An annotated review of the evidence. For each source: what it found, what it supports, and, honestly, where it cuts against me.
A note on method. I've tried to represent each source the way its own author would, not the way that flatters my case. Where the research is mixed, I say so. Where a finding complicates my argument, it's labeled CAVEAT and kept in.
The honest through-line of this literature is not "deregulation fails" and not "supply solves everything." It's narrower and more useful: upzoning is necessary but slow, uneven, and market-dependent, and it produces affordable, non-displacing outcomes only when paired with financing, public land, and protection for existing residents. That is a claim the evidence supports from several directions at once.
Theme IDoes upzoning, by itself, produce housing?
The core empirical question. The most-cited study and its own author's later, more careful synthesis.
Upzoning Chicago: Impacts of a Zoning Reform on Property Values and Housing Construction
FindsUsing difference-in-differences on Chicago parcels upzoned near transit, Freemark detected significant, robust increases in property values, including for existing condos, but, in his words:
"no impacts of the reforms… on the number of newly permitted dwellings over five years… the short-term, local-level impacts of upzoning are higher property prices but no additional new housing construction."
SupportsThe central claim that permission to build is not the same as building, and that upzoning a parcel raises its value immediately, which is where the value-capture and displacement pressure begins.
CaveatThis is explicitly a short-term, local finding. Five years may be too short, since development lags policy. The supply side fairly notes this, and so does Freemark.
Zoning Change: Upzonings, Downzonings, and Their Impacts
FindsReviewing the literature, Freemark concludes upzoning's effects are genuinely mixed: outcomes "depend on market demand, local context, housing types, and timing." Local effects are uncertain; regional-scale upzoning appears more positive. Downzoning, by contrast, clearly worsens affordability.
SupportsThe "necessary but not sufficient, and context-dependent" framing. Citing this alongside his 2020 paper is the honest move: it shows the value-capture finding wasn't cherry-picked, and that the researcher himself holds a nuanced view.
Urban Institute, Housing Matters: How Upzoning Affects Housing Supply
FindsUpzoning does increase supply, but slowly and unevenly: NYC neighborhood upzonings added 4,000+ units within four years, with effects growing 6–12 years out. It works mainly where demand is already strong; weak-market areas saw little. Nationally, municipal upzoning alone is "only inconsistently associated with increases in actual density." Its conclusion: complementary regulations and investments must sometimes run in parallel.
SupportsThe entire thesis, almost verbatim. Upzoning helps, slowly and conditionally, and needs complementary tools to deliver affordability. This is the strongest single source for "pair it, don't just permit it."
CaveatIt also finds that heavy-handed affordability mandates can suppress building (Seattle developers built just outside the mandate boundary). That's a real complication, and it's why I argue for opt-in, incentive-based tools rather than blunt mandates.
The Ann Arbor Comprehensive Land Use Plan (CLUP), the document this essay responds to
Adopted by City Council 2026 (unanimous Planning Commission recommendation; Council adoption after the Draft 3 public-comment period closed Jan. 5, 2026). City of Ann Arbor / Engage Ann Arbor (plan.a2gov.org); Michigan Public and CBS Detroit coverage of the adoption hearings.
FindsThe CLUP is a 25-year guiding policy document required by state law, explicitly not legally binding and not itself a change to zoning. It eliminates exclusively single-family zoning (allowing duplexes/triplexes in a tiered residential/transition/hub structure) and is framed in market/supply terms, but the city's own materials also name owner-occupied missing middle, attainable entry-level homes, and public/private partnership for missing-middle affordability as goals.
SupportsTwo things at once. First, "permission isn't a building": the plan sets goals; the binding zoning rewrite is a separate, district-by-district process that will take years, so for the near term there are no new by-right rules to act on. Second, the displacement worry is on the record, raised by multiple residents at the hearings, and so is the "keep nurses and teachers in town" hope (a resident testified she wanted teachers, nurses, and her dental hygienist able to live here rather than commute from Canton). That hope is the essay's hook: the author is that nurse.
CaveatDo NOT attribute the "nurses and teachers" line to Mayor Taylor as an adoption quote, it was resident hearing testimony. The same framing also appears prominently in the Ann Arbor Education Association's endorsement of Rabhi (nurses/teachers "working without contracts," workforce housing), so it tilts toward the labor side rather than being neutral civic language. The "rewrite not done before end of decade" timeframe is a reasonable inference from how long such rewrites take, not a published city deadline, hedge as "multi-year, just beginning."
Theme IIWhat does the speculative market do with housing?
The argument that housing left fully to leveraged capital is fragile and extractive, grounded in a building you can drive past.
The Village Green / "Haven of Ann Arbor" ownership churn, and the collapse of a $2B apartment empire
Trade and local coverage of the 2017 $125M portfolio sale, the LLC re-ownership, and the subsequent foreclosures and receiverships across the operator's Midwest portfolio.
SupportsThat housing held as a speculative asset is traded in bundles, rebranded for yield, owned at arm's length through LLCs, and ultimately hostage to interest rates, none of which tracks whether people need homes. A concrete, local illustration of commodification.
CaveatThe Ann Arbor building was sold off before the operator's collapse; I have not found a specific Ann Arbor property that was foreclosed by name. The accurate claim is "our stock was inside this portfolio and churn," not "an Ann Arbor tower was foreclosed." Punch at the model, not the man.
City Place & the rejected Heritage Row proposal
Local reporting on the Germantown-block development fight and the as-of-right project that replaced it.
SupportsThat a forced "build-it-or-block-it" binary can produce the worst outcome for everyone: the proposal with affordable units failed, and the as-of-right fallback delivered market-rate student housing with zero affordable units. The middle option existed and died in the crossfire.
CaveatThis is an affordability argument, not a preservation one. Plenty of old stock should come down; the point is about what replaced it and the process that forced the choice, not the loss of the buildings themselves.
Theme IIIThe cost of staying: construction labor.
A displacement vector that's about keeping a home, not buying or building one.
Construction labor shortage & rising repair costs
Home Builders Institute labor reports; NAHB 2026 outlook; ABC workforce estimates; trade and consumer reporting (2025–2026).
FindsA structural shortage of roughly 350,000 construction workers per month; nonsupervisory construction wages up ~9.2% year over year, passed to homeowners through estimates; home-repair costs up nearly 18% over three years; ~7 in 10 homeowners delaying repairs they can't afford. Experts attribute it to an aging workforce, too few new trade entrants, and immigration policy, and call it structural, not temporary.
SupportsThat the cost of maintaining an old home is rising independent of taxes or zoning, a quiet displacement for cash-constrained owners, and that a building boom bids up the same scarce labor, so deregulation can raise the cost of keeping the home you already own.
CaveatThis is national data, not Ann Arbor-specific, and the "a local boom measurably raises local repair costs" link is a sound mechanism rather than a figure measured for one mid-size city. Presented as the mechanism, not a local measurement.
Theme IVThe tools that actually work, and where.
Proof of concept for the "pair it with" half. The strongest examples are towns our size or smaller.
Community land trusts & shared-equity / capped resale
Champlain Housing Trust (Burlington, VT); Summit Hill CLT (Bloomington, IN); national CLT network data.
FindsCLTs hold land permanently and sell the home on a long ground lease with a resale formula (Bloomington gives owners 25% of appreciation). Champlain has made 679 homes affordable for 1,300+ successive owners over 40 years; CLT owners default roughly ten times less often than market owners; 225+ CLTs operate nationally.
SupportsThat capped resale is a real, mature, opt-in mechanism: the owner trades part of the windfall for an affordable entry price, stability, and permanence, and the unit stays affordable for the next family with no new subsidy.
CaveatBy design it limits the owner's resale windfall. That's a feature, not a bug, but it must be stated plainly: this serves people whose goal is to stay, not to flip.
Michigan brownfield housing TIF (Public Act 90 of 2023)
Amendment to Michigan's Brownfield Redevelopment Financing Act; applied locally to the Ann Arbor CLT's Townie Homes.
FindsThe 2023 reform untethered brownfield TIF from contamination: any parcel now qualifies if developed with housing up to 120% of area median income, and eligible costs include site prep, infrastructure, and the financing gap, county-approved, so the DDA cap doesn't apply.
SupportsThat the financing tool already exists and already works here, the strongest local proof. The proposal is to make this pathway navigable for individual homeowners and small groups, not only nonprofits and large developers.
CaveatA homeowner-accessible version is a proposal, not an existing program. The Townie Homes use is the proven part; the broadened access is the ask.
Homeowner ADU finance, pre-approved plans & city-as-coordinator
Montpelier, VT (ADU grants + 0% loans); LA Backyard Homes Project (up to $75K forgivable) & Standard Plan Program; Boston deferred-equity loans; Atlanta Backyard ATL; San Jose modular "buyer's club"; Cleveland modular RFP; YouthBuild.
FindsCities from 8,000 people (Montpelier) up are funding individual homeowners to build, publishing pre-permitted plans to cut cost and time, and exploring bulk-purchase of modular units to capture factory pricing a lone homeowner can't. YouthBuild's model trains young people by building affordable housing.
SupportsThat "the city can help ordinary owners build" is not hypothetical, and that scale (Montpelier < Ann Arbor) is not the obstacle. The pieces exist; what's missing is the decision to assemble them here.
CaveatSan Jose's and Cleveland's are recent/early-stage; frame as "cities are now doing this," not "proven at scale for decades." The trades-apprenticeship pairing is a complementary opportunity, not an existing Ann Arbor program.
NYC "Plus One ADU" homeowner financing program
New York City (reopened March 18, 2026, under the Mamdani administration), with New York State Homes & Community Renewal ("Plus One ADU" statewide program, ~$85M). City + state partnership.
FindsProvides homeowners up to $395,000 in combined support, low/no-interest city loans plus state construction grants, to build an ADU, much of it forgivable over a 10-year owner-occupancy compliance period, paired with a pre-approved plan library and simplified permitting.
SupportsThe single cleanest proof that homeowner-infill financing is a real, current policy choice a government can make, not a fantasy, which is exactly the gap neither Ann Arbor mayoral candidate has filled. Directly answers "but who would pay for an ordinary owner to build?"
CaveatPilot-scale (earlier rounds funded on the order of dozens of homeowners; only ~25% of ~2,800 first-call applicants were eligible) and dependent on a city+state partnership. Frame as "the mechanism exists / a major city just reopened it," NOT "proven at mass scale." Michigan's closest enabling analog is the brownfield housing TIF (PA 90/2023), not a homeowner ADU fund, so an A2 version would still need a funding source/state partner.
Modular construction across scales, and the North Corktown (Detroit) infill project
Oakland six-story modular affordable (Factory_OS); Singapore PPVC high-rises (Clement Canopy, 40 stories); Korea LH 22-story modular (2025); Detroit North Corktown "Tomorrow's Housing Innovation Showcase" (Gilbert Family Foundation + North Corktown Neighborhood Association, factory-built homes on Land Bank parcels via a community land trust, ≤80% AMI). Detroit News (Feb 2024); Bridge Detroit / Detroit Free Press follow-up (Apr–May 2025).
FindsModular is a manufacturing method, not a building type: it scales from infill duplexes to six-story mid-rises (the proven sweet spot) to 40-story towers. North Corktown assembled the full stack locally, modular + public land + community land trust + real mortgage financing at or below 80% AMI.
SupportsThat the whole "pair it" model is being attempted in Michigan, an hour away, not just on the coasts. And the central "two halves of one machine" point, framed as sequencing: North Corktown built homes (2023–24) as a philanthropic pilot ahead of a citywide framework; Ann Arbor adopted the framework (CLUP) first but hasn't built the financing/CLT/modular delivery tools, and the district-by-district zoning rewrite that gives the plan teeth is a multi-year process (realistically not complete before the end of the decade). Neither order delivers alone, and the rewrite window is precisely the runway to build the delivery tools.
CaveatDo NOT claim "Detroit has no zoning plan", it does. Detroit is actively reforming zoning right now (a "Let's Build More Housing" by-right triplex/fourplex + ADU ordinance moving through council in late 2025; an adaptive-reuse ordinance passed Nov 2025; the CHIP missing-middle/shared-equity incentives effective Feb 2025). The accurate point is narrower: the North Corktown pilot ran ahead of, and disconnected from, that reform. Also: North Corktown had a troubled rollout, delays, homes vacant/boarded for months, quality and appearance complaints, so cite it honestly as "why to do this deliberately and at scale," not as a finished success. And high-rise modular has had execution problems elsewhere (NYC's 32-story tower), so mid-rise is the safer claim.