Tuesday, March 21, 2017

The Impact of Rent Control on Oregon's Housing Market

This article is written by Gerard C.S. Mildner, PhD, Director of the Center for Real Estate at Portland State University.


The Oregon State legislature is considering House Bill 2001, sponsored by the Speaker of House Tina Kotek, which would override the state’s prohibition on local rent control ordinances. HB 2001 has three provisions: (1) limit rent increases statewide to 5% for 2017-18, (2) permit any local jurisdiction in the state to enact rent control, and (3) modify the inclusionary zoning legislation that was passed in 2016. The challenge in analyzing this legislation is that no details are provided regarding the nature of the rent control legislation that the local jurisdiction may implement. No limits have been placed on local governments’ rent control power, except that a “fair return” to landlords must be provided. Discussing the efficiency of rent control legislation that hasn’t been enacted is rather like commenting on whether a baby is beautiful while the mother is still pregnant.

The Speaker’s Office has circulated an argument that the rent control that will be implemented after this legislation is passed will be “second generation” rent control, as defined by Canadian economist Dr. Richard Arnott. According to Dr. Arnott, second generation rent control is typified by limits on rent increases (rather than a rent freeze), prohibition of rent control on new construction, allowances for larger increases with rehabilitation and major improvement, and allowances of decontrol between tenancies, none of which is included in HB 2001. I believe the differences between first-generation and second-generation rent control are largely semantic, and if there is a difference, it’s really a degree of the harm that the legislation will generate.


Unlike most countries, the United States has a very limited experience with rent control. Only three states – New York, California, and New Jersey – plus the District of Columbia, have rent control. New York has some form of rent control on a statewide basis, but regulation in California and New Jersey operate at a municipal level. Of the 10 largest metropolitan areas in the United States, New York, Los Angeles, the San Francisco Bay Area, Washington, DC, and by including its New Jersey suburbs, the Philadelophia metropolitan area have rent control. These states and metro areas have some of the highest real estate prices and the highest percentages of tenant households, which explains why those places have rent control. The Boston metropolitan area had rent control until the legislature ended it in 1997.

The lack of rent regulation in the United States is a positive development. Compared to most industrialized counties, the United States has much lower barriers to development than other countries, and a much lower ratio of housing costs to household incomes.

Effects of Rent Control

The main negative effect of rent control is to limit supply, deterring the development of new rental housing and deterring maintenance in rental housing. Multi-family housing developers produce business plans that project net income using assumptions about future rents and expenses, and the persuasiveness of those business plans determines if developers can attract investment capital and bank financing. Legislation that restricts those future increases in rents will reduce the attractiveness of those business plans and will force developers to set higher initial rents to account for the slower increases likely to occur in the future. By raising the price point required for new investment, rent control limits housing supply and raises rents.

A second way that rent control reduces supply is to limit apartment rehabilitation. Mutli-family housing investors will often buy apartment buildings that have deteriorated, and improve those properties over a period of years. The strategy is to offer a higher level of amenities for a higher rent. These investors usually use the normal turnover of tenants to rehabilitate individual units, as the typical tenant in the United States lives in an apartment for 12-18 months. Improving an apartment from a “C” property to a “B” property or a “B” property to an “A” property will often result in substantial increases in rents. Blanket curbs on rent increases (or forcing the developer to apply for permission to achieve higher rents) will deter this sort of investment activity.

The Speaker’s Office argues that new construction will be exempted under “second-generation” rent control, reducing these impacts on housing supply. However, nothing in HB 2001 guarantees this. Moreover, the history of rent control in other jurisdictions suggests that promises to protect new construction from rent control are often not kept. For decades, New York City maintained a distinction between “pre-war” apartments, built before 1947 and governed by rent control, and “post-war” apartments which were not regulated. However, the supply of post-war apartment units and and the political constituency of post-war tenants grew sufficiently large that the New York City Council, under the leadership of then Councilman Ed Koch, enacted rent stabilization in 1969 to cover units built between 1947 and 1969, promising not to regulate construction after 1969. Later in 1974, the New York State Legislature enacted the Emergency Tenant Protection Act to extend rent stabilization to units built between 1969 and 1974, promising not to regulate units built after that date. The legislature later amended the Emergency Tenant Protection Act in 1983 to regulate units built up to that date.

These repeated broken promises have led developers and investors to be wary of investing in states that allow for rent control to be enacted. As noted by economists Ed Glaeser of Harvard and Joseph Gyourko of the University of Pennsylvania, New York, New Jersey, and California have some of lowest housing production rates of any states in the United States. While that lack of housing supply elasticity may have many causes, including excessive zoning regulation and land controls, the threat of rent control is an important factor in deterring new housing investment. And since the reputation (and housing market) of an entire state can be harmed by a single jurisdiction imposing rent control, it’s important for the legislature to maintain the existing prohibition of rent control by local jurisdictions.

A second negative effect of rent control is the politicization of landlord tenant relationships. In unregulated markets, an apartment lease is a voluntary contract between two parties. Landlords are unable to form monopolies or cartels so that competition for tenant demand is vigorous. Apartment owners offering low quality units will earn lower rents. Apartment owners offering improvements in quality (which is a form of housing investment) will be rewarded with higher rents. Once rent levels become statutory, all aspects of leasehold terms become a matter for legal adjudication. If rents are held below market levels, a landlord will face many potential tenants, and will lose the incentive to maintain established quality levels. This might include reducing thermostat temperatures in winter, reducing frequency of common area maintenance, reducing hours of security personnel, or reducing the frequency of garbage collection. Arnott estimates that direct government expenditures in Ontario for enforcing the housing laws are $5 per year per person due to rent regulation. In the City of Portland would amount to $3 million per year and a major expansion of our municipal court system. This amount doesn’t include the legal cost to landlords or tenants.

A common reaction to the politicization of landlord-tenant relationships is for the landlord to convert their building to condominium ownership. If full value of the unit cannot be achieved in the rental market, the landlord can sell their apartments in the ownership market. Under New York City’s regulations, for example, rent stabilization applies to apartments of 12 units or more. Under the current legal interpretation, an apartment owned as a condominium is treated as a single housing unit, even if the building has hundreds of apartments. As a result, the housing stock in New York has changed in the last 50 years from being 90% rental and 10% ownership to 70% rental and 30% ownership and about one-third of those rentals are unregulated, either because they are in small buildings or units in condominium ownership. By comparison, HB 2001 allows for regulation of all rental properties, unless it is a duplex with one unit owner-occupied. My interpretation of the proposed statute is that a rented condominium unit would be subject to regulation.

The politicization of the landlord-tenant relationship also extends to the relationship between tenants and sub-let tenants. It’s fairly common in New York for tenants to take on roommates to share the costs. However, sublet tenants do not enjoy tenant protections, so one of the oddities of New York’s regulation is that “primary” tenants (those tenants named on the lease) can charge market rents to sublet tenants, even to point where the sublet tenant is paying more than the statutory rent. My interpretation of the proposed statue is that sub-let tenants will not be protected.

The Welfare Impact of Rent Control

Rent control can be seen as a transfer of income from landlords to tenants, which in the traditional cost-benefit framework of economics is not a social cost. Social cost is measured by economists as a burden on consumers or suppliers not counterbalanced by any transfer. In my PhD dissertation, I estimated the lower bound of these social costs for the housing market in New York City at between 0.9% and 1.4% of income in the region. That may not seem like much, but that is approximately one-third of the cost of the payroll of New York City government, including its schools and City University system.

It’s also important to understand the distribution of those benefits among landlords and tenants within the community. It’s natural to think of landlords as wealthy and tenants as poor, so any transfer of income from landlords to tenants would appear to make the income distribution more equitable. However, only about one-third of households within the state are tenants and not every tenant will enjoy the protection of rent control. The likely outcome of HB 2001 will be the imposition of rent control within the city of Portland and no regulation in its suburbs. It also seems likely that the City of Portland will respond to demands by small landlords and allow exemptions for small unit buildings, where gaining the votes of a few tenants won’t be worth the cost of gaining the certain enmity of one landlord.

With rents held below market, only a fraction of tenants wanting those units will receive an apartment. Under every rent control statute that I’m aware of, existing tenants in place are protected, in favor of new applicants. Therefore, by its nature, rent control will protect older, white, more established tenants over newly-formed households, young adults, and newly arrived ethnic minorities and immigrants. When new vacancies occur, landlords will find multiple applicants offering the same rent and be able to select higher income and more stable households over applicants with more uncertain prospects. This discrimination will inevitably shift demand to unregulated markets, either among small, unregulated properties in the city and towards the suburbs. As rents rise in the unregulated sector, tenants in those markets will suffer. And even for the tenants lucky enough to find a rent regulated apartment, they will likely pay search costs and side payments as a means to find a below market apartment. It is customary in New York to make payments or “key money” to building superintendents and managers who often have the first knowledge of an apartment vacancy. It is customary for apartment seekers in New York to hire and pay commissions to apartment brokers, who in other markets are compensated by landlords looking to fill vacancies.

There is also an economic burden to the community from the reduction in value of the existing housing stock. As rents are held below market, apartment property values will fall and the burden of local property taxation will shift towards single-family households. This problem may be limited in Portland, where under Measure 5 and 50, assessed values of property are typically 30% below real market values and assessed values grow by a statutorily-determined rate of 3% per year. However, legislation to return Oregon to full-market valuation has been introduced in the current legislature, so the risk of a shifting of the property tax burden to single family households is real.

Finally, HB 2001 is inconsistent with the policies enacted by the 2016 legislature to respond to the housing crisis with inclusionary zoning. Under the policy sanctioned by the state and enacted by the City of Portland, developers of housing properties of 20 units or more will have to insure that rents are kept below market for 20% of the units at a level that is 30% of the income of household at 80% of area median. This policy involves a cross subsidy of below market units by market rate units. While this policy is of dubious value, given that it will act as a development tax, imposing rent control on the market rate units will only make these inclusionary zoning projects less likely.

Housing Prices and Rents in Portland

HB 2001 has been introduced as an emergency response to the problem of rising rents in the Portland metropolitan area. Rising rents and rising house prices in the metropolitan area is a function of high demand and inelastic supply. To understand that problem, we need to understand the impact of Oregon’s land use system.

When I moved to Oregon in 1991, my boss, Dr. Nohad Toulan, Dean of the Portland State University College of Urban and Public Affairs described to me the fundamental trade-off of Oregon’s land use planning system. Developers were restricted to build in the rural areas of the state to protect the agricultural and forestry economy of the state. Housing development was largely constrained within urban growth boundaries, and farmland was restricted in how it could be subdivided. However, within urban areas, developers were promised a 20-year land supply and quick approval of development proposals that met local zoning. Under the Metropolitan Housing Rule, suburban jurisdictions were required to allocate land for all housing types, even apartment buildings in the suburbs.

25 years later, that compact between environmental protection in rural areas and housing development within urban growth boundaries has broken down. Since the late 1970’s, the population of metropolitan Portland has grown by 78%, while the urban growth boundary managed by Metro and DLCD has grown by only 10%. It’s as if we were continuing to wear the size M shirt that our mother gave us in junior high school. As population has grown within a limited land supply, land prices have risen dramatically inside the urban growth boundary to the point that land prices are 10-15 times higher inside the UGB than outside.

This problem grew out of decisions in the 1990’s when our Region 2040 Plan was developed. Citizens were given the choice of “growing up or growing out”, which meant that we needed to densify the region if we wanted to avoid suburban sprawl. The choice was presented as a matter of urban design, with the subtext that “growing up” would be a mature decision and “growing out” implied gluttony. What was not understood at the time was the trade off that “growing up” would mean higher housing costs.

In a recent study that I completed for Holland Residential, we studied the rents charged for new housing developments in the Portland metro area built between 2010-15 and found a consistent relationship between density and rent levels. For garden apartments that were two-stories, usually wood construction with tuck-under parking or surface parking, apartment rents were $1.00 to $1.20 per square foot. For five story apartments that were built as four stories of wood construction over a concrete podium, rents were typically $1.70 to $1.90 per square foot. With few exceptions, these buildings were constructed in Portland west of 39th Street (Caesar Chavez Boulevard) and south of Killingsworth Street. Finally, true high-rise apartments bigger than 5 stories that were built with concrete and steel construction all came in at rents of $2.70 to $3.00 per square foot. And all of those properties were built in the Portland central city, including the Pearl District, South Waterfront, and the Lloyd District. As we shift to a “growing up” strategy, we require more production of higher cost units.

Density at Any Cost

The UGB’s restriction on land supply in the region is reducing housing production in the region and forcing what is developed to be much higher cost. Since the Great Recession, housing production in 2013-16 is about 20% below the average level in 1990-2007, when apartment rents and housing prices were much lower. When examined on a county by county basis, production is up in Multnomah County but down in each of our three suburban counties who rely upon the less expensive and more land intensive models of garden apartments and single family construction. The irony is that apartment construction has returned to pre-recession levels, giving Portland residents the impression that we are in a housing boom. Yet the boom in apartment construction doesn’t make up for the decline in the development of new subdivisions in the suburbs. And without the suburban single family construction, high income consumers have steered to traditional neighborhoods in the city, leading to gentrification and displacement.

This problem has been compounded as Metro has adopted a more unrealistic definition of a 20 year land supply that the planning system requires. Metro has followed a state-mandated procedure to estimate future population and employment growth. A buildable lands analysis determines if the existing UGB can handle the growth. Over time, Metro has developed models of infill development and redevelopment that supplies growing fractions of our housing supply. Infill and redevelopment often require using less attractive parcels that were left over from earlier waves of development, or the tearing down of existing structures, both of which require higher costs and higher rents. Metro also accepts at face value the zoning capacity that is presented by local jurisdictions, even when that capacity is uneconomic. The City of Portland, for example, has zoning in Gateway, Montevilla, Rockwood, and Lents that allows five-story construction along major arterials. However, these districts are the lowest rent areas within the Portland metropolitan area with rents of $1.00 to $1.10, much lower than the $1.70 to $1.90 needed to support that type of construction. For these districts to develop at five stories within 20 years, rents would need to double (accounting for inflation), which would mean astronomical rent levels in other parts of the region with more amenities.

In its 2015 decision to freeze the urban growth boundary, Metro justified this decision through its Metro Scope computer model, which constrained development within the existing UGB. Under this scenario, apartment rents and housing prices were projected to more than double in 20 years, reaching levels in Los Angeles and San Francisco. Now, before that scenario would happen, of course, economic growth in the Portland region would come to a grinding halt. However, the 2015 Urban Growth Report is a stark reminder that “growing up” has its costs, and that we need to accept the tradeoff in our liberal values of environmental protection and housing affordability. As a region and a state, we cannot afford to subsidize the housing costs of a third of our population. The challenge of housing is poor is compounded by our errors in land use planning.

Moving to a More Equitable Region

As advocates for low income households, we need to be sensitive to the problem of housing supply. As a necessity, housing is a much larger fraction of the income of low-income households than higher income households. Barriers to housing development create a disproportionate impact on the wellbeing of the poor than the middle income or the wealthy. We need to urge state government to increase housing supply by expanding the land supply in the region for human settlement. Allocating land in Washington County for producing grass seed, strawberries and hazelnuts at 7% of its market potential as land for housing is a misallocation of resources and a planning error.

We also need to recognize that we cannot and should not provide subsidized housing for every low-income household. If you add all the various housing subsidy programs (including low-income housing tax credits, public housing, housing vouchers, inclusionary zoning units), that represents less than half of all households in the United States below the poverty line. Most low-income people rent at market levels. Building subsidized housing creates a targeted benefit for a few households, but most low-income households find housing themselves. Those low-income households need the community to build large amounts of housing, which by its nature will be built for middle and upper income households, which will free up older housing at prices they can afford.


Finally, we need to avoid policies like inclusionary zoning and rent control that essentially tax new development to help produce a few below-market units. Subsidized housing units can be thought of as a public good for the community to help people with particular needs, but that kind of community resource should be paid by general taxation. Taxing new development may seem politically expedient, but it only perpetuates and compounds the problem.