Population Growth and Residential Construction Activity in BC Part 3

(Warning this is an analysis post; if you want the conclusion skip to the last paragraph. The two of you who actually care about this stuff, please read on.)

This is a third post on population growth and construction activity in BC. This one attempts to round out the core numbers of household formation versus housing supply.

The first post highlighted an empirical relationship between housing starts and population growth and noted that population growth is cyclical.
The second post highlighted that, based on said empirical relationship and household size data, decreases in average household size has accounted for about 25% of residential construction activity (based on units produced) in the past 13 years.

An important point is that there can be additional dwelling formation beyond completions minus demolitions: dwellings can be formed by re-purposing a set of existing dwellings to be a larger or smaller set of dwellings, or by failing to include dwellings in the completions data (for example a "single family" dwelling is actually a 2 or 3 unit dwelling but not reported as such).

Another point is that dwelling completions need not necessarily be occupied. (Dwelling counts include actual dwelling counts as well as those "occupied by usual residents") They can remain vacant (so goes the rumour) or they can be used as vacation or second homes. It should be noted such homes can be either rented or occupied; to wit imagine someone living in Vancouver renting a suite in the Peace region for work reasons. There can be many reasons why dwellings can grow faster than household formation would suggest, and can be either sustainable or unsustainable.

This post analyses household formation, population, (estimated) demolitions and housing completions. Attempts are made to measure how accurately completions align with household formation.

The data are as follows (source: BC Stats and CANSIM 027-0008)

Year   Population Households Completions Net Completions*
2000   4039230    1572086    -           -
2012   4615096    1876404    -           -
Change  575866    304318     334599      301139

* "Net Completions" assume 10% demolition rate

This is broadly consistent with analysis performed by CMHC on the household formation/completions discrepancy in their spring 2013 housing market outlook publication (PDF). Based on this cursory analysis, once accounting for demolitions (about 2500/year) and assuming completions are the sole source of dwelling formation, completions align well with household formation.

Another dataset is the semi-decennial census dwelling and population counts (source Statistics Canada Census data):

Year   Population Dwellings Dwellings** Completions Net Completions*
2001   3908000    1643969   1534335     -           -
2011   4400000    1945365   1764637     -           -
Change  492000     301396    230302     291837      262653

** Occupied by usual residents

Here we can see that net completions are broadly in line, though slightly ahead of, "usual resident" dwelling formation. Census collection methods will do actual on-the-ground dwelling counts so would tend to capture all forms of dwelling formation. Given the uncertainty in demolition rates we don't have enough evidence to support overbuilding, however the change in the total number of dwellings indicates another source of dwelling formation beyond summing completions minus estimated demolitions.

Percentage of dwellings occupied by usual residents dropped from 93.3% in 2001 to 90.7% in 2011; this drop has added about 51,000 dwellings over the 2001-2011 census interval.

As mentioned, there are other forms of dwelling formation not included in CMHC completions data, namely unit conversions and unreported dwellings (e.g. basement suites). It is unclear how significant including these adds to dwelling supply. If such modes of dwelling formation is on the order of net thousands, that would indicate dwelling supply has outpaced household formation.

In conclusion, should CMHC-reported completions have been the sole means of dwelling formation in the province, and demolitions are accounted for, dwelling growth has been broadly consistent with household formation over the past 13 years. Census data indicate diverging dwelling units compared to those occupied by usual residents of approximately 55,000 over the period of 2001-2011, or about 19% of construction activity on a per unit measure. Once other sources of dwelling formation, such as unregistered basement suites or unit conversions, are considered and assumed a significant contributor, that would indicate that there are now relatively more dwellings in BC than there are households than was the case 13 years ago. For reasons discussed this does not necessarily mean an oversupply of dwellings.

Population Growth and Residential Construction Activity in BC Part 2

A followup to yesterday's post on population growth and residential construction activity in BC is in order when thinking about the future path of household formation, household size, and construction activity. We have the basic function of construction activity versus population growth over the past decade:

Since 2000, every person growth in population has resulted in an additional 0.33 dwelling completions. If population were to remain fixed, annual completions would be 9,500.

More interesting is to dig a bit deeper into this "baseline" 9,500 completions. Based on historical demolitions data (CANSIM 026-0012 1965-1999 **terminated**), demolitions are about (10±2)% of housing starts (and completions). What this means is that even with no population growth there is a net 8,500 annual increase in dwellings, which can only mean that household size is shrinking and/or there is oversupply.

Indeed household size has been shrinking since 2000, with average household size falling from 2.580 to 2.473. If we account for the base population's change in household size in a year, this results in a net demand averaging 6,200 per year, which is on the same order, albeit slightly smaller than the estimate of 8,500.

There are reasonable error bars on the estimates, so we can state that, after accounting for demolitions and more importantly changing household size that housing completions have broadly aligned with population growth over the past decade.

More interesting, though, is looking forward to the next 13 years (and longer). The important takeaway from this post should be that shrinking household size has added approximately 6,200 additional residential construction units per year, or about 25% of total residential construction activity, since 2000. If household size were to plateau that would mean a 23% drop in average construction activity compared to the past 13 years (or significant oversupply). BC Stats projects household size to drop from 2.53 in 2006 to 2.37 in 2035. Under this scenario, using BC stats population projections, this would mean the projected decreasing household size will add 3,400 additional residential construction units per year, which would mean, all else equal, a 13% decrease in average construction activity from the previous decade.

Another factor to consider is the utilization of existing dwellings as the household size has dropped. While I don't have data to support one way or another, it may be true that existing dwellings are not being utilized to their full capacity (think empty nesters who haven't downsized yet). If this is true that means, once turned over, these dwellings will act as "phantom" supply that would further decrease residential construction demand.

Population Growth and Residential Construction Activity in BC

(This is the first in a three part series analysing population growth and construction activity in BC. See the second and third parts.)

There is an obvious correlation between housing activity and population growth. Since the 1980s one can see the ebbs and flows of population growth, alongside housing starts and completions, have been a normal occurrence in BC's history. The second graph puts population growth in a longer historical perspective for posterity.

 

The best correlation is between housing completions and population growth, however there are two behaviours, one pre-2000 and another post-2000. Pre-2000 there were approximately 0.22 completions per person of growth, with a 1 quarter lag (ie population growth rises and falls one quarter before completions). Post-2000 there were approximately 0.33 completions per person of growth with a 3 quarter lead (ie population growth rises and falls three quarters after completions).

Some clues as to why there is a difference should be evident by observing the fraction of detached completions of the total has been dropping since the early 1990s in Vancouver that makes up the majority of residential construction activity in the province.

Notes:

• Though it's not immediately evident from the data, by visual inspection, there is arguably a "baseline" 10,000 units built each year to maintain the existing stock of housing for the existing population. That is, if population growth were zero there would still be a turnover demand of around 10,000 units per year.
• The number of completions per person of growth after 2000 corresponds to 0.33, or about 3 people per completion. Before 2000 this was 4.5 people per completion. It looks as if the shift to multi-unit construction at least partially accounts for this difference.
• As a rough rule of thumb, if population growth is 30,000 this results in annual housing demand of 20,000. Over the course of BC's history since the 1960s, 30,000 is a typical nadir. Most recent population growth has been around 40,000, which corresponds to a demand of about 22,000-25,000 units. Current completion rate is near the high end of this range but the number of starts suggests the number of completions will not markedly deviate from current population growth.
• Population projections typically do not account for the variability in year-to-year population growth but concentrate on averages. Population growth has averaged 60,000 and completions have averaged 28,000 since 1965. Growth projections are about 55,000–60,000 until 2035 based on BC Stats analysis. Population growth cycles with a period of about 10–15 years.

I track population growth closely as it aligns with residential construction activity, GDP growth, and new and used housing sales and for-sale inventory.

Vancouver CMA Intraprovincial and Interprovincial Migration

Refrains I have heard, as long as I can remember, is families migrating away from expensive cities into other regions due to high housing costs. This would make sense simply from a "diffusion" perspective where areas of growing population will invoke migration to less populated areas. To capture this effect I have plotted intraprovincial (to other parts of the province) and interprovincial (to other provinces) migration by lumped population cohorts:

In terms of migration due to high housing prices, while there may be some effect, out migration in the early 2000s was heavily negative even with prices and affordability lower than today's. Other observations:

• There has been smaller intraprovincial out-migration of the 30-44 cohort since the recession. This meant that from 2009 through 2011 there would have been additional housing demand for this cohort
• The older cohorts, including those with school-age children, tend to have net migration out of Vancouver CMA.
• The fundamental mode of interprovincial migration patterns, regardless of cohort, is on the order of 15–20 years.

International migration graphs are below, with an interesting decomposition of the 15-29 year cohort:

BC Population Projections and Cohorts

Something to consider is the projected increase in population of BC in the coming decades and what types of housing will be required. Caveats abound, not least from the projectors (BC Stats), nonetheless I have graphed their projections below by age category. I have included the "implied migration minus deaths" in the first graph, it can be seen that migration patterns do not look out of line, and note that over the next 25 years the mode will shift markedly lower due to in-migration of members of the 20-45 year old cohort, then start trending higher again.

For the purposes of housing mix discussion, for discussion only, I have included five cohorts below:

• 20-65 working age
• 35-75 family with child dependents
• 0-20 dependents
• 25-35
• 75+ downsizing

There can be debate about whether these cohorts are appropriate for their designations. I am assuming

• Family with child dependents will be seeking, and remaining in, family style housing from ages 35-75. A contention, perhaps, here is that I have extended this age to 75, if only to question how many households will indeed downsize upon retirement and how many will remain in their houses until required to move. This age could be lowered if it turns out households downsize sooner. It is unclear to me this downsizing will occur en masse as some pundits have claimed so I'll take the other tack for argument's sake. 
• The 25-35 cohort will be primarily childless, or with young children in which smaller accommodations are suitable. This cohort will demand smaller dwellings, either suites or apartments.
• There will be a baseline of families who, due to economic, family size, or lifestyle reasons, choose smaller accommodations than the standard 20th century Canadian pattern. Nonetheless these will be a baseline subset of housing mix and I am assuming there will still be significant traditional demand for dwelling size as before.

Here are the cohorts:

A few observations:

• The working age population as a percentage of total population will decrease but not substantially due primarily to immigration.
• The 35-75 cohort will continue to increase in percentage until 2020 at which point it will plateau and then start to decrease.
• The 25-35 cohort will increase until mid-decade and will decrease as a percentage of total population.
• The 75+ cohort will start increasing faster starting about 2020

These data are missing some pieces to draw any conclusions regarding future housing mix, not least the actual need for family-style housing embedded within the 35-75 cohort (eg not all people in this age range will have children), nonetheless there are some starts on the road to understanding what housing types BC will require over the coming decades, and, perhaps more importantly, what housing types it won't.

Central 1 B.C. Housing Forecast 2011-2013

Report available here (PDF). Excerpts (emphasis mine):

On the bifurcation of BC's real estate market:

A key characteristic of the post-recession housing market has been the divergent housing strength between the Lower Mainland and most other areas of the province. While the Lower Mainland-Southwest and, to a lesser extent, the Capital region had shown relatively stronger post-recession sales activity, most other regions remained at recessionary levels. The impact of low interest rates was more benefi cial for real estate markets in larger, diversifi ed economies with a higher proportion of local area buyers. In addition, employment growth was generally weaker outside of the Metro Vancouver region.

On overvaluation in the Lower Mainland:

It has become fashionable to suggest that price levels in Lower Mainland-Southwest region of the province, and particularly Greater Vancouver, are set to correct substantially due to the significant price gains in recent years and a de-linking of home prices relative to income and rental rates. Central 1 does not subscribe to this view, but does expect price gains to slow considerably over the forecast horizon. While price levels may turn lower in the near term, the annual Lower Mainland-Southwest median resale price level in 2012 is forecast to surpass 2011 by 1.4% to reach $497,000. A further gain of 3.6% is forecast in 2013 Central 1 deems a significant price correction in the Lower Mainland-Southwest to be unlikely for various reasons. First, much of the price growth in the region has been attributed to disproportionately strong demand for higher priced single-detached product in localized regions such as the west side of the City of Vancouver and Richmond. In contrast, price gains have been less substantial in other markets and product types, meaning this has not been a region-wide price surge. Moving forward, demand will likely remain stable as economic growth, albeit slow, persists and mortgage rates remain low.

In addition, speculative demand in the region remains low. The proportion of units re-sold within six months of purchase can be used a proxy for speculative activity. In theory, speculators look to gain through capital appreciation over a shorter time-frame relative to home-owner occupiers. In a period of higher speculation, which is generated by strong market activity and price gains, this proxy generally rises. However, this metric has exhibited a declining trend since early 2008, currently hovers near 2% and operates near normal levels. In contrast, this proxy surpassed 10% in the late 1980s, and was closer to 6% in 2006 when markets were overheated. The lack of excessive speculation suggests that we are unlikely to see a speculation-induced bust in pricing.

Meanwhile, price levels will be further supported by supply-side adjustments. Sales activity and the flow of new listings are positively correlated – when demand increases, new listings tend to follow in the months that follow. The opposite is also true. This reflects the tendency of sellers to capitalize on strong markets and rising prices, and sit tight when market conditions weaken. In the absence of any major shock in the economy such as a large and unexpected increase in interest rates or another recession, Central 1 expects the recent slowdown in demand to be met by declining listings activity, which will mitigate growth in standing inventory of resale product.

On population growth:

Weak population growth through 2013 will be a limiting factor for housing over the forecast horizon. The provincial population is forecast to expand at a lackluster rate of 1.1% this year, and fare only slightly better in 2012 and 2013 with 1.2% growth. The slow pace of growth will reflect a drop in the number of landed immigrants to B.C. from international markets this year and increased net outflow of residents to other provinces, primarily Alberta, in 2012 and 2013. This interprovincial net outflow reflects the stronger rebound in Alberta’s economy and improved labour market conditions.

On interest rates:

Mortgage rates will remain low and edge up beginning in the latter half of 2012 and through the remainder of the forecast horizon. This reflects a compression of bond yields, which have recently declined sharply in the U.S., Canada, and Germany during the latest round of market concerns and volatility. The U.S. central bank has stated that it expects no rate increase until mid-2013 and only then if conditions warrant.

I do not necessarily endorse this view in full; nonetheless the data presented are worthy of review. Personally I would give greater weight to price-rent ratios but to each his own. House purchases come with obligations lasting longer than to 2013. Place your bets.

Vancouver Housing-Related Statistics

This post is intended to summarise data sources that can be used to analyse the metropolitan Vancouver housing market. Many of these data sources can be used for other housing markets too. Unfortunately there is no one good source of data, however the Sauder School of Business has summarised much of the key historical data.

House Prices

Real Estate Board of Greater Vancouver House Price Index
jesse's REBGV tracking spreadsheet
Fraser Vally Real Estate Board House Price Index
Teranet National Bank House Price Index
Sauder data

Inventory and Sales

REBGV Press Releases
FVREB Press Releases and statistics
AgentWill's statistics (tracks a sub-area of the Vancouver CMA)
jesse's REBGV tracking spreadsheet

Rental

CMHC reports "Rental Market Reports — Major Centres"
Sauder data

Employment

Statscan labour force survey
BC Stats labour and income
Sauder data

Wages

GVRD city median incomes
City of Vancouver income map (2000)
City of Vancouver local area statistics
BC Stats taxfiler income tables
BC Stats labour and income

Population

BC Stats Migration

BC Stats Population Highlights
BC Stats Population Estimates
GVRD key facts
City of Vancouver statistics

Metro Vancouver housing data book

Housing and Construction

CMHC reports "Monthly Housing Statistics"
GVRD data
City of Vancouver statistics (from Census data)

Immigration

Citizenship and Immigration Canada statistics

Interest and Mortgage Rates

Sauder data
Bank of Canada data
Government of Canada 5 year bond quote

Economic Accounts
BC Stats


Economic Reports

BC Stats economic statistics
RBC Housing Trends and Affordability
TD economic reports

General

CMHC statistics
GVRD key facts
GVRD census bulletins
City of Vancouver statistics
Sauder Centre for Urban Economics and Real Estate data
BC Stats
Statistics Canada

Population report Q2 2010

The Statcan quarterly population report is released today. Get your pdf here. This is always covered breathlessly in the local media, picking out data points that confirm the well-known fact that not only does everyone WANT to live here, they actually ARE moving here. Beyond confirming our own natural attractiveness, it also apparently provides some justification for our rising house prices.

Some context.

First, net migrants from other provinces is barely over zero. In fact, we have just barely gained back what was lost from 1998-2003. In TOTAL since Q1 1998, we are up a net 6,118 bodies from interprovincial migration.

Next, immigration. Immigration is doing fine. We are at the same level in terms of number of bodies that we reached in the mid 1990s. As a percent of population, it is lower than the mid-90s.
Combine these together and throw the natural growth (births - deaths) into the mix and you get total pop growth. Here it is. For new bodies, we have a cyclical peak in the late 90s, but lower than previous cyclical peaks. As a percent of population, however, this cyclical peak was around half of previous peaks--and lower than previous troughs!

How remarkable the 2000s were for population growth can be made clear by ranking the 5-year periods from 1960 to 2009 by population growth. 2000-04 and 2005-2009 were dead last. So much for the population boom.

School Closures in Vancouver

The Vancouver School Board has announced (PDF) closures of 11 schools.

As you may be aware, trustees faced difficult financial decisions this spring, and last evening approved a budget that includes approximately $16 million in spending reductions and eliminates 137 positions. We also received a report earlier this month from a government appointed special advisor recommending we consider school closures to reduce our costs.

It is in this context that district management prepared a list of schools that trustees might consider for possible closure. The schools on the pre-notification list were selected because of a number of factors, including low enrolment, under utilization and the availability of space at neighbouring schools.

Without delving into the politics of the matter there has been a general trend towards lower public school enrollment in the past number of years. This is not to say that the school-age population is necessarily dropping -- parents may be opting for private or home schooling instead. Statistics do indicate a dropping school-age population in Vancouver but, interestingly, the number of family-sized dwellings in the city has not commensurately dropped.

Looking at recent birth rates, it is likely these school closures may be a bit premature. But beyond even that, a look at the city's available housing stock shows a strange dichotomy between dwellings designed for families with school-age children and the actual residents of these dwellings. At some point it would seem reasonable the utility of these dwelling should match the residents' needs by way of demographic shift back towards more children or re-development away from the usual family-oriented architecture.

Demographic Effects on House Prices

100 posts, 57,000 visitors and counting. Thank you to all of the readers and commenters who make blogging here a joy and good exercise for my brain.

I just finished reading a respectable article on the impact of demographics (baby boom) on interest rates, consumption, and housing prices. The article, published November 2005 by Robert Martin, is available at the US Federal Reserve website.

I won't bore you with the detailed analysis - it is extensive and barely understandable for a layman (including me) - but essentially the outcome of his analysis is that US home prices are affected dramatically by demographics because of the effects on productivity, consumption, and interest rates. The conclusion of the study is that US house prices are due to fall approximately 30% in real terms over the next 50 years.

To quote (on pages 26 and 27 ):

"The model gives a gloomy view of house prices going forward. . . . In the near term, house prices will peak in level terms sometime between 2005 and 2010, the exact date of the peak turns out to be sensitive to the calibration. . . . Indeed the date of the collapse itself . . . can be moved as early as 2008. . . . Following the peak, (real) house prices decline over 30 percent in value over the next 50 years. While this number seems quite large, it must be placed in perspective. Real house prices fell almost 20 percent in the three years following the 1979 peak. Real house prices also fell around 30 percent in the United Kingdom following their 1989 peak.

Will this decline in house prices occur? There are several channels through which the results of the model could give a misleading prediction of future prices. The first and perhaps most unlikely is that the rate of productivity growth in consumption goods picks up sufficiently relative to productivity in housing output to offset the decline in labor input. Second, and more likely, is that the model assumes that agents over the age of 65 have zero productivity. This is an unrealistic assumption. Agents over the age of 65 can work and if they chose to do so in mass numbers then the timing of the downturn might be quite different. I would note that they are most likely not as productive as younger workers and almost every body agrees that at some point they will no longer be productive. Keep in mind, that while 30 percent seems fantastically large, I will show data for Japan in which the model exactly replicates the 34 percent fall in Japanese real house prices over 15 years." (look at the study for further details)

To comment on the study, it is probably the most elegant study on house prices I have ever read. It comes from a respectable source and is not likely influenced by outside interests. It is purely academic but profound in the study model's conclusions.

What do you think?