The staff at Downton Abbey

The staff at Downton Abbey

Scraped clean of rationalizations, roommate caps are simple. They are tools that privileged people use to exclude from their neighborhoods people without much money, such as immigrants and students. To reveal this elitist reality fully will require this full article, but one example shines a bright light on part of it: how land-use codes treat servants.

At least six Cascadian cities specifically exempt live-in servants from the residential caps they impose on everyone else. For example, in the region’s fifth largest city, Burnaby, BC, a house may hold only five unrelated people. As everywhere in the region, families may pack as many members as they like into their residences, but for unrelated people, Burnaby allows no more dwellers in its tens of thousands of single-family homes, condos, and apartments than five. Regardless of the size of the home, no extras may move in: no friends in need, no additional roommates to help cover the rent, not even a parent or child.

But if you can afford servants? Well, by all means, invite them! All the butlers, housekeepers, gardeners, cooks, chauffeurs, and nannies you can afford are welcome to share your roof in Burnaby (see the definition of “family” in Burnaby code). The same is true in the region’s seventh largest city, Boise (“dwelling unit” at 11-01-03.01); its tenth largest city, Yakima (“family” at 15-02); and in Gresham (the 18th largest, [“family”]), Hillsboro (21st, [“family”]), and Beaverton (23rd, “family” in ch. 90]), Oregon. Between them, these cities hold nearly 900,000 people, all governed by a housing rule that is more Downton Abbey than Portlandia.

Other rules favor the rich and propertied less overtly but no less egregiously, as this final installment on occupancy limits explains. First, though, let’s circle back and set the stage properly.

Same Tune, Different Lyrics

"Room for rent" sign on picket fence in front of house.

Room for Rent, flickr, Doug Waldron

You may not want roommates. That’s fine. I would not wish them on you. Still, your city should not forbid you to rent out your spare rooms, and in many cases, cities do. Current land-use rules criminalize shared quarters, and shared quarters offer the simplest fix for housing affordability. Occupancy limits constrain the most profitable, abundant, and sustainable housing option available: the more than one-third of Cascadia’s existing bedrooms that sit empty each night.

So far in these articles, I’ve illustrated the absurdity of occupancy limits, tallied the excess of bedrooms, detailed Cascadia’s battery of arbitrary roommate restrictions and showed how illogical and hard-to-enforce they are. In this article, I aim to debunk the excuses that pose as arguments in their support. Occupancy limits serve no legitimate public-policy purpose, at least none that is not — or would not — be better served by other rules. Consequently, Cascadia’s cities would do well to simply delete occupancy limits from their law books.

Roommate caps have sailed under many colors over the years. A parade of rationales, each tuned to the times, is evident in the minutes of political bodies and in court decisions. The late University of Oregon planning professor Marsha Ritzdorf documented this procession of excuses (see, for example, “Not in My Neighborhood” in Lifestyles: Family and Economic Issues Vol. 9, No. 3). Proponents have argued for limiting occupancy to protect singles’ virtue by curbing promiscuity, safeguard the young from unsavory influences, limit noise and filth and polluted air, stop crowding, prevent disease, preserve neighborhood character, temper overflow parking, ensure safety and sanitation, defend vulnerable renters from exploitation by slumlords, and more.

Any policy that stands immutable while its purpose shifts merits suspicion, and many of the historic arguments are no longer mentioned. But some of these rationales survive, including preventing crowding and its health effects, preserving neighborhood character, stemming noise and parking, and defending vulnerable renters. Let’s take the modern arguments for occupancy limits in turn; none is even close to compelling. By teasing them apart — reverse engineering them — we can discern occupancy limits’ true purpose.


You might suppose that roommate maximums’ primarily serve to prevent too many people from crowding into a dwelling, and at some level, that common-sense-sounding rationale underlies many of the others. It’s therefore worth unpacking this idea.

Studied in their particulars, occupancy limits reveal that they are not crowding policies — not even close. Crowding policies would set maximum ratios of people to indoor space. A number of people with no reference to an area is not a crowding policy: Seven people in a phone booth is crowded; seven people on a soccer field is not. And occupancy limits set no ratios. In the numerator (number of people), they exempt all family members — the very groups that most commonly share dwellings — and only fence out unrelated people. Thirty distant cousins may share a Vancouver, BC apartment, but three family members plus three friends may not inhabit a mansion. Occupancy limits control not people but certain kinds of people.

In the denominator (amount of space), occupancy policies are effectively blanks. They indicate how many unrelated people may share a “dwelling unit,” which could be, for example, a 300-square-foot micro-studio apartment or Bill Gates’s 66,000-square-foot home.

What’s more, every housing code I’ve reviewed includes a separate provision that regulates crowding as a ratio of people per square foot of floor area. That’s right: occupancy limits are not crowding rules, but cities do have crowding rules. In Seattle, for example, the “floor area” section of the housing and building code says that two people may not share a room of less than 150 square feet and that for each additional resident another 50 square feet are required. That means, for example, that in a 400-square-foot studio apartment, Seattle allows two people in the first 150 square feet, and one additional person in each additional 50 square feet. That comes to seven people. Unlike occupancy limits, these floor-area rules apply to family members. (March 1, 2013 UPDATE: British Columbia’s provincial building code employs a simpler, and more-restrictive, crowding rule: two people per bedroom or sleeping area.)

Crowding rules, furthermore, are usually minimal and rarely breached. Seven people in a 400-square-foot apartment is crowded, by contemporary standards in the Northwest (though not perhaps by the standards of Coast Salish longhouses or pioneer cabins). Yet Seattle’s crowding rule is aligned with the 1986 recommendations of the US Centers for Disease Control and Prevention (page 43). It’s also aligned with the 1997 Uniform Housing Code (section 503.2, as adopted by Medford, Oregon), a product of the International Conference of Building Officials (since merged into the International Code Council). This official suggestion, widely followed across North America, employs a formula of minimum floor area per person for sleeping quarters that, if simplified and averaged, comes out to about 50 square feet. Strikingly similar standards are on the books in many other places, including faraway locales such as the United Kingdom and New Zealand.

Crowding and Health

Do occupancy limits — or even the less-demanding floor-area standards on the books across the Northwest — flow from a body of research that reveals the threshold of unhealthy crowding? Far from it. To my initial surprise, no such threshold is evident in the research literature. In fact, experts have been searching for decades for crowding’s fingerprints in causing disease and other harm. They have met with little success. For example, researcher Alison Gray, who studied this question in New Zealand, summarized her extensive global literature review:

The debate about the relationship between crowding and health is long standing and inconclusive. The complexity of relationships makes it difficult to separate the effects of crowding from confounding variables such as the physical condition and type of housing, socio-economic factors and lifestyle choices. Issues of measurement and other methodological difficulties limit the ability to establish causality. Many researchers are left concluding that in practice it is not possible to move beyond the level of statistical association.

I have reviewed almost a dozen summary reports from public health and housing experts in the United States and abroad. They all echo Alison Gray’s words. A UK report, for example, finds an abundance of weak statistical associations between crowding and various health problems but no conclusive evidence of crowding as a root problem. A World Health Organization study from Europe made a rigorous estimate of the role of crowding in spreading diseases such as tuberculosis but acknowledged that in affluent countries such as the United States and Canada, researchers have found hardly any correlation between crowded housing and infection.

The challenge for researchers is that people who live in crowded dwellings almost always do so because they are poor or otherwise disadvantaged. When crowded people have more money, they move into larger accommodations. Poverty and disadvantage are a root cause of illness, chronic health conditions, depression and other mental health challenges, family stress that leads to violence, and more. The effects of poverty are so big that they tend to obscure any independent effect from crowding.

Statistically speaking, searching for the fingerprints of crowding on human health is like looking for the effects on soldiers’ health of the stress they experience when they are under fire. The stress is presumably bad for their health, but the main threat to their health is the bullets. And bullets cause the stress anyway. So it probably makes most sense to focus on the bullets rather than the stress.

Researchers do find an abundance of correlations between crowding and health problems.

They assume that some of those associations reflect causation, and they have guesstimated guidelines on how much space people need. Those guesstimates are reflected in housing codes as floor-area rules, but no one really knows if they are right. And everyone knows that poverty and disadvantage — the bullets — are the main problem.

As a result, setting crowding standards is risky. Banning close-quarters living forces people to spend more money on housing, which leaves them less for everything else. It impoverishes them further: more bullets. In Cascadia, however, most floor-area requirements are so minimal that they are unlikely to constrain inexpensive housing. Fifty square feet per person isn’t much: it’s about the footprint of a king-size bed. One community with a more burdensome floor-space standard is Salem, Oregon. It requires 200 square feet per resident in short-term dwellings such as rooming houses, where migrant workers might live, but not in apartments and houses (Definition of “occupancies,” 59.035). (March 1, 2013 UPDATE: Tacoma, Washington, has an even more egregiously restrictive crowding rule: 300 square feet per person, and British Columbia’s 2-people-per-sleeping space rule is also more constraining than the customary 50-square-feet-per-person standard.)

Most cities’ occupancy limits, however, do constrain inexpensive housing. They close off the market for empty bedrooms in large houses and apartments and make neo-rooming houses like Seattle’s eight-roommate aPodments illegal in most of the region.

What Is Crowding, Anyway?

Crowding, what’s more, is in the eye of the beholder. Professor Dowell Myers of the University of Southern California and his coauthors wrote (see here) in 1996, “After a century of debate it is still in question whether so-called overcrowding is harmful to the people affected, or merely socially distasteful to outsiders who observe its presence.” Alison Gray, the New Zealand policy researcher, writes: “Standards generally reflect the values of dominant or decision-making groups and do not necessarily incorporate the views of householders or minority groups.”

Those dominant group values, moreover, are not static. They rise with the middle-class housing norm. In the United States, official government surveys of housing in the early 1940s counted any dwelling with more than two people per room (all rooms, not just bedrooms) as crowded. By 1950, officials had lowered the threshold to 1.5 people per room and, by 1960, to one person. Residential space was expanding far faster than population throughout this period, so the norm changed. Nowadays, in Cascadia, average occupancy is less than 0.5 people per room. Even as crowding diminished, many cities tightened or created their first occupancy limits for unrelated roommates. The 1970s were the peak period for adopting these limits, which effectively banned shared living arrangements — letting out multiple rooms or doubling up in houses — that had been normal in previous generations.

All evidence and logic suggest crowding is not the problem that occupancy limits are actually designed to prevent. Occupancy limits exempt families and do not control crowding. Even if they did, crowding is socially defined. Health research provides little reason for regulating crowding, and in any event, Cascadian cities already have separate crowding rules. What about other rationales for occupancy limits?

 “Neighborhood Character”

Houses in a Seattle Neighborhood

Seattle Neighborhood, flickr, Bev and Steve

Occupancy limits cap roommates, not families. That’s been especially true since the mid-1970s, when two key decisions from the US Supreme Court created precedents under which occupancy limits persist. In the first, the court upheld a New York town’s authority to stop a property owner from renting a single-family house to a group of six college students. Justice William O. Douglas, a Northwesterner, wrote for the court, “A quiet place where yards are wide, people few, and motor vehicles restricted are legitimate guidelines in a land use project addressed to family needs. . . .The police power is . . . ample to lay out zones where family values, youth values and the blessings of quiet seclusion and clean air make the area a sanctuary for people.”

It sounds good, but it’s a curious argument. Aren’t students people? Aren’t they youth? It’s a dated argument. We know now that suppressing population density makes for more traffic and air pollution, not less. It’s a classist argument, focused on the idyll of fortunate families raising children among other fortunate families amid “the blessings of quiet seclusion” where “yards are wide” (in other words, expensive). The ruling affirms that localities may write land-use rules to exclude people who cannot afford to live in a single-family neighborhood unless they double up with roommates. That’s not exactly “family values, youth values” so much as it is class values.

A year later, the Supreme Court extended its own distinction between family members and roommates. It ruled that a community could not restrict members of an extended family from sharing housing. Cities across the United States revised their occupancy limits to eliminate bans on extended family members. Canadian land-use rules now reflect the same principle.

The pro-family bias of these decisions may have reflected mainstream views in the mid-1970s; it is now an anachronism. Household structures have changed dramatically in the intervening years, with higher divorce rates, delayed marriage and childbearing, extended lifespans, and the proliferation of cohabitation, melded households, same-sex partnerships, and shared housing. US housing laws now make it illegal to discriminate against potential tenants or buyers on the basis of their family status, yet land-use codes persist in precisely this kind of discrimination. State courts in jurisdictions including California, Michigan, and New Jersey have since thrown out roommate caps entirely. Cascadian courts have not yet followed their example.

Nowadays, defenders of occupancy limits are less likely to argue for them as protecting youth and families than as safeguarding “neighborhood character,” but that’s just code. It’s a class marker. It means “middle- and upper-class families who own detached houses.” It’s a polite way for the privileged residents who tend to dominate city politics to say that they don’t want people in their neighborhoods who need to share housing with several roommates to afford the rent. They don’t want students, recent immigrants, or blue-collar workers.

Noisy Neighbors

Occupancy limits are sometimes rationalized as dampers on nuisances such as noise. Consider the implicit reasoning. If occupancy limits were effective ways to ensure quiet, then unrelated roommates must, even in modest numbers, be much noisier than family members. Family members are welcomed in infinite numbers, but roommates are strictly limited. Are the kind of people who have roommates noisy? No actual evidence has ever been adduced to support the proposition that roommates are categorically noisier — much noisier — than family members. This rationale trades in sweeping stereotyping.

  • In any event, if the concern is noise, the appropriate response is a noise rule. Such ordinances are simple to adopt and easy to enforce. Neighbors call police; police knock on the door; residents hush. Regulating the number of roommates as a noise-control measure is like trying to stem litter by capping the size of groups of friends who may walk the sidewalks: it’s an insane, or at least inane, policy. Why not just put out some trash cans and fine litterbugs?

    The Parking Problem

    More roommates may bring more cars, which may clog on-street parking spaces, inconveniencing neighbors. Roommates may share cars less than do family members, justifying limiting roommates more than family members. This argument is, morally speaking, inverted: if airlines operated by the same logic, they would not limit or charge for luggage and would fly with empty seats to avoid overfilling the overhead bins and cargo bay. Refusing affordable housing in unoccupied bedrooms to perhaps tens of thousands of people across the Northwest because they may make on-street parking harder to find for current residents in some places some of the time is a perversion of our better values as a region. Still, the parking argument is at least plausible, unlike others I’ve examined, and managing on-street parking is a legitimate goal for public policy.

    Still, the logic of the parking argument holds only in places where on-street parking is a free for all, an unregulated commons. The argument’s validity evaporates in neighborhoods that use the emerging kit of technologies and policies for solving on-street parking problems. I plan several articles soon on the revolution in urban parking that is within Cascadia’s reach, and the big benefits it promises for housing affordability, local economies, and neighborhood improvement. Rather than capping roommates in hopes of alleviating parking pressures — a morally dubious proposition — cities can resolve their parking woes head on, making the parking argument for occupancy limits moot. In fact, with the right parking policies, neighborhoods profit financially from roommates who move in and bring cars. More on this another day.

    Defending Vulnerable Renters

    A planner in a Canadian city told me, “We need occupancy limits or we’ll have Chinese immigrants packed in, sleeping in shifts.” He didn’t indicate what the problem was, exactly, with sleeping in shifts, or why his city government had any legitimate role in determining how people slept, whether Chinese or not. He seemed to think it obvious. But rather than simply writing off his comment as racist, I am going to assume he was giving voice to a final, common argument for occupancy limits: that occupancy limits prevent unscrupulous slumlords from exploiting vulnerable tenants by packing them into unsafe housing.

    It is a frustratingly vague argument. Does it mean that, absent occupancy limits, landlords would crowd tenants beyond what they would choose for themselves? If that’s what it means, it’s nonsense. As I’ve explained, occupancy limits are no good at preventing crowding. That’s a job for square-footage limits. And besides, crowding is one tactic that poor people can use to cope with poverty — meaning that occupancy limits don’t expand the housing options of the impoverished, they constrain them.

    Perhaps, then, people think that high-occupancy housing is simply unsafe. But the real safety problem lies in housing that’s dilapidated or that violates code. Unscrupulous landlords feel no particular need to maintain their housing or to follow city codes, subjecting their tenants (crowded or not) to safety and health hazards: lack of fire exits, fire doors, and smoke detectors; faulty ventilation and inadequate heating; unstable floors and railings; compromised electrical and plumbing systems; moist walls that grow black mold; radon seeping up from the ground; and more.

    Yet occupancy limits do absolutely nothing to end such code violations; in fact, they encourage them. By excluding many empty bedrooms in existing homes from use as rentals, occupancy limits eliminate slumlords’ competition. Roommate caps enhance landlords’ market power and leave them less eager to maintain their rental properties. Removing occupancy limits, conversely, could bring new housing options onto the market, loosening slumlords’ grip on their tenants by giving those tenants new alternatives.

    The way to prevent exploitation of vulnerable tenants is to enforce safety rules and to cultivate competition at the bottom of the rental market. Occupancy limits, because they tighten rental markets, make exploitation more likely, not less.

    Erasing Roommate Caps

    Crowding, disease, noise, parking spillover, exploitation of renters — none of these ills is a reason for occupancy limits. Cities have tools for addressing these problems, and some of these tools, such as safety codes and floor-area provisions, are already in force across Cascadia. Assuming occupancy limits are not an epidemic of insanity but are crafted for a purpose, perhaps unspoken, that purpose must be something other than these rationalizations.

    The real purpose must be the one thing that they actually do. They exclude from neighborhoods the kinds of people who would need to share housing with more roommates in order to afford the rent: people who are not members of middle- and upper-class families; people who are students and therefore lack money; people who have recently immigrated and are still climbing up from the bottom of the wage ladder; people who, for whatever other reason, are poor.

    This urge to exclude is likely motivated by a noxious but powerful blend of self-interest (protecting property values by making neighborhoods more uniformly affluent), stereotyping (of renters as noisy, for example), and racism or classism (“packed in, sleeping in shifts”). It’s  precisely this vile mixture that explains the mockery of Northwest values with which I began this article: the fact that elected leaders in cities that are home to almost 900,000 Cascadians have left untouched in their law books provisions that cap roommate numbers while welcoming all live-in servants.

    All of this analysis leads inescapably to one and only one conclusion: occupancy limits deserve elimination. Cascadia’s cities can simply erase them, strike them from the law books, cross them out, consign them to history. That’s the recommendation of Professor David Ormandy of the University of Warwick in Coventry, England, perhaps the world’s leading expert on housing, crowding, and health. He argues for “decriminalizing” inexpensive housing by eliminating all occupancy standards. Instead, he suggests that authorities employ a rating tool to inform housing inspectors’ own professional judgments about safe and healthy housing. The rating tool does not completely ignore crowding, but it puts it in context and focuses on the real issues — the bullets.

    For city clerks, revoking occupancy limits would be a few seconds’ work. The limits mostly live in the complicated definitions of “family” that cities have adopted to specify who can live in a dwelling. All that cities need to do to remove occupancy limits is to replace their family definitions with the single sentence employed by Surrey, BC, Cascadia’s fourth largest city: “Family means 1 or more persons occupying a dwelling unit and living as a single non-profit housekeeping unit.”

    Or, if they prefer, they could use Victoria’s: “Family means one person or a group of persons who through marriage, blood relationship or other circumstances normally live together.” Or they can do what Bend, Oregon, has done: entirely eliminate the definition of “family” and references to it.

    Adopting either definition, or no definition, in all the Northwest’s cities could unfetter thousands — or perhaps tens or even hundreds of thousands — of unoccupied bedrooms for rental: rental to anyone the owner choses to house. They would also be available, you might notice, for live-in servants. A no-limits policy would not discriminate, even against the propertied and privileged.

    Thanks to Mieko Van Kirk for research assistance for this article.