Will the new federal housing bill meaningfully expand affordable housing supply in the markets where costs are most burdensome?

Advocate A’s weakest point is not the claim that more supply can reduce rents; I concede that as an empirical matter. The weak point is the leap from “supply works in Austin and Minneapolis” to “this federal bill will meaningfully expand affordable supply in the most burdened, exclusionary markets.” Austin is evidence for my side as much as A’s: rents fell where the pipeline was enormous and local rules allowed large-scale building. The core question is whether federal incentives can create that outcome in coastal metros and high-opportunity suburbs where zoning, permitting, land costs, neighborhood vetoes, and construction capacity are the binding constraints.

The bill appears to contain real but limited levers: federal regulatory streamlining, environmental-review changes, financing expansion, manufactured-housing support, and reporting or best-practice frameworks for local land-use reform. But reporting restrictive zoning, publishing HUD best practices, or offering grants is not the same as preempting minimum lot sizes, apartment bans, parking mandates, discretionary review, or litigation delays. If localities can qualify by planning, reporting, using pattern books, or building at rates already above the median, the likely geographic effect is selection: money flows to places already politically and administratively willing to build. That may increase output somewhere, but not necessarily where affordability pressure is worst.

The second problem is capacity and cost. NAHB reported in 2026 that builders still face elevated mortgage rates, rising material and labor costs, development fees, environmental regulations, and labor shortages; it also estimated material costs up 46.1% since February 2020. Harvard’s 2026 housing work similarly found construction activity softening and new market-rate construction largely unaffordable for lower-income households. In that environment, additional credits and grants can close some project gaps, but they also bid against scarce land, contractors, and entitled parcels. In high-cost markets, the marginal subsidy may be capitalized into land prices or absorbed by higher total development costs rather than producing many net new affordable units.

Third, LIHTC precedent should make us cautious. LIHTC has produced millions of restricted units, but after nearly four decades the U.S. still has record renter cost burdens and a sharply shrinking low-rent stock. Eriksen and Rosenthal’s Journal of Public Economics study found substantial crowd-out of unsubsidized rental construction in some geographic estimates, suggesting that subsidized units are not always one-for-one net additions. Other LIHTC research finds location incentives and profitability often pull projects toward lower-market-rent areas because the opportunity cost of restricted rents is lower there. That is not a fatal indictment of LIHTC, but it undercuts A’s assumption that expanding the tool automatically solves scarcity in the most expensive markets.

The best test is not national unit counts. It is whether completions, not just allocations, rise in high-cost, low-vacancy metros and exclusionary suburbs; whether permitting times fall; and whether rents fall outside the already supply-heavy Sun Belt. Current rent relief is concentrated in markets with large pipelines, while constrained coastal markets remain tight. That pattern suggests Washington can subsidize production, but without local land-use change it is more likely to subsidize scarcity than overcome it.

A’s rebuttal concedes the central weakness in its own case: “federal incentives cannot by themselves preempt local apartment bans, discretionary permitting, litigation delays, or other exclusionary zoning barriers.” That is not a marginal caveat; it is the mechanism question. If the highest-cost markets are expensive because they combine scarce legal building sites, slow approvals, high land prices, neighborhood veto points, and high construction costs, then a bill that mostly offers financing, best-practice frameworks, environmental-review streamlining, and grants can only work where local governments voluntarily convert those incentives into buildable capacity. The implication is selection bias: the bill is most likely to fund places already willing and able to build, while the most exclusionary suburbs and coastal metros remain supply constrained.

A tries to escape this by saying the bill can succeed “if” incentives are conditioned on actual production and permitting reform. But that “if” is doing the work. CRS’s summary describes zoning frameworks, CDBG-related reporting or eligible-use changes, grants, and federal review changes; it does not show a federal override of minimum lot sizes, apartment bans, parking mandates, discretionary local approvals, or local litigation. AP likewise describes a broad supply bill that reduces federal regulations and streamlines environmental reviews, but federal streamlining does not create local zoning capacity. A’s rebuttal therefore proves only that Washington can accelerate projects already legally and politically viable, not that it can produce affordable homes at scale in the places where local exclusion is the binding constraint.

The remaining evidence reinforces B’s position. First, construction capacity and costs limit conversion of subsidy into units. NAHB reported in May 2026 that construction materials were up 46.1% since February 2020 and that the sector still had more than 200,000 unfilled jobs. In that setting, additional credits and grants may raise bids for scarce entitled land, contractors, and materials. That is especially likely in high-cost metros where land-use barriers restrict the number of feasible parcels.

Second, LIHTC’s scale is real but insufficient to solve scarcity. HUD reports 54,102 LIHTC projects and 3.7 million units placed in service through 2023, yet Harvard reports 22.7 million cost-burdened renter households in 2024 and a major decline in low-rent units from 2014 to 2024. The issue is not whether LIHTC produces units; it does. The issue is whether expanding a rationed, place-based subsidy can materially change market scarcity in exclusionary regions. Eriksen and Rosenthal’s Journal of Public Economics evidence on LIHTC crowd-out shows that subsidized construction is not always one-for-one net new rental supply. That finding directly weakens A’s reliance on gross LIHTC production counts.

Finally, the rent evidence still favors B’s test. Rent declines are most visible where pipelines are large and local rules permit construction; A’s Minneapolis example is a local reform story, not proof that federal money can force resistant jurisdictions to build. Even with A’s rebuttal, the core issue remains unanswered: will the bill change land-use constraints in the most burdened markets, or merely subsidize units in places already predisposed to add them? The evidence so far points to the latter risk.

A’s strongest argument is that the bill is not a pure demand subsidy; it contains supply-side tools that can lower project costs, shorten federal reviews, expand financing, and reward jurisdictions that actually build. A can also fairly argue that LIHTC is a functioning production platform, with HUD reporting 54,102 projects and 3.7 million units placed in service through 2023, and that the disappearance of low-rent units creates a real need for restricted-rent production, not just vouchers or buyer aid. In the strongest version, federal money does not need to coerce every exclusionary suburb; it only needs to change incentives for enough high-pressure jurisdictions at the margin to generate measurable starts, completions, preservation, and rent moderation.

I still believe B’s position is stronger because the bill’s central tools do not appear to bind the central constraint. The AP description emphasizes federal regulatory reduction, environmental-review streamlining, expanded financing, manufactured or innovative housing, renter protections, and money for localities that build above the median rate. CRS describes zoning best-practice frameworks, planning-oriented approaches, grant eligibility, CDBG-related changes, environmental-review changes, and finance provisions. Those mechanisms may accelerate projects with legal sites, but they do not directly repeal apartment bans, minimum lot sizes, parking mandates, discretionary local review, local litigation, or neighborhood veto points. If those are the binding constraints in the most expensive metros and suburbs, then the bill’s most predictable effect is selection: funds flow to jurisdictions already politically willing and administratively able to approve housing.

A’s LIHTC point also proves less than A needs. HUD’s 3.7 million-unit count shows that LIHTC can produce and preserve restricted units, but Harvard reports 22.7 million cost-burdened renter households in 2024 and more than 7 million real sub-$1,000 rental units lost from 2014 to 2024. That gap after decades of LIHTC suggests the program is important but insufficient as a market-wide scarcity solution. Eriksen and Rosenthal’s Journal of Public Economics evidence that LIHTC can crowd out unsubsidized rental construction in some settings further weakens reliance on gross subsidized-unit counts as proof of net supply expansion.

Finally, the cost environment supports B’s caution. NAHB reported material costs up 46.1% since February 2020 and more than 200,000 unfilled construction jobs in 2026. When entitled land, contractors, labor, and materials are scarce, added subsidies can be capitalized into land prices or absorbed by higher development costs. The specific evidence that would change my mind would be causal, metro-level post-enactment data showing that bill-funded awards caused large net increases in permits and affordable completions in high-cost, low-vacancy, historically exclusionary markets—not merely in already-permissive Sun Belt or reform cities—and that rents or rent burdens fell relative to matched controls without major crowd-out or land-price capitalization.