Rent control: does it really destroy housing supply?

That single number has become ammunition in one of the most entrenched debates in housing policy — but it is also the least interesting part of the story. The real question is not whether rent control shrinks supply. It is how it reshapes supply, for whom, and whether the mechanism of that reshaping can be engineered differently.

The answer, as nearly every serious empirical study in the past decade has shown, depends almost entirely on the architecture of the policy itself.

The Mechanics of Supply Contraction: Beyond New Construction

The popular framing of rent control's supply effect fixates on a single narrative: developers stop building. It is clean, intuitive, and incomplete. The empirical record shows that the primary supply-side responses to rent regulation are not about cranes leaving the skyline. They are about what happens to buildings that already exist.

When landlords face capped returns on a regulated property, they do not simply accept lower margins. They restructure. The documented mechanisms read like a decision tree:

• Condominium conversion. Units exit the rental market entirely by being sold as individually owned properties — a permanent removal from the rental stock.
• Redevelopment. Older regulated buildings are demolished or substantially renovated into uses that fall outside the rent-control ordinance — often luxury units, commercial space, or owner-occupied housing.
• Sale to owner-occupants. A landlord sells a treated building to someone who intends to live in it, eliminating multiple rental units in a single transaction.
• Deferred maintenance. When rental income is capped, the incentive to invest in upkeep diminishes. Properties deteriorate, code violations accumulate, and units eventually become uninhabitable or are withdrawn from the market.
• Tenure shifts. Some landlords convert rental units to short-term vacation rentals, office space, or other non-residential uses that escape regulation.

These are not hypothetical adjustments. They are the observed pathways through which regulated rental inventory contracts. A 2020 review of the rent-control literature found that condominium conversion, redevelopment, and changes in the existing rental stock are the commonly identified supply channels — not a wholesale halt in new construction. In fact, the same review noted that some studies found little effect of rent regulation on new building permits.

This distinction matters. If the primary supply response is conversion and withdrawal of existing units rather than a freeze on new construction, then the policy design lever sits in a different place than most political debates assume.

The damage rent control inflicts on housing supply is rarely a dramatic halt in construction. It is a quiet, structural withdrawal of existing units from the rental market — conversion by conversion, building by building.

The 2025 Cross-City Evidence: Quantifying the Rental Stock Shift

The most comprehensive recent data comes from the June 2025 study published in the Journal of Housing Economics. It examined rent-control policy changes across 27 U.S. metropolitan areas over the period from 2000 to 2021, using panel and staggered-treatment models to isolate the association between regulatory shifts and rental housing outcomes.

The headline finding — a roughly 10% reduction in the total number of rental units in cities that adopted more restrictive rent-control reforms — is striking but demands careful parsing. This is an average association across a large and heterogeneous sample. It is not a universal law, and it should not be read as a causal estimate that will replicate identically in every jurisdiction. The study's authors used newspaper-based identification of reform timing, with an endpoint of April 2021, and the results reflect the particular policy designs and market conditions of the cities in the dataset.

What makes this study structurally important is not the aggregate number but the disaggregation. The researchers tracked not just the total rental stock but its composition by income affordability tier — and the results there tell a far more complicated story than either side of the rent-control debate typically acknowledges.

The Affordability Paradox: Winners and Losers in Regulated Markets

Here is where the standard narrative fractures. The same 2025 study that found a 10% aggregate rental-stock reduction also found that:

Rent regulation, in other words, did not uniformly destroy affordable housing. It restructured it. The number of units accessible to the poorest renters — those earning below 30% of area median income — increased by approximately half. At the same time, the inventory of units affordable to higher-income renters declined by nearly the same magnitude.

This is the affordability paradox at the heart of rent control's empirical record. The policy simultaneously created affordability at the bottom of the income distribution while eroding it in the middle and upper tiers. The net effect on total supply was negative, but the distributional consequences were asymmetric.

The critical caveat is the lower-bound estimates. The 52% increase in extremely-low-income-affordable units could be as low as 11%; the 46% decline in higher-income-affordable units could be as low as 4%. The precision of these figures varies substantially across model specifications and city characteristics. What remains robust across specifications is the direction: more restrictive regulation shifts the affordability composition of the rental stock downward, even as it reduces total inventory.

This creates a genuine policy dilemma — not the cartoonish version where rent control is either salvation or catastrophe, but a structural trade-off where gains for the most vulnerable tenants come at the cost of overall market contraction and reduced options for households above the poverty line.

San Francisco and Cambridge: Lessons from Natural Experiments

Two of the most cited studies in the rent-control literature exploit specific policy changes as natural experiments, and both illuminate different facets of the same structural dynamic.

San Francisco: The Conversion Cascade

In 1994, San Francisco expanded its rent-control ordinance to cover small, older multifamily buildings that had previously been exempt. A landmark study published in the American Economic Review in September 2019 used this expansion as a treatment event and tracked its effects over subsequent years.

The findings were precise. Landlords of treated properties reduced rental housing supply by 15%, primarily through two channels: sales to owner-occupants and redevelopment into properties exempt from regulation. The study also documented a 20% reduction in renters' mobility — tenants in controlled units stayed longer, which is both a stability benefit for incumbents and a friction that reduces market dynamism.

The most consequential finding was directional. The reduction in rental supply likely contributed to higher market rents in the long run, even as incumbent tenants benefited from lower rents and reduced displacement in the short run. Rent control, in the San Francisco case, acted as a transfer from future renters to present ones — a structural intertemporal trade-off that is rarely made explicit in political debates.

The San Francisco evidence also complicates the "landlords are greedy" narrative. The supply response was not driven by spite or profit-maximization in the abstract. It was a rational reallocation of capital. When the regulated return on a rental property falls below the return available from converting it to another use, capital flows accordingly. The landlord does not need to be villainous; the incentive structure does the work.

Cambridge: The Decontrol Signal

Massachusetts provides the mirror image. In 1995, the state eliminated stringent rent controls that had been in place in Cambridge, creating a sharp policy discontinuity. Research using property-level data from 1988 to 2005 found that decontrol generated substantial price appreciation at formerly controlled properties — and, crucially, at nearby never-controlled properties as well.

The numbers are large. Between 1994 and 2004, Cambridge residential property values increased by approximately $7.8 billion. About one-quarter of that increase — roughly $1.8 billion — was attributed to the end of rent control. The spillover effect onto never-controlled properties suggests that rent regulation had been suppressing values across a broader geographic radius than the regulated addresses alone.

This finding carries a specific implication for policy design. Rent control does not merely affect the buildings it covers. It radiates outward, shaping expectations, investment patterns, and property valuations in surrounding areas. The regulatory boundary of an ordinance is not the boundary of its market effect.

Policy Design as the Deciding Factor in Market Outcomes

The most important sentence in the 2024 Urban Institute research brief on rent regulation is also the most unglamorous: outcomes depend on the specific design and implementation of local rent-control laws.

This is not an evasion. It is the central empirical finding that decades of research have converged upon. Rent control is not a single policy; it is a spectrum of regulatory architectures, each producing different supply-side and affordability outcomes. The variables that matter include:

• Rent-cap formula. Whether increases are tied to CPI, fixed percentages, or negotiated benchmarks — and whether the cap resets between tenancies (vacancy decontrol) or persists regardless of occupancy.
• Coverage and exemptions. Which buildings are covered, by age, size, and type. New construction exemptions are particularly consequential for maintaining the development pipeline.
• Owner-occupancy provisions. Whether landlords can reclaim units for personal use, which affects the conversion pathway.
• Enforcement intensity. A law that caps rents on paper but lacks enforcement infrastructure produces different outcomes than one with active monitoring and penalties.
• Interaction with other regulations. Zoning, permitting timelines, and density restrictions shape the feasibility of the redevelopment and conversion pathways that landlords use to exit the regulated market.

The 2020 literature review's observation that findings differ across jurisdictions and policy regimes — including studies that found little effect on new construction — is not a sign of sloppy research. It is a sign that rent regulation is a policy instrument with highly context-dependent outcomes. Treating it as a binary (good or bad, builds or destroys) is analytically useless.

The empirical record does not support the claim that rent control universally destroys housing supply. Nor does it support the claim that rent control painlessly delivers affordability. What it demonstrates, with increasing precision, is that restrictive regulation reshapes the rental stock in specific, measurable ways: it reduces total inventory, shifts affordability toward the lowest income tiers, benefits incumbent tenants at the expense of future ones, and triggers structural responses — conversion, redevelopment, tenure change — that are predictable if policymakers choose to look for them.

The question for any jurisdiction considering rent regulation is not the ideological one. It is the engineering one: which design parameters produce which supply and affordability outcomes in this specific market, at this specific density, with this specific housing stock? The evidence is clear enough to answer that question with far more precision than the political debate typically allows. The difficulty is not in the data. It is in the willingness to read it.