or construction; SF's version has been struck down in court and never collected a dollar; and land-value taxation has no viable path through Prop 13.

Supply impact
1/5
Evidence strength
3/5
Public cost
5/5
Speed
3/5
How it works, the tradeoffs, and where it stands

How it works

Vacancy taxes raise the cost of holding habitable units empty, activating existing stock (a one-time effect; nothing is built). Land-value taxation would genuinely incentivize construction by taxing land rather than improvements — in a state whose constitution didn't preclude it.

The honest tradeoffs

The best-case result (Vancouver) activated roughly 1-2% of housing stock, once, with zero measured rent benefit to tenants — so even success is mostly a revenue program funding affordable housing. The taxes fall on second homes, probate properties, and units mid-renovation; self-declaration invites evasion; and the SF Superior Court's takings/Ellis Act ruling (under appeal, fully briefed as of mid-2026) casts legal chill statewide. Split-roll's 2020 defeat and Prop 5's 2024 defeat mark the political boundary for anything land-tax-shaped.

Who bears it

Owners of intermittently used or hard-to-rent properties; cities bear litigation defense.

Where it stands here

Oakland's Measure W and Berkeley's Measure M are collecting (single-digit millions per year, marginal unit effects); SF's Prop M is suspended pending appeal with zero dollars collected.

Sources: C.D. Howe study of Vancouver's empty homes tax · SF Examiner on Prop M suspension · Berkeleyside on Measure M first-cycle data · Updated August 21, 2026

The rubric behind these scores is on the methodology page.