What we talk about when we talk about a “housing crisis”

What do we talk about when we talk about a housing crisis? People doing without any kind of housing? People living in inadequate housing? Crowding together? Spending too much on rent? People struggling to get into home ownership? People not being able to afford ownership of the single-family detached house they always dreamed about?

To me, a “crisis” suggests fundamental needs unmet. But just what’s a “need”? How should this be separated from a “want” or “dream,” if at all? Addressing these questions and trying to figure out how they matter was the subject of my keynote talk at the PartnerLife conference last week in beautiful Cologne, Germany.

I illustrated my talk with my favorite case study: Vancouver. In the process, I ran some numbers to compare how Vancouver is doing relative to other metropolitan areas if we address some of the different things we mean when we talk about a housing crisis. In particular, I was interested less in the kinds of “dream” measures used by organizations like Demographia (oh, it’s painful to even link to them!), and more in the fundamental measures of need (note: I’ve compared rents elsewhere, though I need to update the comparison!).

How is Vancouver, long considered the most unaffordable housing market in North America using Demographia’s single-family detached house measure, doing when we look at homelessness? How about when we look at providing basic standards, avoiding crowding, and insuring affordability?

To answer the first question, we can look at homeless counts. I’ll work on building this measure further, but for now I’ll just compare Vancouver with our near neighbours to the south (Seattle and Portland) and to the east (Calgary). Is homelessness a major crisis here?

The first answer to this question is indisputably: YES. Homeless is a major crisis wherever it occurs, with large effects, for instance, on the risk of dying. But a more nuanced answer, of more use in thinking through solutions and sorting out what’s working, is to consider the relative size of the homelessness crisis. Though it’s far from a definitive comparison, I started looking into this question by comparing homeless count data by relevant population size, across the regions of Vancouver, Seattle, Portland, and Calgary.* Here’s what I get:


Is homelessness in Vancouver a crisis? Yes. But when compared to other nearby metro areas, Vancouver looks like it’s doing better. This is important in terms of judging the response so far and thinking through how to continue dealing with this crisis.

Let’s address the second question: How are we doing in terms of insuring people are living in adequate housing, not feeling too crowded, and not spending too much money on rent? In Canada, we have a nice comparative measure of “core housing need” that gets at these components of housing crisis. Importantly, these aspects of a “housing crisis” remain detachable, revealing, for instance, different sorts of crises between the North of Canada (where the issue at hand tends to be crowding) and the South (where it tends to be affordability). Overall core housing need is worst in the North and on reserves, where we can talk about some serious housing crises. But here let’s just look at how Vancouver is doing by comparison with other metro areas in Canada given the most recent data available.**


How’s Vancouver doing by core housing needs? Not so great. We’ve got a lot of people feeling the pain of unmet housing need, as defined by Canadian standards. Mostly these are people spending more then 30% of their income on rent. I’ll be the first to suggest that this is a funny standard, but it still indicates a real problem, especially for those at the bottom of the income distribution. At the same time, by comparison Vancouver is not actually the worst Canadian metropolis. The worst is tiny Peterborough, Ontario! What’s going on there? I’ve no idea, though now I’m quite curious (and it might just be the small sample size of the income survey). After Peterborough, Toronto is also worse than Vancouver.

While both homeless counts and core housing needs remain open to critique in terms of their conceptualization and measurement, they’re also the best measures of need we’ve got. As such, I’d argue they remain the best measures of when we’re seeing a real housing crisis. Using these measures, we can see that there are indeed housing crises at play in Vancouver. At the same time, in comparative context we can recognize that Vancouver’s doing much better at addressing these real crises than it’s typically been given credit for.

Why doesn’t it get credit for what it’s doing right? I think the unaffordability of the single-family detached house in Vancouver sucks up a lot of attention. I’ll continue to argue that this is a very BAD measure of a housing crisis. After all, if we want to reduce the size of our ecological footprints, if we want to support our great cities, if we want to combat isolation and obesity, and arguably if we want to sustain our democracies, then we want to discourage everyone from living in single-family houses. This means not everyone can get what they want. But it’s not a housing crisis if everyone still gets what they need. And by that measure, Vancouver’s doing better than most people think, even if it’s still got a lot of work to do.



*- It’s worth noting that the administrative basis for count data here varies between metro region (Vancouver), county (Portland and Seattle), and city (Calgary). I’ve used the relevant administrative data from the 2010/2011 census year as the denominator in each case. This means population has been kept constant for comparison purposes, while the homeless population has been allowed to grow, resulting in a slight underestimate of homelessness per 10,000 people in early years and overestimate in later years. Also of note, King County and Multnomah County are smaller than the metro areas of Seattle and Portland (accordingly). The City of Calgary is relatively co-terminous with its metro area. This could bias overall estimates of relative counts for metro areas. But even if we were to just use central cities (where the populations of Vancouver, Seattle, and Portland are quite similar at @ 600,000) or metro areas, the overall results would still be about the same – there are a lot more homeless people showing up in other nearby cities and metro areas relative to Vancouver. A caution also remains in the possibility for different definitions and methods in each region, particularly with respect to the meaning and coverage of “transitional housing.” Also of note: the big drop observed in Portland between 2011 and 2015 might be worth following up on!

**- the data here come from the CMHC, and are based on Canada’s income survey rather than census data. It’s possible the sample size of the surveys explains some of the variation, and rankings here should be considered preliminary until we get something more definitive, like the 2016 Census data! In past census years, Vancouver and Toronto usually compete in the metropolitan title for greatest proportion in core housing need, and Peterborough tends to be more middle-of-the-road.


Vancouver’s Transformation: from Phoenix to Montreal in under 60!

I can never quite get over just how dramatically the residential stock of Vancouver has transformed away from its former dominance by single-family detached houses. As recently as 1961, proportionately more of Vancouver’s residents lived in detached houses than in Phoenix, Arizona today (well-known for its inescapable sprawl). Less than sixty years later, Vancouver has now toppled both New York and Montreal as the least house-dominated metropolis in North America! That’s a dramatic transformation!


Above I’ve updated an older figure of mine that made its way into my (award-winning!) book with the newest census data from 2016, when Vancouver definitively displaced Montreal as the least house-dominated metropolis in North America. What’s led to and resulted from this dramatic change in Vancouver’s housing stock? That’s all in the book!


Surveying Realtors

I’m always both fascinated by and wary of the data produced by real estate associations. I initially had a whole chapter in my book devoted to taking apart survey data on consumer preferences put together by real estate organizations (sadly, but probably correctly, it got cut). Here’s one of my favourite such survey questions (see slide 8) based on what Vancouverites might want to buy if, inspired by the Bare Naked Ladies, they had a million dollars. (Nearly a quarter chose to keep the $1 million and rent!)

I notice that such data is back in the news again, this time based on surveys of realtors, from April 2016 to April 2017, who’ve recently represented buyers in sales. The write-up leaves a lot to be desired in terms of methods (what’s the sample size of realtors and buyers? what’s the response rate? are there warning flags in terms of representation of realtors and buyers?) It’s also unclear whether this represents entirely re-sale or also sales of new residential real estate. This makes it difficult to evaluate the quality of the data. But it’s still kind of fun to play around with it.

I’ve broken the data, as presented by REBGV, down into my own categories. Here’s type of sale:


According to recent surveys of REBGV realtors, investment purchases make up about one in five sales. The role of foreign investment (largely, but not entirely post-Foreign Buyer Tax) is relatively small. But survey quality, about which we know little, likely matters a lot for these estimates. Are some realtors and real estate companies more likely to respond than others (especially those, like New Coast, likely to especially target overseas buyers)? Other important details are also missing: Are sales of newly constructed properties included? How do realtors decide who counts as a foreign investor vs. a domestic one?

Setting investment purchases aside, first-time buyers, targeted by a much-derided recent BC Liberal finance assistance program, make up nearly a third of buyers. That’s a pretty big chunk of sales! But here it’s not clear quite what counts as “first-time.” First time in Vancouver, first time in Canada, first time at all anywhere? Other moves, making up nearly half of all purchases, tend to be from buyers moving around from one dwelling to another.

Finally, there’s really interesting data breaking down moves of owners moving from one property to another by type (condo apartment, townhouse, and detached house) at old home and new. I simplified this into lateral moves, moves to likely bigger units (upsizing), and moves to likely smaller units (downsizing). Many general life cycle models of housing assume households tend to upsize over time as they grow, the better to fit with children. Downsizing only (maybe) occurs after retirement or when children move out. But with Vancouver steadily moving away from single-detached houses, upsizing is the least likely type of move between owned units. Instead, most moves are either lateral (e.g. apartment to apartment) or downsizing. That’s pretty interesting, and likely reflects, in part, how people moving here from elsewhere in North America typically find a house out of reach.

And just where are people coming from?


Hmmm… returning to the data quality issue, it’s a little concerning to me that the “investors” category in this question is so much smaller (14%) than in the previous question about type of sale (20.8%). Where did the extra investors go? Did some of them move as they made investment purchases? Were others counted as living in the same community? Weird.

But we get some idea about what proportion of sales represent people moving here from beyond the Metro area, and it’s about 12%. That could account for many of the downsizers, as they reckon with the realities of Vancouver’s pricey market (esp. for single-family detached homes). Another healthy chunk might involve retirees (more on that in a second).

Setting aside investors, we can actually do a comparison of where moving buyers are coming from by looking to Census data (or more accurately, National Household Survey data). The 2016 data on mobility and migration aren’t out yet, but the 2011 data (limited access here, but also recently out in IPUMS) provides a breakdown for those who’ve moved in the past year. Limiting the sample to those in Metro Vancouver, I looked at household heads who’d moved in the past year and owned their own home. How did where came from match up to REBGV data in 2016-2017?


That’s actually a pretty good match! There is some difference in terms of who the Census thinks is moving within their own community relative to who realtors think of as moving within their own community. This likely relates to shifting definitions of communities (again, not defined in the REBGV data). But looking at the proportion of new buyers moving within the metro area (in green) relative to those moving in from away (blue and pink), the figures are actually quite close, at about 86% of non-investment residential sales being to local buyers.

The Census from 2011 would suggest slightly more recent buyers moving to the area came from outside Canada than the REBGV data from 2016-2017, but not by a lot (7.4% to 5.8%), and the disparity could arise from either historical change (including the imposition of foreign-buyer tax) or from issues with data quality (see above). Still, a pretty good match.

It’s actually harder to match up the “demographic” categories used by REBGV data to census equivalents. But playing around with the community profile data from BC Stats, I did my best. Here’s how new buyer households in the REBGV surveys from 2016-2017 kinda, sorta stacked up against all households in Metro Vancouver by household types in 2011.


Again, it’s tricky to make sense of REBGV categories and match them up to Census categories (the census, for instance, does not differentiate between “young couples without children” and “empty-nesters,” and I’ve no idea how these were defined for the realtor survey either). I also don’t know how demographics on investors were tabulated, or where they fall relative to households looking to buy a place to live. But the general match-up between all households (from 2011 Census) and new buyer households (from 2016-17 REBGV survey) looks plausible to me in terms of what I might expect. New household formation drives a lot of sales. So couples without children are disproportionately likely to buy a place while retirees (or those age 65+ in the Census) don’t actually move all that much (there’s a lot of aging-in-place).

I don’t know that I have a big takeaway from all of this data exploration. I think the REBGV data remains kind of sketchy for estimating investment purchases until we get some basic information about data quality and representativeness out of the way. But setting aside investors, the data on where new buyers are coming from when they move within or to Vancouver lines up well with what I’d expect from the census, which is reassuring and kind of cool.

Bright Lights, Big City, Bigger Metro Area

The young are generally drawn to the central cities of big metropolitan areas in North America. That’s where the bright lights can be found! The action! The scene! And also most of the universities, the majority of metropolitan rental stock, lots of the jobs, etc.

Vancouver is no exception. As revealed in previous posts looking at net migration patterns, the City of Vancouver really draws in the young people, both from the surrounding suburbs and further abroad. Then many of those young people gradually have kids and settle down into the surrounding suburbs again. Others remain, renewing the population of children and middle-aged stalwarts in the city. For the metro region as a whole, just how many people remain in the central city as they age and how many move out to the suburbs might not matter a great deal. But for a variety of municipal purposes (including the vitality of school districts), these patterns can matter a lot!

So is Vancouver weird in its age distribution with respect to what proportion of people live in its central city?

One issue in answering this question is the relative size of the City of Vancouver relative to the metro area as a whole. The City of Vancouver, in 2016, reached just over 630,000 people. But the metro area (CMA) of Greater Vancouver is much larger, reaching nearly 2.5 million people. In other words, the City of Vancouver is at the centre of a fragmented metro area much larger than itself. This is quite different from the situation of many other central cities where their boundaries contain large portions of their own suburbs (like Edmonton or even Toronto), obscuring the mobility of the young. Ideally we’d want to compare Vancouver with other metropolitan areas with central cities of about the same size to get a feel for how weird we might be.

Cascadia to the rescue! Let’s look to our sister cities to the south. The City of Portland’s population, at just over 610,000 in 2015, is quite comparable to the City of Vancouver. The Metro population of Portland, at just over 2.3 million, is also nearly the same. For good measure, let’s also throw in Seattle, with a city of similar size (just over 650,000), but a significantly larger metro area (3.6 million).

So how does the proportion of various age groups residing in the eponymous central cities compare across these three metropolitan areas?


There’s definitely a pattern here. Central university cities within much larger metro areas tend to concentrate young adults. They also concentrate the elderly, though to a lesser extent. But they consistently lose those in middle age (and their school-aged children) to the more suburban parts of the metropolitan area.

Overall Vancouver looks a lot like Portland and Seattle to me. This despite quite different housing markets and housing stock. If anything, what stands out is that the City of Vancouver seems to be retaining MORE of its school-aged children than its similarly positioned neighbours to the south. And it’s also attracting a greater proportion of the elderly.

What else have I learned? Cascadian comparisons are actually pretty cool! Bring on the Cascadia Urban Analytics Cooperative!

My award-winning book wins an award!

The Death and Life of the Single-Family House just won the Canadian Sociological Association‘s John Porter Tradition of Excellent Book Award. I’m truly honoured by this award, especially by the company it allows me and my book to keep! My thanks go out to the review committee, and also to my very supportive team at Temple University Press., especially the editors of the Urban Life, Landscape, and Policy series.

I’m joined by three other members of the UBC Sociology Department on the list of awards handed out by CSA this year, including my good friend Sean Lauer who won the Angus Reid Applied Sociology award as a Practitioner. Incidentally, he and fellow faculty member Carrie Yodanis also have a book out this year, called Getting Married: The Public Nature of Our Private Relationships. Two of our graduate students, François Lachapelle and Patrick Burnett also won an award for their paper, “Canadianization Movement, American Imperialism, and Scholastic Stratification: Professorial Evidence from 1977 to 2017.” To be sure, this is important stuff (and I say that as an American immigrant to Canada and beneficiary of the processes described).

But back to the company my book gets to keep! Through the CSA Awards, I’m excited to discover Dalhousie Prof. Karen Foster‘s new book, Productivity and Prosperity: A Sociological History of Productivist Thought, published by the great team at Univ. of Toronto Press. (I’m also looking forward to reading their recently published Gentrifier, but that’s another story). Prof. Foster’s work looks right up my alley in terms of trying to get at how economic concepts like “productivity” get measured and talked about in ways that are both socially constructed (often in a problematic fashion) and also highly consequential in terms of how they shape public policy. Really cool stuff – I’m looking forward to reading it!

The list of past award winners of the John Porter Award also places my book in brilliant company. I’ve been longing to get a copy of Vic Satzewich‘s book, Points of Entry: How Canada’s Immigration Officers Decide Who Gets In, from the excellent UBC Press, for quite some time now. This is an area of real interest to me (for instance, I regularly have students read some of the training documents for how immigration officers determine which marriages are “real”). Prof. Satzewich’s influential work has been cited by Canadian policy-makers in recent leadership debates (if wrongly), and it’s another I’m really looking forward to reading.

The rest of the list is equally as brilliant (Lesley Wood’s Direct Action, Deliberation, and Diffusion ; Elke Winter’s Us, Them, and Others ; Andrea Doucet’s Do Men Mother? ; Kay Anderson’s Vancouver’s Chinatown – which I cite in my own book! – and many more besides! See the list to fill out your own bedside table).


Now I’ll go back to pettier concerns, like listening to myself repeat the phrase “my award-winning book” over and over again, and envisioning how it will look the next time I revise my CV and find myself with something lovely to plant in the vast, otherwise empty landscape of its “Awards” section.


Good Age-Specific Net Migration Estimates Come in Threes!

Recently I posted on how we’re still not seeing any big age-specific losses in net migration figures in Metro Vancouver following the release of 2016 Census data. To summarize, there is STILL no flight of the millennials, BUT maybe there’s a slow leak of the Baby Boomers, which might be seen as evidence of “cashing out” of the local real estate market.

Today I wanted to provide both some metropolitan comparisons to note how Vancouver’s patterns fit with a couple of similar places, and also some municipal comparisons within the Metro Vancouver area. I also wanted to make some technical adjustments in how I modeled mortality* as I aged people through the past five years to estimate net migration, which really matters for older adults (not so much for the young). Again, I’m using 2011 and 2016 age distributions drawn from census profiles to get at age-specific net migration estimates for each of the metro areas and municipalities below.

First let’s compare Vancouver as a metropolitan area to two other metro areas: Edmonton and Toronto. I like this comparison primarily because Vancouver is nestled nicely between these two areas in terms of size, and they’re all big university towns.


For Vancouver, you may notice that the figure looks very similar to what I posted two days ago, up until you get to folks in the 70s and above. That’s where mortality effects really start to matter! I think the above is a better approximation of those effects, but it’s tricky to get them right.

Comparing Vancouver to Toronto and Edmonton, what stands out most for me is just how similar these three metropolitan areas look! Metro Edmonton has grown faster over the last five years in % growth terms, but age-wise, the basic pattern of growth is the same as in Metro Vancouver or Metro Toronto. Young people (including Millennials) pour into all three of these areas, and then mostly stick around.

I noted in Vancouver there was new evidence (at least new to me) of a slow leak of Baby Boomers over the last five years. It appears this leak is also showing up in Metro Toronto, with a very similar pattern. It appears there are fewer folks in their late fifties and sixties than might be expected, suggesting they’re leaving town (cashing out?). Then people in their seventies and above start returning (probably for the good health care & related facilities).

There is also a later-life leak of Metro Edmontonians, but it starts later and never quite stops until the latest age. This could reflect more of a straightforward retirement and return home effect for the many folks drawn to the region, but it’s hard to say. At any rate, all later life migration patterns are dwarfed by the influx of younger adults (and their children) into these growing regions. I don’t see a lot of cause for concern about any particular age-groups shying away from our rapidly growing metro areas.

What about within Vancouver’s metro area? I’m somewhat ambivalent about emphasizing municipal differences in age-specific net migration patterns insofar as metropolitan areas tend to be tightly integrated. When a group disproportionately moves over the border from one municipality to another, it doesn’t have a big impact on the vitality of the region as a whole. Nevertheless, it’s worth tracking, and it certainly can have big implications for quite local livability, diversity, development, and transportation questions.

Here I’m just going to compare Vancouver and Surrey, the Lower Mainland’s biggest two municipalities, with Maple Ridge, a smaller suburb further out.


Here you really get a sense of how tightly connected central cities and their suburbs can be. As the region’s central city (and biggest university town), Vancouver receives an ENORMOUS influx of young people. Then, as they move into their thirties (and often start having children of their own), they tend to move out again, slowly leaking out of the City thereafter. Nevertheless, so many young adults move to the City of Vancouver that they overwhelm the later leavers. In net terms, the majority of young adult arrivals stick around in the City of Vancouver all through their later lives.

But back to the leavers – where do they go when they leave? Mostly to the suburbs. Maple Ridge is the City of Vancouver’s mirror image in this regard. People in their thirties and beyond account for most of this suburban municipality’s growth. By contrast, young adults, especially of university age, but extending into the twenties, flee Maple Ridge. Where are they going? (see above).

What about Surrey? It’s still a suburb, but also increasingly a centre of action in its own right within a multi-polar metropolis. At the moment it’s hit a sort of demographic sweet spot where it’s gaining people at all ages. Nevertheless, it’s worth noting that while young adults aren’t exactly fleeing Surrey, their contribution to its growth isn’t as strong as for older adults or their children, and it remains nowhere near as strong as what we see in the City of Vancouver.

On the whole, these net migration patterns are not too surprising for a relatively large metropolitan area. Young people tend to leave home and move toward the vibrant city centre. Later they tend to move back to the suburbs as they settle down and start families of their own. If anything, what’s striking here is just how many young people remain in the City of Vancouver as they age, living on their own or in diverse families across a wide array of the different housing options the City is working to provide – if still, typically, at too great an expense!



*- my mortality modeling from my earlier post was really crude – simply applying five years of the expected death rate to the starting (2011) population. Bad demographer, bad! Now I’m using BC Deaths data to apply a survival rate and age the population from 2011 year by year, for each of the past five years, allowing one-fifth of the population in any given age group to age to move to the next mortality risk with each year and then applying the survival rates to the surviving population in sequence. This still doesn’t account for the mortality of recent migrants (in other words, recent arrivals could die and never be counted by the census, and I don’t take into account their mortality in any separate fashion – if I did it would boost the net migration estimates, especially for older adults). I’m also twiddling a bit with my estimates for 0-4 year olds and 85+ year olds, as needed by modeling (infant mortality is much higher than any year afterward until quite late in life, and after 85 I’m dividing the population into about half experiencing 85-89 vs. 90+ mortality). But I think I’ve got most of the technical details now closer to realistic for estimation purposes. As noted previously, none of this really matters much for younger population groups.


Update: The Lifeblood of Vancouver still isn’t leaving!

New Census results out today from the 2016 Census! They include dwelling type, age, and sex figures. The former is of great interest to me, but I’m going to concentrate on the latter just to update my older posts on migration patterns for Metro Vancouver.

Behold, the lifeblood of Vancouver still isn’t leaving!


I followed the same basic procedure here as I described in previous posts, comparing 5-year age groups across 5-year census periods. For example, given how many 20-24 year olds we had in Vancouver in 2011, how many 25-29 year olds would be expect to be here in 2016? Without any net migration, we’d expect roughly the same number, subtracting a few who died. So if we compare population figures, and make minor adjustments for mortality (I used 2013 figures, drop me a line for details [UPDATE: I think I’ve made better technical assumptions about late life mortality effects in this later post, reducing net migration estimates from age 70+]), then we can estimate net migration by how many more (or less) people show up in 2016 than we’d expect. I use the intervening age intervals (e.g. 23-27 year olds) as labels to demonstrate where most of the in-out movement is taking place between census years, which I find really captures, for instance, those university years (18-22) well.

The big takeaway, given the frequent concerns expressed over millennials leaving Vancouver,* is that it’s STILL NOT HAPPENING! Young people continue to pour into the region (University town, vibrant urban scene, etc.), and they tend to stay well into their forties.

What does appear to be new this year, at least according to my calculations (which are heavily dependent upon mortality assumptions as the population gets older), is that we’re starting to see a net loss of our late-career / early retirees. These are folks in their fifties and sixties. Yes, yes, the slow leak of our Baby Boomers is upon us! Apocalypse Now! (to be fair, it is their movie…) It’s quite possible these are predominantly people cashing out on their investments in the local real estate market and leaving for elsewhere. But if so, that’s about the only age-specific migration trend I’m seeing that seems driven by Vancouver’s widely unaffordable real estate.

*- I’ve still not seen any calculations or corrections on this issue from Bloomberg. Show your work! Tell us where your bad data is coming from! It’s ok to get stuff wrong, but not ok to keep false stories running!