Rotterdam's investor rule shifted buyers but did not show cheaper homes

The federal experiment had not begun at the July 10 cutoff. AP reported that Trump would let H.R. 6644 become law without signing it, while GovInfo still labeled the measure an enrolled billenrolled billAn enrolled bill is the final text passed by both chambers and prepared for presentation to the president, but it is not yet necessarily a law.. Even after conditional enactmentenactmentEnactment is the moment a bill becomes law through a presidential signature, a veto override, or expiration of the action period while Congress can receive a veto. on July 11, Title XTitle XTitle X is the tenth major section of H.R. 6644 and contains the institutional-investor rules discussed here.'s purchase prohibition and enforcement would wait 180 days. [AP](https://apnews.com/article/trump-housing-bill-77ec340dcdd676c46c458813b461b1af) [GovInfo status](https://www.govinfo.gov/app/details/BILLS-119hr6644enr) [Enrolled Title X](https://www.govinfo.gov/content/pkg/BILLS-119hr6644enr/html/BILLS-119hr6644enr.htm)

Rotterdam supplies the clearest directly observed test in the sources reviewed here, but its answer separates two outcomes that are often blended together. The rule changed the share of eligible transactions won by different kinds of buyers. Its first-year sale-price estimate did not show that homes became cheaper, and it also could not establish that the true average price effect was exactly zero. [Rotterdam working paper, Tables 3 and 5](https://www.eur.nl/en/media/2025-07-paper-m-korevaar-buy-live-buy-let-1107251)

Changing who wins a home is not the same as changing what the home costs.

Think of a sale as an auction. Removing one class of bidder lowers the winning price only if no replacement bidder is willing and able to offer nearly as much. Rotterdam's estimates suggest replacement owner-occupierowner-occupierAn owner-occupier owns a home and lives in it. demand was substantial, which can change the name on the deed without producing a lower sale price. This is an interpretation of the paper's buyer and price results, not a separately measured mechanism. [Rotterdam working paper](https://www.eur.nl/en/media/2025-07-paper-m-korevaar-buy-live-buy-let-1107251)

Rotterdam's purchase protectionpurchase protectionRotterdam's purchase-protection policy is a permit-based restriction on renting a newly bought qualifying home during its first four years after transfer. took effect on January 1, 2022. In 16 of the city's 71 neighborhoods, a newly transferred qualifying home could not be rented without a permit for four years. The 2022 cutoff was a WOZ tax appraisal strictly below EUR355,000, not a sale-price ceiling, and the working paper says more than 90% of homes in the selected neighborhoods fell below it. [Official 2021 ordinance](https://zoek.officielebekendmakingen.nl/gmb-2021-484369.html) [Rotterdam working paper, Sections 2.2 and 2.3](https://www.eur.nl/en/media/2025-07-paper-m-korevaar-buy-live-buy-let-1107251)

The legal mechanism matters. An investor was not forbidden to buy, resell, occupy, or hold a covered home; the rule made a new buy-to-letbuy-to-letBuy-to-let means purchasing a home in order to rent it to someone else. conversion unattractive by restricting rental use. Permits covered first- or second-degree relatives, one non-tourist letting of at most 12 months after at least 12 months of owner occupancy, inseparable business premises, and several local social-purpose cases. [Official 2021 ordinance](https://zoek.officielebekendmakingen.nl/gmb-2021-484369.html)

That is already a larger practical reach than Title X in one dimension: Rotterdam imposed no investor portfolio threshold. Yet it was narrower in other dimensions because it applied only to qualifying existing homes in selected local areas and below a local appraisal cap. [Rotterdam ordinance](https://zoek.officielebekendmakingen.nl/gmb-2021-484369.html) [Enrolled Title X](https://www.govinfo.gov/content/pkg/BILLS-119hr6644enr/html/BILLS-119hr6644enr.htm)

The study is a matched difference-in-differencesdifference-in-differencesDifference-in-differences subtracts the before-and-after change in a comparison group from the before-and-after change in the group affected by a policy. analysis, not a randomized trial. The researchers predicted investor activity from 2019-2021 transactions, removed neighborhoods without comparable overlap, and retained 12 treated neighborhoods and 29 controls. They then compared the before-and-after change for eligible homes in the treated areas with the simultaneous change in the controls. [Working paper, Sections 3.2 and 3.3 and Table 2](https://www.eur.nl/en/media/2025-07-paper-m-korevaar-buy-live-buy-let-1107251)

In plain language, the method subtracts the movement both groups shared, such as a citywide housing-market shift. A causal reading still requires the groups to have followed parallel paths without the policy, no different shock to have hit only the treated areas, and spillovers into controls to be limited. The paper reports no statistically detectable pre-policy trend differences in its event-study checks, which supports but cannot prove those assumptions. [Working paper, Figure 2](https://www.eur.nl/en/media/2025-07-paper-m-korevaar-buy-live-buy-let-1107251) [World Bank difference-in-differences guide](https://dimewiki.worldbank.org/Difference-in-Differences)

Buyer labels also need care. The paper calls a buyer an owner-occupier if someone registered at the property within two years; any other buyer is classified as an investor. A first-time buyerfirst-time buyerIn the Rotterdam study, a first-time buyer is an owner-occupier with no earlier home purchase in the linked records. is an owner-occupier with no earlier purchase in the linked records. Those research definitions do not match Title X's legal definition of a large institutional investorlarge institutional investorUnder Title X, this is a specified for-profit entity that is in the housing-investment business and directly or indirectly controls at least the statutory number of covered homes.. [Working paper, Section 4.1](https://www.eur.nl/en/media/2025-07-paper-m-korevaar-buy-live-buy-let-1107251) [Enrolled Title X](https://www.govinfo.gov/content/pkg/BILLS-119hr6644enr/html/BILLS-119hr6644enr.htm)

Across 4,182 transactions in the buyer model, the fully adjusted estimate put the investor-buyer share 16.6 percentage points lower and the first-time-buyer share 9.3 points higher. Their approximate 95% intervals exclude zero. Because every sale is classified as an investor, first-time owner, or repeat owner, the remaining +7.3 points are an arithmetic residual for repeat buyers. The paper did not estimate that residual separately, so it has no published standard error or confidence interval. [Working paper, Table 3](https://www.eur.nl/en/media/2025-07-paper-m-korevaar-buy-live-buy-let-1107251)

A broader appendix using participating Dutch municipalities points in the same direction: its treatment-weighted adjusted estimates were -13.0 points for investor buyers and +9.0 points for first-time buyers. This is a robustness extension by the same authors using related Dutch data, not an independent replication, and the paper contains inconsistent counts of participating municipalities. [Working paper, Appendix Table A.2](https://www.eur.nl/en/media/2025-07-paper-m-korevaar-buy-live-buy-let-1107251)

The most adjusted pooled price coefficient was 0.0143 with a standard errorstandard errorA standard error measures how much an estimated result would be expected to vary across comparable samples under the model. of 0.0096. Converting the log-price coefficientlog-price coefficientA log-price coefficient expresses an estimated change in the logarithm of price and must be converted to describe the corresponding percentage change. gives a +1.4% point estimatepoint estimateA point estimate is the model's single best numerical estimate of an effect.. Cumulant's normal-approximation 95% interval runs from -0.5% to +3.4%, so zero remains inside the range. [Working paper, Table 5](https://www.eur.nl/en/media/2025-07-paper-m-korevaar-buy-live-buy-let-1107251) [NIST confidence-interval guide](https://www.itl.nist.gov/div898/handbook/eda/section3/eda352.htm)

The disciplined reading is that the model found no statistically distinguishable pooled averagepooled averageA pooled average combines all post-policy observations in the stated window into one overall estimated effect. sale-price effect. Under the study's assumptions, the data are inconsistent with a short-run average decline much larger than about 0.5%, but they remain compatible with a small decline, no change, or an increase as large as about 3.4%. A confidence interval describes values compatible with the model and data; it is not proof that every value inside it is equally likely. [Working paper, Table 5](https://www.eur.nl/en/media/2025-07-paper-m-korevaar-buy-live-buy-let-1107251) [NIST](https://www.itl.nist.gov/div898/handbook/eda/section3/eda352.htm)

The price regression contains 3,967 transactions across the whole 2021 Q2-2022 Q4 window, not 3,967 post-policy sales. That window contains three quarters before the January 2022 start and four quarters after it. The paper's method and table notes describe that full year after the rule, although one nearby sentence incorrectly says three post-policy quarters. [Working paper, Sections 3.3 and 4.3 and Table 5](https://www.eur.nl/en/media/2025-07-paper-m-korevaar-buy-live-buy-let-1107251)

Quarter-specific estimates add texture: the authors describe a short-lived positive response of roughly 2% to 3% in the first two post-policy quarters, which later became statistically indistinguishable from zero. That is a local transaction-price pattern after implementation, not an announcement-day financial-market study, and it still offers no evidence of cheaper homes. [Working paper, Figure 4](https://www.eur.nl/en/media/2025-07-paper-m-korevaar-buy-live-buy-let-1107251)

Sale price is only one outcome. The paper does not estimate mortgage payments or rents, and the buyer-share increase does not show that households qualified for cheaper financing. The result supports better access to eligible purchases for first-time and repeat owner-buyers, not a general claim that housing became more affordable. [Working paper, Tables 3 and 5](https://www.eur.nl/en/media/2025-07-paper-m-korevaar-buy-live-buy-let-1107251)

Title X generally defines a large institutional investor as a specified for-profit entity in the housing-investment business that, alone or in concert with other entities after enactment, directly or indirectly controls at least 350 statutory single-family homes. Homes acquired through post-enactment excepted purchases are excluded from that threshold calculation. The statutory home can contain one or two units, but cannot be a manufactured homemanufactured homeA manufactured home is factory-built housing covered by a separate federal definition, and Title X excludes it from the homes subject to the purchase rule.. [Enrolled Title X, definitions](https://www.govinfo.gov/content/pkg/BILLS-119hr6644enr/html/BILLS-119hr6644enr.htm)

The prohibition would block future nonexempt acquisitions, not require existing portfolios to be sold. Exceptions cover build-to-rentbuild-to-rentBuild-to-rent means constructing homes to keep and operate as rentals rather than selling them to individual occupants., qualifying renovation, certain creditor and foreclosure transactions, homeownership programs, certain compliant transfers from other large investors, communities for households with a member aged 55 or older, and other listed cases. One temporary path permits purchases from investors not covered by the section for up to two years after the effective dateeffective dateThe effective date is when a legal requirement starts operating, which can be later than the date the law is enacted.. [Enrolled Title X, subsections 1001(a) and 1001(b)](https://www.govinfo.gov/content/pkg/BILLS-119hr6644enr/html/BILLS-119hr6644enr.htm)

The comparison supports one mechanism-level inference: when a rule actually removes investor demand from a set of homes, owner-occupiers can replace at least some of those bids. It does not supply a U.S. effect size. The Rotterdam paper says large commercial investors were not active on the buy-to-let margin it studied, while Title X is built around large controlled portfolios. Different homes, buyers, financing systems, exceptions, and geographies could produce a different substitution rate and a different price response. [Rotterdam working paper, Section 2.1](https://www.eur.nl/en/media/2025-07-paper-m-korevaar-buy-live-buy-let-1107251) [Enrolled Title X](https://www.govinfo.gov/content/pkg/BILLS-119hr6644enr/html/BILLS-119hr6644enr.htm)

The enrolled text delays the purchase prohibition and enforcement until 180 days after enactment. If H.R. 6644 became law without a signature on July 11, those provisions would begin January 7, 2027. The temporary seller exception and the first statutory GAOGAOGAO is the U.S. Government Accountability Office, an independent congressional audit and research agency. and HUDHUDHUD is the U.S. Department of Housing and Urban Development. study deadlines reach their two-year point on January 7, 2029; the purchase prohibition and enforcement would repeal 15 years after they start. [Enrolled Title X](https://www.govinfo.gov/content/pkg/BILLS-119hr6644enr/html/BILLS-119hr6644enr.htm) [Congressional procedure](https://www.congress.gov/legislative-process/presidential-action)

A credible evaluation should first count which attempted purchases were actually barred and which exception applied. It should then compare changes in covered-entity purchases, first-time and repeat owner-buyer shares, sale prices, rents, transaction volume, new construction, and rehabilitation across markets with different pre-policy exposure. The comparison should separate the pre-effective anticipation period from the period after January 7, 2027. This is Cumulant's measurement recommendation based on the enrolled rule and the Rotterdam study's identification logic. [Enrolled Title X](https://www.govinfo.gov/content/pkg/BILLS-119hr6644enr/html/BILLS-119hr6644enr.htm) [Rotterdam working paper](https://www.eur.nl/en/media/2025-07-paper-m-korevaar-buy-live-buy-let-1107251)

Rotterdam's best-supported lesson is therefore narrow but useful. The rule shifted who acquired eligible homes, with first-time and repeat owner-buyers replacing investor buyers. Its first-year price evidence did not show cheaper homes. For Title X, expanded access is a hypothesis worth testing; cheaper housing is a separate result that must be measured, not assumed. [Rotterdam working paper, Tables 3 and 5](https://www.eur.nl/en/media/2025-07-paper-m-korevaar-buy-live-buy-let-1107251) [Enrolled Title X](https://www.govinfo.gov/content/pkg/BILLS-119hr6644enr/html/BILLS-119hr6644enr.htm)