Design Overview
Policy Events
Oregon SB 608 (effective 2019-02-28) — caps annual rent increases at CPI + 7 percentage points on most rental units over 15 years old. First statewide rent cap in U.S. history.
California AB 1482 (effective 2020-01-01) — caps increases at CPI + 5 pp on most units over 15 years old. Applies statewide to ~8 million rental units.
Washington HB 1217 (effective 2025-05-07) — 7% annual cap. Included for descriptive tracking only; almost no post-policy data exist.
Comparison Group
11 donor states with no statewide rent cap during the sample period:
Arizona, Colorado, Florida, Georgia, Idaho, North Carolina, Nevada, Tennessee, Texas, Utah, Virginia.
Sensitivity checks restrict to western-only donors (AZ, CO, ID, NV, UT) and run
leave-one-donor-out (dropping AZ, CO, or FL in turn) to test whether results depend on any single state.
Research Questions
Q1–Q10| Q | Question | Domain | Source | Coverage | Status |
|---|---|---|---|---|---|
| Q1 | Did caps reduce occupied-rent growth? | Rent | ACS | 2015-19, 21-24 | Limited |
| Q2 | Did caps reduce renter cost burden? | Affordability | ACS | 2015-19, 21-24 | Limited |
| Q3 | Did caps increase residential stability? | Mobility | ACS | 2015-19, 21-24 | Limited |
| Q4 | Did caps reduce housing supply? | Supply | BPS | 2010-2024 | Answered |
| Q5 | Did caps change housing prices? | Capitalization | FHFA | 2010-2025 | Answered |
| Q6 | Did caps change labor-market outcomes? | Labor | QCEW | 2014-2024 | Limited |
| Q7 | Larger effects where burden was higher? | Heterogeneity | ACS+ES | CA vs. OR | Suggestive |
| Q8 | Larger effects where supply was constrained? | Heterogeneity | BPS+ES | CA vs. OR | Suggestive |
| Q9 | Did impacts differ across outcome domains? | Cross-domain | All | Varies | Summary |
| Q10 | Were CA and OR similar enough to pool? | Poolability | All | Full | Answered |
Cross-Outcome Results
Baseline pooled estimatesPost-Treatment Shift by Outcome
Each bar shows the average post-treatment event-study coefficient as a percentage of the treated pre-policy mean. Outcomes are grouped by data source and domain.
| Domain | Outcome | What it measures | Source | Pre-mean | Post shift | % of mean | Grade |
|---|---|---|---|---|---|---|---|
| Supply | Total permits | New privately-owned housing units authorized by building permits (annual, state-level) | BPS | 51,845 | -12,819 | -24.7% | Answered |
| Supply | 5+ unit permits | Permits for buildings with 5 or more units (multifamily construction) | BPS | 23,890 | -5,044 | -21.1% | Answered |
| Supply | Multifamily share | Fraction of total permits going to 5+ unit buildings | BPS | 0.459 | -0.041 | -8.9% | Answered |
| Capitalization | FHFA HPI | Purchase-only house price index (seasonally adjusted), measuring home sale price trends | FHFA | 261.4 | -18.7 | -7.1% | Answered |
| Rent level | Median gross rent | Median monthly rent paid by current occupants (not asking rents or new-lease prices) | ACS | $1,270 | -$9 | -0.7% | Limited |
| Affordability | Rent burden 30%+ | Share of renter households paying 30% or more of income toward rent | ACS | 53.1% | -0.5 pp | -1.0% | Limited |
| Stability | Same house 1yr | Share of population living in the same house as one year ago (residential stability) | ACS | 85.0% | +0.3 pp | +0.4% | Limited |
| Stability | Moved last year | Share of population that moved within the past year | ACS | 15.0% | -0.3 pp | -2.1% | Limited |
| Stability | Moved diff state | Share that moved from a different state in the past year (interstate migration) | ACS | 2.3% | +0.7 pp | +31.7% | Limited |
| Tenure | Renter share | Share of occupied housing units that are renter-occupied | ACS | 42.2% | +0.4 pp | +0.9% | Limited |
| Labor | QCEW weekly wage | Average weekly wage across all covered employers (total private + government) | QCEW | $1,115 | +$42 | +3.8% | Exploratory |
| Labor | QCEW employment | Total covered employment level (workers covered by unemployment insurance) | QCEW | 9,999K | -235K | -2.3% | Exploratory |
Evidence grading: Answered = full coverage, clean pre-trends, all checks run. Limited = informative but coverage gaps or pre-trend issues reduce confidence. Exploratory = short pre-period or contaminated placebo; directional only.
California vs. Oregon
State-specific interactionsPost-Treatment Effects by State
Average post-treatment coefficient as percent of each state’s own pre-policy mean, from state-specific event-time interactions. With only two treated states, these are descriptive contrasts.
| Outcome | Domain | CA pre-mean | CA post (%) | OR pre-mean | OR post (%) |
|---|---|---|---|---|---|
| Total permits | Supply | 85,114 | -12.9% | 14,879 | -96.4% |
| 5+ unit permits | Supply | 40,020 | -11.7% | 5,968 | -95.6% |
| Multifamily share | Supply | 0.503 | -6.7% | 0.410 | -12.3% |
| FHFA HPI | Capitalization | 215.8 | -20.1% | 312.1 | +1.0% |
| Median gross rent | Rent | $1,453 | +0.6% | $1,042 | -2.5% |
| Rent burden 30%+ | Affordability | 54.8% | -2.0% | 51.0% | +0.3% |
| QCEW employment | Labor | 16.8M | -1.7% | 1.8M | -8.7% |
| QCEW weekly wage | Labor | $1,248 | +6.3% | $956 | +1.4% |
California has higher rents, more renters, and lower per-renter permit rates. The FHFA signal is driven almost entirely by California (−20.1% vs. +1.0%). Oregon shows a massive proportional permit decline (−96%) but near-zero HPI movement. These reflect different baselines, not a powered test of moderating mechanisms.
Pre-Policy State Profiles
2015–2018 meansPermits per 1,000 Renter Households
New housing permits relative to the renter population, showing how supply-constrained each state was before policy adoption. Lower values indicate tighter supply relative to demand.
| State | Role | Burden 30%+ | Renter share | Median rent | Permits/1K | FHFA HPI |
|---|---|---|---|---|---|---|
| California | Treated | 55.2% | 45.8% | $1,453 | 18.0 | 249.2 |
| Oregon | Treated | 51.0% | 38.0% | $1,042 | 32.0 | 370.3 |
| Arizona | Donor | 48.0% | 36.4% | — | 39.6 | 276.5 |
| Colorado | Donor | 51.6% | 35.3% | — | 51.4 | 390.8 |
| Florida | Donor | 56.5% | 35.3% | — | 45.9 | 265.1 |
| Georgia | Donor | 49.1% | 37.5% | — | 37.3 | 209.9 |
| Idaho | Donor | 46.2% | 30.5% | — | 69.0 | 274.4 |
| North Carolina | Donor | 47.6% | 35.4% | — | 45.8 | 212.8 |
| Nevada | Donor | 49.5% | 44.4% | — | 36.1 | 217.1 |
| Tennessee | Donor | 47.5% | 34.4% | — | 40.6 | 230.3 |
| Texas | Donor | 47.9% | 38.5% | — | 48.0 | 256.3 |
| Utah | Donor | 44.5% | 30.2% | — | 78.4 | 349.1 |
| Virginia | Donor | 48.5% | 34.3% | — | 29.2 | 235.0 |
Credibility Checks
Robustness batterySpecifications Completed by Outcome
Number of distinct robustness specifications run for each outcome. FHFA and QCEW include quarterly timing checks; permit and ACS outcomes are annual only.
Placebo Timing
Treatment shifted two years early. If the placebo produces significant post effects, the original estimate is less credible.
Donor Sensitivity
Western-only donor pool (AZ, CO, ID, NV, UT) and leave-one-donor-out (AZ, CO, FL each dropped).
Single-State Models
CA-only and OR-only event studies test whether pooled results are driven by one state.
State Interactions
Separate CA and OR event-time coefficients reveal heterogeneous responses.
Quarterly Timing
FHFA and QCEW under preferred (OR: 2019Q2) and alternative (OR: 2019Q1) treatment dates.
Pre-Trend Tests
Lead coefficients for each outcome. Clean pre-trends mean treated and donor states were on parallel paths before policy.
Pre-trend summary: Permits and FHFA have the cleanest pre-trends. Median gross rent has strongly negative leads at event times −4 and −3 (pre-treatment divergence). QCEW employment has a borderline negative lead at −4. ACS outcomes are compressed by the narrow 2015–2019 pre-window.
Data Sources
All free, all federalCoverage Windows
Available data years by source. The ACS gap in 2020 is real: the Census Bureau did not release 1-year profiles that year due to pandemic-related data-collection disruptions.
| Source | Agency | What it provides | Frequency | Coverage |
|---|---|---|---|---|
| Building Permits Survey (BPS) | Census Bureau | New privately-owned housing units authorized by building permits, broken down by structure size | Annual | 2010–2024 |
| House Price Index (HPI) | FHFA | Purchase-only, seasonally adjusted index of repeat-sale home prices | Annual + Quarterly | 2010–2025 |
| ACS 1-Year Profiles | Census Bureau | Median gross rent, rent burden, renter share, residential mobility (3 measures) | Annual | 2015–19, 21–24 |
| QCEW Area Slices | BLS | Total covered employment and average weekly wages for workers covered by unemployment insurance | Annual + Quarterly | 2014–2024 |
Methods
Specification
Two-way fixed-effects (TWFE) event study at the state-year level. State fixed effects absorb permanent state differences. Year fixed effects absorb national trends. Event-time leads and lags identify the treatment path.
Treatment Timing
Oregon: annual 2019, quarterly 2019Q2 (preferred) / 2019Q1 (alt).
California: annual 2020, quarterly 2020Q1.
Washington: descriptive only (2025).
ACS 1-year profiles were not released for 2020. QCEW starts 2014. Only two treated states. ACS measures occupied rents, not asking rents. All estimates are state-level. Washington is descriptive only. This project does not claim to measure “rent control” as a general policy category.