Institutional landlords hold about 3% of US single-family rentals but 25% of Atlanta's, and one rent algorithm added $3.8 billion to a year of rent. Who pays.
The Largest Asset Class on Earth
How shelter became collateral
The world's real estate was worth $393.3 trillion at the start of 2025, roughly four times global GDP, and $286.9 trillion of it was housing ✓ Established [1]. Housing is not merely the biggest store of wealth in existence. Over three decades it has become the biggest financial asset class, one whose price is set by the return it can be made to yield rather than by what local incomes can carry [2] [3]. This report follows that shift from the foreclosure auction to the bond market and the algorithm, and asks a narrow question: who pays when homes become an asset class.
Begin with the scale. Savills values the planet's residential stock at $286.9 trillion at the end of 2024, more than all commercial property, farmland, listed equities and government debt combined ✓ Established [1]. A quarter of that value sits in China, and falling Chinese prices were enough to pull the global total down by 0.5% in a year even as most other countries rose [1]. No other asset comes close to this weight, and no other asset is simultaneously a human necessity. Housing is the one market where the balance sheet and the bedroom are the same object.
Financialisation, in the definition used by the geographer Manuel Aalbers, is the increasing dominance of financial actors, markets, practices, measurements and narratives, at various scales, resulting in a structural transformation of economies, firms, states and households [2]. Applied to housing, it describes a change in what a home is for: a dwelling once priced as a consumption good whose cost tracked wages is now priced as a yield-bearing asset whose value tracks interest rates, expected rent growth and the global supply of capital seeking collateral. Aalbers describes a wall of money searching for high-quality collateral and finding that housing is one of the few asset classes large and durable enough to absorb it ◈ Strong Evidence [2].
The divergence between wages and prices is the fingerprint of that change. Economists at the Federal Reserve Bank of St. Louis calculate that between 2000 and 2024 median per-capita income in the United States grew about 155% in nominal terms while median home prices rose around 207%, and a repeat-sales index that strips out compositional change still shows cumulative growth of 171% ✓ Established [3]. Most US counties moved from house prices worth roughly three to five times local income in 2000 to substantially higher multiples by 2023, and the average age at which Americans buy a first home rose by about ten years [3]. In the authors' words, the typical home has simply outrun the typical paycheck.
The pattern is not American. Across the OECD, real house prices rose by almost 60 index points on average over three decades, climbing gradually after the 2007 to 2008 financial crisis and then accelerating during the pandemic ✓ Established [4]. The OECD's price-to-income ratio averaged 114.7 index points across member countries in 2025 against a 2015 base of 100, and Portugal, the Netherlands and Canada all exceeded 130 [4]. In each of those countries the same explanations are offered: too little building, too much demand, interest rates too low for too long. All three are true. None explains why the buyer at the margin is now so often a fund rather than a family.
Savills puts total global real estate at $393.3 trillion at the start of 2025, about four times world GDP, with residential property accounting for $286.9 trillion [1]. That figure is the denominator for everything that follows. Even a fund owning 100,000 homes controls a rounding error of the stock; the significance of financialisation lies in who sets the marginal price, not who owns the median house [7].
The costs land on tenants first. Harvard's Joint Center for Housing Studies counts a record 22.7 million renter households, 49% of all renters, paying more than 30% of income for housing and utilities in 2024 ✓ Established [5]. Eleven million extremely low-income households compete for 3.8 million rental homes that are both affordable and available to them. Between 2014 and 2024 the stock of units renting for less than $1,000 a month shrank by more than 30%, a loss of 7 million homes [5]. These are the outcomes of a market that has done exactly what an asset market is supposed to do: reprice toward the return investors demand.
Housing has lost its social function and is seen instead as a vehicle for wealth and asset growth.
— Leilani Farha, UN Special Rapporteur on the right to adequate housing, presenting report A/HRC/34/51 to the Human Rights Council, March 2017When Leilani Farha presented her report on the financialisation of housing to the UN Human Rights Council in March 2017, she framed the problem as one of function rather than ownership ◈ Strong Evidence [6]. Two years later she and the UN Working Group on Business and Human Rights wrote to Blackstone's chief executive, stating that unprecedented amounts of global capital were being invested in housing as security for financial instruments and that the firm had used significant resources and political leverage to undermine domestic laws and policies [45]. Blackstone rejected the characterisation within three days [45]. The debate has since moved from human-rights bodies to antitrust divisions and, in July 2026, to the United States Code.
Four distinct phenomena are usually folded into the single word financialisation, and this report keeps them apart: the institutional single-family landlord, a post-2008 invention that owns several hundred thousand American houses; build-to-rent, the construction of neighbourhoods designed never to be sold; the listed or private-equity apartment owner, from Vonovia in Germany to Blackstone in Spain, which is older and larger; and rent-setting software, which requires no ownership at all to move prices. What unites them is a statistic that both sides of the debate cite and neither fully absorbs: institutional investors own about 3% of America's single-family rentals nationally, and about 25% of them in metropolitan Atlanta ✓ Established [8]. Both numbers are correct. The argument is about which one matters.
From Foreclosure to Bond Coupon
The mechanism of the single-family asset class
In 2011 no single investor in the United States owned more than 1,000 single-family homes. By 2022, 32 investors owned about 450,000 between them ✓ Established [8]. The asset class was assembled in a decade, out of the wreckage of the previous one, and the tool that made it possible was not the purchase of houses but the securitisation of their rent.
The single-family rental industry was born in the foreclosure crisis. Between 2007 and 2011 millions of American homes passed from defaulted borrowers to banks and agencies, and the buyers with the cash and speed to absorb them were funds, not households [7]. Blackstone began purchasing in 2012 through a new subsidiary, Invitation Homes, and was soon buying thousands of homes a month in Atlanta, Phoenix, Tampa and Southern California ✓ Established [45]. Institutions could pay cash for distressed properties, clear title, renovate and re-let in a fraction of the time available to a mortgage-dependent purchaser, and research reviewed by the Government Accountability Office finds that their purchases helped stabilise prices in the hardest-hit metros [7]. The rescue came with a permanent change of ownership.
The decisive innovation came in November 2013, when Invitation Homes bundled 3,207 of its rental houses as collateral for a $479 million bond, the industry's first single-family rental securitisation ✓ Established [10]. Investors in the notes were buying a share of future rent, exactly as investors in a mortgage-backed security buy a share of future mortgage payments, and the deal set a template competitors copied within months [10]. From that point the economics of the business were fixed by the bond market: a portfolio was worth the capitalised value of its rent roll, and every decision from renewal pricing to maintenance spending fed the yield the notes were priced against. Real estate investment trusts completed the structure: by distributing at least 90% of taxable income, REITs avoid corporate tax and are obliged to maximise distributable income [7].
The resulting firms are large by any standard except the size of the housing stock. Invitation Homes owns about 86,000 homes and American Homes 4 Rent about 61,000; among private asset managers Pretium Partners owns and manages about 82,000, Blackstone about 63,600 and Amherst Group about 50,000 ✓ Established [7]. Invitation Homes reported average monthly rent per occupied home of $2,439 in 2025, up 2.2%, with average occupancy of 95.0% [11]. The composition of that growth is instructive: same-store renewal rents rose 4.6% while new-lease rents fell 0.6% [11]. Existing tenants, who face moving costs, paid more; new tenants, who could shop, paid slightly less. That gap is the market power of the incumbent landlord expressed in one line of an earnings release.
The Government Accountability Office reports that no single investor owned more than 1,000 single-family homes in 2011, and that by 2022 more than 30 investors each held over 1,000, for a combined 450,000, with the five largest holding nearly 300,000 [8]. Those holdings are concentrated: institutions are estimated to own 25% of single-family rentals in metropolitan Atlanta, 21% in Jacksonville, 18% in Charlotte and 15% in Tampa, the markets foreclosures hit hardest after 2008 [8].
Why a fund wants to own 80,000 houses when the return on any one is modest has a one-word answer: cost. Joshua Coven, using Census Bureau cost data for small landlords and earnings supplements for the listed REITs, finds that institutional landlords operate at lower average costs and scale more efficiently than the individuals who traditionally owned single-family rentals ◈ Strong Evidence [13]. Their portfolios are spatially concentrated, up to 85,000 homes against the one to three of a typical small landlord, so maintenance, leasing and procurement are amortised across dense clusters [13]. Coven also documents a channel of market power specific to real estate: large landlords adjust the number of units they hold in a market by as much as 14.7% in a year, in a pattern consistent with profit maximisation, rather than adjusting occupancy [13].
Geography is therefore not incidental to the model but constitutive of it. The GAO's follow-up analysis of six metropolitan areas finds that between 2018 and 2024 institutional investors added at least 16,000 homes each in Phoenix and Dallas, increases of 177% and 114%, at least 8,000 each in Jacksonville and Nashville, increases of more than 145%, about 3,000 in Cincinnati and fewer than 2,000 in Seattle ✓ Established [9]. In some suburban ZIP codes around Atlanta, Phoenix and Tampa institutional investors have bought up to 8.5% of the entire housing stock [13]. A national ownership share of 3% is compatible with a local share high enough to set prices, and both descriptions of the same industry are accurate [8].
Once rent is the collateral for a bond, the house is priced backwards from the coupon. A portfolio's value is its net operating income divided by the yield investors will accept, and every practice that lifts income or cuts cost, from renewal pricing to automated fees to deferred maintenance, flows straight into valuation. The conduct findings in later sections are therefore not aberrations. They are the financial logic of the securitised rental working as designed, whether the owner holds 86,000 homes or 350.
The second phenomenon, build-to-rent, is where the industry has moved its growth. Single-family homes built to be held as rentals reached 84,000 starts in the United States in 2024 before falling 19% to 68,000 in 2025 as financing costs rose and a multifamily glut competed for tenants ✓ Established [12]. Even after that decline, built-for-rent is about 7% of single-family starts, against 2.7% between 1992 and 2012, excluding homes sold to investors after completion, which the National Association of Home Builders puts at a further 3% to 5% [12]. John Burns Research tracks about 500,000 build-to-rent units with 160,000 more in the pipeline [43]. Build-to-rent is the version of financialisation that adds supply, and the version legislators have been most careful to exempt.
What the Studies Found
Rents, prices, homeownership and the scale trade-off
The peer-reviewed evidence on institutional landlords is more precise, and less comfortable for both camps, than the public debate. Institutional entry raised local house prices, lowered homeownership, expanded rental supply and, where firms consolidated, raised rents ◈ Strong Evidence [13] [14]. Those findings come from different studies and they are not in conflict. They describe a single trade-off.
The clearest evidence of pricing power comes from mergers. Umit Gurun and co-authors, writing in the Review of Financial Studies in January 2023, use mergers between institutional single-family landlords to identify neighbourhoods where the two merging firms both owned homes and therefore gained local market share overnight ◈ Strong Evidence [14]. Rents rose in those overlapped neighbourhoods relative to comparable ones where only one firm was present, and crime fell significantly, which the authors read as institutions using scale to improve the rental service, notably safety, while extracting more of the surplus from tenants [14]. The Federal Reserve Bank of St. Louis summarises the finding as institutions gaining a degree of price-setting power in the areas they dominate after consolidation [46].
Coven's structural model, the most complete attempt to net out the competing effects, fits neither side's talking points ◈ Strong Evidence [13]. For every home an institutional investor purchased, local rental supply rose by 0.5 homes, because institutions convert owner-occupied houses into rentals and run them cheaply enough to keep them let. For every home purchased, the stock available for owner-occupation fell by 0.22 homes, because construction and sales by small landlords offset most but not all of the demand shock. In the top decile of markets by institutional entry, that entry explains about 20% of the house-price increase since 2012 [13]. Renters on net gained lower rents, prospective buyers lost, and the driver was economies of scale rather than market power. Coven concludes that a purchase ban or a rent cap on institutions would raise rents by shrinking rental supply [13].
A separate strand of research finds that the price effects usually attributed to Wall Street belong mostly to Main Street. Carlos Garriga, Pedro Gete and Athena Tsouderou, in Real Estate Economics, show that small and medium-sized investors, largely local operators with about two-thirds of their portfolios in one metropolitan area, increased house-price growth and the price-to-income ratio and shifted construction toward multifamily units ◈ Strong Evidence [15]. Those investors bought 18% to 24% of single-family homes in every year of the past two decades, while institutional purchases never exceeded 3% even at the 2022 peak [41]. The Federal Reserve Bank of Philadelphia, examining the listed single-family REITs directly, finds their effect on local price growth modest [7].
The studies converge on an uncomfortable point: institutional landlords are cheaper per unit than the landlords they replace, and cheapness at scale is what lets them outbid families for a house while offering renters a slightly lower rent for it. The harm to buyers and the benefit to renters are the same phenomenon seen from different sides of the lease. A policy that removes the buyer without replacing the supply transfers the benefit from one group to the other rather than creating it.
Market fundamentals complicate every attribution. The GAO's March 2026 study of six metropolitan areas chosen for heavy institutional activity finds that all six experienced population growth between 2018 and 2024, accompanied by rising median rents, rising counts of single-family units and rising counts of rental units of all types ✓ Established [7]. Homeownership rates increased in four of the six metros and fell only where population growth was weaker [9]. The CRS observes that rent increases attributable to investor pricing power would be more visible if population had been stagnant and rental supply had fallen; neither was the case, so demand growth and investor behaviour are entangled in the same data, and the GAO study makes no formal attempt to separate them [7].
On tenant outcomes the evidence is older but consistent. Elora Raymond and colleagues at the Federal Reserve Bank of Atlanta examined eviction filings in Fulton County, Georgia, in 2015 and found that 22% of all renters faced an eviction proceeding that year ◈ Strong Evidence [16]. A filing was 8% more likely if the home was owned by a large corporate landlord with more than 15 properties, and some institutional owners were 18% to 19% more likely to file than small landlords, controlling for property and neighbourhood characteristics [16]. Filing is not removal, and research on outcomes remains thin [7]. But a filing is itself a financial event for a tenant, generating fees and a court record, and at institutional scale it can be automated.
What remains genuinely contested is whether institutional landlords hold market power in the economic sense at all ⚖ Contested. Gurun's merger evidence says yes in specific overlapped neighbourhoods [14]. Coven's model says the dominant force is cost advantage and that market power, where present, is dampened by supply responses [13]. A March 2026 paper by Felipe Barbieri and Gregory Dobbels, cited by the CRS, finds that institutional landlords can raise rents through market power but that the increases are constrained by rental supply growth [7]. The positions reconcile: pricing power exists at the level of a ZIP code and dissolves at the level of a metro, and the geography of institutional ownership populates both scales.
The Algorithm
How rent-setting software coordinated landlords without a meeting
In October 2022 ProPublica reported that property managers using RealPage's YieldStar software accepted roughly 90% of its daily rent recommendations, built on non-public lease data from competing landlords ✓ Established [19]. The White House Council of Economic Advisers later estimated that algorithmic rent-setting added about $70 a month to affected leases and $3.8 billion to US rents in 2023 alone ◈ Strong Evidence [20]. Financialisation, it turned out, did not require owning the houses.
The fourth phenomenon is the one that most directly separates financialisation from ownership. RealPage, a property-management software company based in Richardson, Texas, sold revenue-management tools that ingested each client's non-public lease terms, rents and occupancy, pooled them and returned a recommended price for every unit every day [19]. ProPublica found that landlords accepted roughly nine in ten of those recommendations, that the software discouraged the concessions leasing agents had traditionally used to fill units, and that in some markets a majority of large apartment buildings were priced through the same system ✓ Established [19]. The company's own materials described an approach in which landlords could tolerate higher vacancy in exchange for higher rents [19].
The Council of Economic Advisers quantified the effect in December 2024. Its analysis estimated that rent-setting algorithms added an average of $70 a month to the cost of units in buildings that used them, or about $3.8 billion in additional rent across the United States in 2023 ◈ Strong Evidence [20]. In metropolitan Atlanta, where institutional single-family ownership is also highest, the estimated premium exceeded $180 a month [20]. The mechanism is the one antitrust law has always targeted, coordination among competitors on price, stripped of the meeting and the paper trail. Each landlord could truthfully say it had never spoken to a rival. The software had spoken for all of them.
The White House Council of Economic Advisers' December 2024 analysis attributes an average premium of $70 a month to units in buildings priced with commercial rent-setting software, roughly $3.8 billion of additional rent nationally in 2023 [20]. RealPage disputed the methodology, and the Justice Department's later settlement contains no admission of liability and no penalty [21]. The estimate is best read as an order of magnitude: a pricing layer with no ownership of housing moved rents by an amount comparable to the annual revenue of the largest single-family landlord.
The Justice Department and a coalition of states sued RealPage in August 2024, alleging that it used non-public, competitively sensitive information from rival landlords in its recommendations and discouraged independent pricing, in violation of Section 1 of the Sherman Act [21]. On 24 November 2025 the department filed a proposed settlement ✓ Established [21]. RealPage agreed to stop using competitors' non-public pricing information, to train models only on data at least twelve months old, to restrict geographic pricing models to the state level or broader, to remove features that limited price decreases or aligned prices, to end its market surveys, to accept a court-appointed monitor and to cooperate in the government's cases against the property managers that used the product [21]. There was no fine and no admission of wrongdoing.
Competing companies must make independent pricing decisions, and with the rise of algorithmic and artificial intelligence tools, we will remain at the forefront of vigorous antitrust enforcement.
— Abigail Slater, Assistant Attorney General, Antitrust Division, US Department of Justice, 24 November 2025The landlords followed. Greystar, the largest apartment manager in the United States, reached its own agreement with the Justice Department and a $7 million settlement with nine state attorneys general under which it agreed to stop using any software that relies on competitively sensitive information to align rents, while maintaining that its use of RealPage's tools had complied with the law ✓ Established [22]. In July 2026 plaintiffs' lawyers announced proposed class settlements of $359.9 million covering everyone who paid rent on a multifamily lease at a RealPage-licensed property between 18 October 2018 and 21 November 2025 [23]. Against the CEA's $3.8 billion estimate for a single year, the settlements recover roughly one dollar in ten of one year's premium.
The algorithm matters here because it shows that the ownership share, the number both defenders and critics of institutional landlords reach for first, is the wrong denominator for financialisation. RealPage owned no apartments. Its clients managed millions of units, and the product's value to them was that it converted a fragmented, competitive market into one that behaved as if it were concentrated [19]. The single-family REITs are large landlords in a few metros; the pricing software was a small company with a large footprint in every metro. Both express the same logic, that the rent roll is a financial instrument to be optimised, and the second required no capital at all.
The RealPage record resolves a question the ownership debate cannot: financialisation is a pricing regime before it is an ownership pattern. A market in which shared data lets a 1% owner set prices for the other 99% behaves like a cartel regardless of how the deeds are distributed. That is why the antitrust track has produced more measurable change in two years than a decade of hearings on institutional landlords, and why the RealPage remedies, twelve-month-old data, state-level models, no shared surveys, are the most transferable policy template this field has produced.
Whether the software raised rents above the level a competitive market would have produced remains formally contested ⚖ Contested. RealPage has argued that its recommendations were advisory, that landlords accepted them because they were accurate, and that rents rose in 2021 and 2022 for reasons of supply and demand no algorithm created [21]. Because the Justice Department settled without trial, no court has ruled on the merits, and the class settlements likewise resolve claims without findings. What is not contested is the remedy: RealPage, Greystar and the other settling landlords have agreed to stop doing the specific things the complaints described, the clearest available indication of how the parties themselves judged the litigation risk.
Who Pays
The renter's ledger
The costs of housing as an asset class are paid in rent, in fees, in deposits not returned, in eviction filings and in a first home that recedes faster than a deposit can be saved. Half of American renters were cost-burdened in 2024, and the enforcement record of the largest single-family landlord shows how the margins were made ✓ Established [5] [17].
Start with the aggregate. The Joint Center's count of 22.7 million cost-burdened renter households in 2024 is a record for the third consecutive year, and the burden is steepest where incomes are lowest ✓ Established [5]. The disappearance of 7 million units renting below $1,000 between 2014 and 2024 is the supply-side counterpart: the cheap end of the market has been renovated, repriced or demolished, and it is precisely the segment that value-add investment strategies target [5]. Residential mobility has fallen to a record low because a household that moves must re-enter the market at today's rent [5]. Immobility is the tenant's only hedge, and landlords price it, which is why renewal rents at Invitation Homes rose 4.6% in 2025 while new-lease rents fell [11].
Now the conduct. In September 2024 the Federal Trade Commission announced that Invitation Homes, which it called the nation's largest single-family landlord, would pay $48 million to settle allegations that it deceived renters about lease costs, charged undisclosed junk fees, failed to inspect homes before move-in, withheld security deposits unfairly and used unfair eviction practices ✓ Established [17]. The mandatory fees, for items such as smart-home technology and air filters, added more than $1,700 a year to advertised rents. Residents of 33,328 homes filed a maintenance work order within the first week after moving in between 2018 and 2023. Between 2020 and 2022 the company returned 39.2% of security deposit dollars against a national average of 63.9% [17]. In March 2026 the FTC mailed cheques totalling more than $47.2 million to affected tenants [47].
FTC chair Lina Khan said at the time that Invitation Homes had preyed on tenants through a variety of unfair and deceptive tactics [17]. The company admitted no wrongdoing and agreed to include all mandatory monthly fees in its advertised prices and to reform its deposit practices [17]. What the settlement documents is not villainy but optimisation. Each practice the FTC listed lifts net operating income by a small percentage, and net operating income, capitalised at the yield the bond market demands, is what the enterprise is worth. A landlord with three houses has no reason to design a smart-home fee. A landlord with 86,000 has a spreadsheet showing what one adds.
The pattern recurs at smaller scale. In March 2024 the Minnesota attorney general settled a 2022 lawsuit against HavenBrook Homes and six other defendants connected to Pretium Partners and Progress Residential, which the state accused of systematically under-maintaining more than 600 single-family rental homes in the Minneapolis area while falsely promising repair services ✓ Established [18]. The companies agreed to pay $2.2 million into a restitution fund and to forgive up to $1,987,015 of rental debt owed by former tenants, and Pretium and Progress disclosed plans to transfer the properties to affordable-housing entities [18]. Formal research comparing the maintenance practices of institutional and small landlords is largely nonexistent; what exists is enforcement records and the pattern within them [7].
The cost that appears in no settlement is paid by the household that never becomes an owner. The Federal Reserve Bank of New York's Survey of Consumer Expectations finds that 71.5% of renters in 2025 said they would prefer to own, essentially unchanged from 71.3% in 2019 ✓ Established [46]. Over the same six years the share who believed obtaining a mortgage would be difficult rose from 25.8% to 42.9%, and the share who expected ever to own fell from 52.6% to 33.9% [46]. Preference held constant while expectation collapsed. Coven's estimate that each institutional purchase removed 0.22 homes from the owner-occupied pool is one mechanism by which that expectation was foreclosed, and the decade added to the first-home age is its consequence [13] [3].
At the bottom of the ledger is homelessness. The Department of Housing and Urban Development counted 745,652 people experiencing homelessness on a single night in January 2025, 3% fewer than in 2024 and the first year-on-year decline since 2016, but still close to the record ✓ Established [24]. The number of individuals not in families rose 0.6% to the highest ever recorded, and 266,320 people were unsheltered, 27% more than in 2013 [24]. The Joint Center draws the connection explicitly: the shortage of affordable rentals feeds homelessness [5]. No study attributes homelessness to institutional landlords, and none should. The link is indirect, through the loss of the cheap units that financialised strategies reprice.
Nine Countries, One Logic
How the asset class travels
The same capital meets different housing systems and produces different symptoms: single-family portfolios in the United States, build-to-rent towers in Manchester, a listed landlord with 540,000 flats in Germany, foreign buyers in Tokyo's central wards ✓ Established [26] [35]. The policy responses are as varied as the symptoms, and the results are beginning to come in.
The United States is the only large economy in which the single-family house became an institutional asset class, for reasons of history rather than design: no other country produced millions of detached foreclosures in a four-year window and then sold them in bulk [7]. Elsewhere financialisation arrived through the apartment block, the pension fund and the foreign buyer. What travels is the logic, that housing is priced from its yield; what differs is the regulatory inheritance the logic meets. The comparison below runs from the market that most resembles the American case to the one that least does.
The United Kingdom is building the American model rather than buying it. Build-to-rent stock passed 139,000 completed homes in the third quarter of 2025, up 14% on the year, with 52,500 under construction and a further 106,500 in the planning pipeline, and investors committed 2.6 billion pounds in the first nine months of the year, matching 2023 and 2024 ✓ Established [25]. The sector is still only 2% of the private rented sector nationally, but in Manchester its share has reached a fifth [25]. Because British build-to-rent is almost entirely new construction, it has attracted little of the hostility directed at American single-family buyers; the political question in Britain is not who owns the homes but why so few are built.
Germany is the oldest and largest case of listed residential ownership, and the site of the most radical proposed remedy. Vonovia owns about 540,000 flats across Germany, Sweden and Austria; its German in-place rent reached 8.19 euros per square metre at the end of 2025 against 7.89 a year earlier, and its rental revenue rose 2.8% to 3,417.2 million euros ✓ Established [26]. In September 2021, 57.6% of Berlin voters backed expropriating landlords with more than 3,000 units; four years later the campaign published a draft socialisation law covering about 220,000 flats, with a binding vote not expected before 2027 [28]. In July 2026 the federal coalition of CDU, CSU and SPD announced legislation to prohibit the states from socialising private rental housing at all [27]. The federal rent brake has meanwhile been extended to the end of 2029 [48].
The Netherlands ran the cleanest natural experiment. A 2022 national law allowed municipalities to bar investors from buying homes below a price threshold in order to rent them out, and Rotterdam applied it first ✓ Established [29]. Economists at Erasmus University Rotterdam and the University of Amsterdam found that investor purchases in the regulated neighbourhoods fell by 73%, that the owner-occupied share rose by one percentage point in 2022, and that prices did not fall but rose slightly in the two quarters after the ban [29]. The replacement buyers were more affluent than the tenants who would otherwise have rented the same homes, more often native-born Dutch, and stayed longer [29]. The ban delivered what it promised, more owner-occupation, and what its critics predicted, less rental housing for the lower-income and more mobile households who had used it.
Spain and Ireland illustrate the difference between announcing a policy and enacting one. In January 2025 Spain's prime minister proposed a tax of up to 100% on home purchases by non-resident buyers from outside the European Union; by March 2026 it had never been debated in parliament and the government's own January 2026 housing package had dropped it ✓ Established [30]. Catalonia's rent caps, in force since 2024, produced an ambiguous result: average rents still rose 2.5% in the third quarter of 2025, new contracts fell to 26,962 from 34,495 in the first quarter of 2024, and listings of seasonal lets exempt from the cap jumped 45% [31]. Ireland raised the stamp duty on bulk purchases of ten or more houses in a year from 10% to 15% in October 2024, a tax on the American model at the point of entry [32].
Canada and Australia chose the foreign buyer as the target. Canada's Prohibition on the Purchase of Residential Property by Non-Canadians Act, which bars foreign commercial enterprises and non-residents from buying homes, was extended to 1 January 2027 ✓ Established [33]. Australia banned foreign persons, including temporary residents and foreign-owned companies, from buying established dwellings from 1 April 2025, and in its 2026 to 2027 budget extended the ban by two years and three months to 30 June 2029, with exemptions for investment that adds supply [34]. Canada sits above 130 on the OECD's price-to-income index [4]. Neither ban addresses domestic institutional capital, which in both countries flows chiefly into purpose-built rental and is welcomed for doing so.
Japan is the outlier that proves the rule. Average prices for new condominiums in Tokyo's 23 wards reached a record 137.84 million yen in fiscal 2025, up 18.5% in a year, while the wider metropolitan average rose 15.3% to 93.83 million yen ✓ Established [35]. A Mitsubishi UFJ Trust survey found foreign buyers accounted for 19.0% of purchases in Chiyoda, Minato and Shibuya in the first half of 2025 against 12.7% in the rest of the 23 wards [36]. Japan has almost no restrictions on foreign ownership and no institutional single-family sector, yet its capital displays the same symptom as Atlanta and Amsterdam: a marginal buyer who prices the asset for its yield in a foreign currency rather than for its use in the local one. Financialisation does not need a Wall Street landlord. It needs a yield differential and an open door.
The Response
A ban, an order and a stamp duty
On 11 July 2026 the 21st Century ROAD to Housing Act became law in the United States without a presidential signature, the first broad federal restriction on institutional ownership of single-family homes in the country's history ✓ Established [38]. From 7 January 2027 any for-profit entity controlling 350 or more single-family homes is barred from buying another, subject to exemptions the industry spent the spring negotiating [39]. Whether it will lower a single price is the question the previous sections were written to answer.
The federal response came in two steps. On 20 January 2026 Executive Order 14376, titled Stopping Wall Street From Competing With Main Street Homebuyers, directed federal housing agencies to stop approving, insuring, guaranteeing, securitising or otherwise facilitating sales of single-family homes to large institutional investors, to adopt first-look policies giving owner-occupiers priority on foreclosed properties, and instructed the Treasury to define a large institutional investor within 30 days ✓ Established [37]. The order worked at the margin: federal programmes already refuse mortgage insurance to investors, and the government-sponsored enterprises had ended their single-family rental pilot in 2018 [7]. Its significance was political. A Republican administration had adopted the framing Democratic senators had used since 2021.
Congress then legislated. The Senate passed the 21st Century ROAD to Housing Act by 85 votes to 5 on 22 June 2026 and the House by 358 to 32 the next day; it became law on 11 July when the President neither signed nor vetoed it within ten days ✓ Established [38]. Section 1001, titled Homes Are for People, Not Corporations, defines a large institutional investor as any for-profit entity that alone or in concert controls 350 or more single-family homes of one or two units, and prohibits such entities from purchasing any further single-family home except under a statutory exemption [39]. Violations carry a civil penalty of the greater of $1 million or three times the purchase price. The prohibition takes effect on 7 January 2027 and sunsets fifteen years later [39].
The 350-home threshold, the one-or-two-unit definition, the penalty formula and the fifteen-year sunset are set out in Section 1001 of H.R. 6644 and summarised by the Congressional Research Service and the law firms advising the sector [7] [38] [39]. John Burns Research estimates that investors above the threshold own about 0.7% of America's 92 million single-family homes, about 5% of its 14 million single-family rentals, and bought about 1% of the 4.7 million homes sold in 2025 [43].
The exemptions define the law as much as the prohibition does, and they track the distinction this report has drawn between buying existing homes and adding new ones. Purchases in the following categories remain lawful for covered investors [39] [7]:
- Homes newly built, renovated or converted for resale to individual buyers rather than held as rentals.
- Build-to-rent homes, newly constructed and retained as managed rentals, with no obligation ever to sell them.
- Substantially rehabilitated homes where the investor spends at least 15% of the purchase price on improvements before renting.
- Homes in rent-to-own or homeownership programmes offering market-comparable rent, positive rent reporting to credit bureaus and a purchase right for the tenant.
- Homes acquired through foreclosure, deed in lieu or other satisfaction of a debt, and newly built rentals in communities for residents aged 55 and over.
- Homes bought from another covered investor at any time, and homes bought from non-covered sellers during a two-year transition window ending 7 January 2029.
The most consequential change happened between the versions. The bill the Senate passed by 89 votes to 10 in March 2026 required homes acquired under several exemptions, including build-to-rent, to be sold to individual buyers within seven years ✓ Established [7] [40]. Industry groups argued that forced sales would make purpose-built rentals, in John Burns Research's phrase, largely unbuildable and uninvestable, and the Terner Center at Berkeley warned the provision would reduce construction of new single-family rentals [43] [7]. The House removed the clause in May and the enacted version requires no sales at all [39]. The law therefore freezes existing portfolios, channels new institutional capital into construction and leaves the 450,000 homes already held where they are, while requiring covered investors to report holdings to HUD annually and to give tenants a dispute-resolution channel [7].
The market had begun adjusting before the law existed. Parcl Labs data compiled by ResiClub show that the eight largest institutional single-family landlords were net sellers of 3,011 homes in the second quarter of 2026, up from 593 a year earlier, an increase of 408% ✓ Established [40]. VineBrook Homes had 1,900 of its 20,560 houses listed for sale and disclosed that it lacked the liquidity to meet $265.9 million of debt falling due within a year [40]. Institutional buying had been shrinking since interest rates rose in 2022; Invitation Homes bought 2,072 homes in 2024 and sold 1,501 [41]. The ban arrived after the buying stopped and may accelerate the selling, the one outcome that could lower prices in the concentrated metros, and also the one most likely to depress the equity of households who already own there.
| Risk | Severity | Assessment |
|---|---|---|
| Ownership data remains unmeasurable | No single data source identifies the ultimate owner of a rental home; portfolios held through LLCs, joint ventures and funds are invisible to property records, so the 350-home threshold rests on self-reporting to HUD [7]. | |
| Purchase ban shrinks rental supply | Institutional entry added 0.5 rental homes per purchase; removing the buyer without replacing the supply raises rents for the households least able to buy [13] [42]. | |
| Capital migrates to exempt channels | Build-to-rent, renovate-to-rent and inter-investor sales remain lawful with no divestiture obligation, so the same capital keeps growing through construction and consolidation [39]. | |
| Rent caps push supply into exempt segments | Catalonia's caps coincided with a 22% fall in new contracts and a 45% jump in exempt seasonal listings, the standard leakage pattern for price controls without supply measures [31]. | |
| Divestiture wave hits concentrated metros | Institutional net selling rose 408% in a year; a disorderly exit in Atlanta, Phoenix or Jacksonville would lower prices for buyers and equity for existing owners alike [40] [9]. |
The alternative track is antitrust, and the CRS draws the contrast directly: enforcing competition law against suspected anticompetitive conduct is less likely to compromise liquidity in single-family markets than restricting who may buy [7]. Joe Gyourko of the Wharton School, writing for Brookings in February 2026, argues that localised market power should be addressed by antitrust authorities rather than a purchase ban, and that a sector holding just over 3% of the rental stock is too small for any intervention targeting it to produce meaningful affordability gains ◈ Strong Evidence [42]. The RealPage decree shows what the antitrust track produces: conduct remedies, a monitor and a template that applies to every landlord regardless of size [21]. The ROAD Act shows what the ownership track produces: a freeze on a shrinking category of buyer, with the growth channels left open.
Outside the United States the responses map onto the same two tracks with a third added. The Netherlands, Canada and Australia chose to exclude a category of buyer, and the Dutch data show this works for owner-occupation and against renters [29]. Germany and Catalonia chose to regulate price, and the Catalan data show the leakage that follows [31] [48]. Ireland chose to tax the transaction [32]. Berlin proposed to change ownership itself, and the federal government has moved to make that impossible [27]. None has yet matched the one remedy the evidence most consistently supports: building enough that the marginal buyer, whoever it is, has less to bid for.
The Debate and the Verdict
Symptom or cause
One camp says institutional investors own 1% of the housing stock and cannot be the cause of a crisis thirty years in the making ◈ Strong Evidence [41]. The other says they own a quarter of the rentals in Atlanta, file evictions at higher rates, charge fees the FTC called unlawful and priced rents through a shared algorithm ◈ Strong Evidence [8] [17] [20]. Both describe the same industry accurately ⚖ Contested. The disagreement is about what financialisation is.
The case that institutions are a symptom rests on the denominator. The American Enterprise Institute's housing centre, using Parcl Labs data, finds that investors owning 100 or more properties held 1.0% of the single-family stock in June 2025, that their share of purchases never exceeded 3%, and that small and medium investors accounted for more than 90% of investor purchases in every year for two decades ✓ Established [41]. John Burns Research puts the share above the ROAD Act's threshold at 0.7% of homes and 1% of 2025 purchases [43], and Gyourko notes that single-family rentals are only 11% of the occupied stock [42]. On this reading institutional buying was a reaction to a shortage created by land-use restriction and cheap money, not its cause, and by 2025 the institutions were net sellers anyway [41] [40].
The case that institutions matter rests on the numerator being in the wrong place. Three percent of the nation's single-family rentals is 25% of Atlanta's and 8.5% of the entire housing stock in particular suburbs [8] [13]. Consolidation in exactly those places raised rents [14]. Eviction filings were 18% to 19% more likely under some institutional owners [16]. The largest operator returned 39.2% of deposit dollars against a 63.9% norm and paid $48 million to settle a federal deception case [17]. And the algorithm that added an estimated $3.8 billion to a year of rent required no ownership share at all [20]. On this reading the ownership statistic is a distraction from the pricing regime, and the pricing regime is the point.
The Case That Investors Are a Symptom
Investors with over 100 homes owned 1.0% of the single-family stock in June 2025 and never bought more than 3% of homes sold in any year [41].
Structural estimates find 0.5 rental homes added per home purchased, with rents falling on net because of scale economies [13].
Small and medium investors raised price-to-income ratios and shifted construction toward multifamily; the listed REITs' price effect is modest [15] [7].
All six high-investor metros studied by GAO grew in population, rents and housing units together, and homeownership rose in four of them [9].
The Case That Investors Are a Cause
Institutions own 25% of Atlanta's single-family rentals and up to 8.5% of all housing in some ZIP codes, and entry explains 20% of price growth in the top-decile markets [8] [13].
Where merging landlords overlapped, rents rose relative to comparable neighbourhoods, the signature of local pricing power [14].
A $48 million FTC settlement, 39.2% deposit returns, fees above $1,700 a year and a $2.2 million Minnesota restitution fund are matters of record [17] [18].
Filings were 8% more likely under large corporate owners and 18% to 19% more likely under some institutional owners in Fulton County [16].
The resolution is that the two camps are measuring different things. The symptom camp measures ownership and finds it small, which it is. The cause camp measures conduct and pricing and finds them consequential, which they are. Financialisation, in Aalbers's definition, is the dominance of financial practices and measurements, not the concentration of deeds [2]. A market in which the marginal house is priced from its capitalised rent, renewal increases are calibrated to moving costs, fees are designed as revenue lines and competitors' data flow through a shared model is a financialised market whether funds own 1% or 25% of it. The RealPage record is decisive because it shows the pricing regime at full strength in a sector, multifamily apartments, where ownership is fragmented [19] [20].
That reframing has a practical consequence. Policies aimed at ownership share, whether the ROAD Act's 350-home threshold, Ireland's stamp duty or the Dutch municipal bans, act on the smallest and most visible part of the phenomenon, at the cost of rental supply, which the Dutch and structural evidence agree falls when the institutional buyer is removed [29] [13]. Policies aimed at conduct, the RealPage decree's data-age and geography limits, the FTC's fee-disclosure order, the Minnesota maintenance settlement, act on the pricing regime directly and apply to every landlord regardless of size [21] [17] [18]. The first kind is easier to legislate and to campaign on. The second kind is the one the evidence supports.
Ownership is the distribution of financialisation; pricing is its mechanism. Institutional landlords are a small share of a large asset class that behaves financially all the way down, from the pension fund buying build-to-rent in Manchester to the foreign buyer in Minato to the software recommending Tuesday's rent in Atlanta. Regulating the 350th house changes who holds the deed. Regulating the data that sets the rent changes what the deed is worth, and that is the lever the last two years of enforcement have shown to move.
What remains open is whether any of this lowers the price of a home for the household that wants one. The St. Louis Fed's arithmetic, prices up 207% against incomes up 155% over a quarter-century, was not produced by 450,000 institutionally owned houses, and it will not be reversed by preventing the purchase of the 450,001st [3] [8]. It was produced by a market that reprices shelter toward the return capital demands, wherever capital is free to enter and building is not. Housing became an asset class because it was allowed to be scarce. The landlord with 86,000 homes is the most visible consequence of that scarcity, the algorithm its most efficient exploiter, and the household paying $2,439 a month for a house it once expected to own pays for both [11].