01 Fnance Mnsleve ues Finances Cam Camfi MEMORANDUM DE SERVICE UNCLASSIFIED A--Mkloc IO . Paul Rochon Walsh W4 Jmanmnvois remult A0525 INS $75? Elfin-ling the flouting Stuck in Alberta Ind the Rink n'OverhuMin: 11Iis memo is fur infomulimL it estimues whether there is rn ovenu-pply ofhoms in Alheu us well in testing the Emmi-1 for uvcrbuilding under Iwu mul-tion outflow maxi Key Points . We hrve developed housing stock model for Alhetn The model slim-la the ctnrent stock ofhernee which an then he eonrpered to the slack required given the nuinher of - minutes orthe homing stock end number othouoeholds rot Alhertn ill/slime the excess olhomes in Alberta expmded rnpidly between 2001 Ind ml 1. However, this axes: supply her tinoe receded and overull supply rel-rive to reqtrireinein is now in line with historiul nmms. - Nonetheless. we believe that a decline in and menu] ofimapmvincinl ring-don -- the musl impomm source qulbelu's guwllt over the put sevch yen: A could create Wm: Ind the mnylitud: ufneaicd ndyurnnents in housing activity and prices, Ally housing market weaknesses nssocinted with tn economic decline Ind populnion outflow would extend beyond housing eonmretion. Oil-w, sin-ller fleets ofthc Canadfl m1 Kind- 21(Ura) 5 21mm, v1housing market, such as services associated with selling a new home like land development and processing mortgage insurance, would also weigh on economic growth. The impact from less spending on housing services would be relatively small compared to housing construction from 2008 to 2009 in Alberta the value of investment in new homes declined by $2.5 billion, compared to under $450 million for acquisition costs) but would still create further unwelcome economic drag. a While weaker housing activity will affect prices, we believe that the potential for a signi?cant price shock in Alberta will still largely be determined by labour market conditions. However, up to May 2015 arrears rates in Alberta have remained near their lowest point since before the 2008-09 recession. 0' Labour market weaknesses are especially relevant to Alberta?s housing market than other provinces? since there is generally no legal recourse for uninsured mortgages with low loan-to-value levels in Alberta, in contrast to much of the rest of Canada. These non-recourse mortgages could create incentives for some homeowners facing an income shock to pursue a ?strategic default?, and thus place further downward pressure on prices. One mitigating factor could be the rapid rise in Alberta?s transient worker population working for a part of the month in Alberta but residing elsewhere) living in temporary residences such as work camps, and so are mostly not part of either the rental or owner- occupied markets. Labour market shocks concentrated on these workers would therefore not place further downward pressures on Alberta?s housing market as these transient workers are not a part of either the rental or owner?occupied housing market, this would limit the negative impacts on the local housing market. - Looking forward, we will re?ne our estimates of the excess supply of homes as more data become available. 000002 Document released under the Access to Information Act _3 Document divulgu? en vertu de la Loi sur l'acces l?information Background Current Alberta Housing Market Conditions I Alberta?s housing market was strong up until October 2014, when prices and housing re- sales began to fall. Sales volumes fell by 34% to its trough in February, while prices declined by 6% from their peak. 0 The housing market has stabilized since March 2015. Despite these improvements, sales are still down by 20% ?'om October 2014 while prices are down by almost Future labour market and economic developments will determine when, and to what extent, the housing market recovers. 0 Residential investment in Alberta was relatively slow to respond to economic weaknesses, with housing starts in the ?rst quarter of 2.01 5 being their strongest since the 2008-09 recession. The second quarter, on the other hand, was the weakest in alrnost two years as builders, especially those in Calgary, responded to weaker economic prospects. Estimating the Supply and Demand for Housing An oversupply of homes can be estimated by observing the gap between the number of homes and the number of households (or the number of homes necessary to ful?ll ?demographic requirements?). There will always be more homes than households as a result of new homes being on the market, homes sitting vacant between moves, or apartments waiting for new tenants. This is a normal mpect of a well-functioning housing market, and helps avoid rigidities similar to, for example, Quebec?s July 1" ?moving day?. However, episodes of signi?cant overbuilding arise if the number homes available increases considerably beyond current or expected rates of growth in the number of households. The relative balance between the stock ofhomcs and number of households is an important indicator of the strength of the residential investment sector and for price movements. A surge in the construction of homes relative to population growth, for example, could indicate that residential investment will soon weaken or that there will be downward pressure on house prices. Unfortunately, of?cial and timely data on the housing stock and the number of households are not readily available. Statistics Canada stopped publishing data on the housing stock in 2000. However, a ?housing stock model? can be used to estimate the size of the housing stock since then, using more recent data on completions of new homes, conversions of other properties hotels, barns, warehouses, or gymnasia) into homes, and demolitions of existing homes. The number of households in Alberta is also not readily available in non-census years (the last census being the 2011 one). Nonetheless, using Statistics Canada?s population estimates by age, and census-based estimates of household formation by age group, we can approximate the current number of households in Alberta (see Annex for ?rrther details on this model). 000003 Document released under the Access to Information Act -4 Document divulgue en vertU'de Loi sur [scores a I?I?nformation Current Extent ovaerbuilding Based on this methodology, it appears that Alberta?s housing market went through a period of over-building from 2001 to 2011, with the excess supply of homes increasing from 2.6% to or almost 47,000 units (Figure 1). Much of this period coincided with the rapid expansion of activity in the oil sands sector, and a sharp rise in incomes and population in the province. However, the 2008-09 recession resulted in a major housing market decline in Alberta, with sales falling by 60% while prices fell by almost 13%, which was likely intensi?ed by the large excess supply of homes. Since 2011 though, the excess supply of homes has been falling and was at 2.0% (approximately 30,000 units on a total estimated stock of 1.56 million units) in 2014. This is below the longer- run 1976-2001 average ?excess? supply of The excess supply of homes has decreased since 2011 as a result of stronger population growth which has been suf?cient to absorb the continued expansion in the housing supply (Figure 2). The reduction in the excess supply of homes ?om 201 1 to 2014 in Alberta has been re?ected by a stronger real estate market. Average home prices in Alberta increased by an annual average of 4.3% ?'om 2012 to 2014 (compared to an average annual decline of 0.25% over the previous three years) while the sales-to-new listings ratio also became tighter, rising from 0.51 in 2011 to 0.64 in 2014. Apartment vacancy rates have also declined in Alberta?s major cities since 201 1, with vacancy rates in April 2015 being 2.4% in Edmonton and 3.2% in Calgary, down from 4.7% and 3.4% respectively in April 201 1. Moreover, the latest Home Price Analysis and Framework also found that both Calgary and Edmonton are at low risk of overbuilding?. Sources of Population Loss in Alberta As shown in Figure 3, inter-provincial migration has been the main contributor to population growth in 2013 and 2014, and so is a major source of demand for housing. Inter-provincial migration can be volatile, though, and has also been a major source of population loss during the 19803 and 2008-09 recessions. Therefore, a potential out-migration of current Alberta residents to other provinces is likely to be the primary source of any population declines in Alberta during the current downturn and therefore a key factor underpinning the evolution of the housing market. Many of the out- migrants ?orn Alberta could be recent arrivals ?'om other provinces who may be more likely to leave during an economic downturn than more established Albertans. Inter-provincial migration is more likely to be a source of population out?ows than international migration, since there are fewer barriers to inter-provincial than international migration, with less planning or administrative approval being required. I Housing Now Quarterly Supplement, August, 2015. 000004 -5- Figure 1: Excess Supp? Housing Stock Rda?ve to the qu?er of?eesdeolNewman-m1de1975 1900 1985 1990 1995 2000 ZEUS 2010 2015 6 Households Based on Census Data I Households from Internal Calculations - Average, ma [616053; Figure 2: Annualized Growth in the Housing Stock and Number ofHoyseholds ?any; .. 11. 5 Year periods 1 Year periods 35% 3.091 2.596 - 2.0564 1.5% . . 0.5% -- 0.0% 1 200610 2011 2011 102012 2012 to am 2013mm? 0 Housing Stock I Number of Households Sun-cc: Su?sics Canal. CMHC. 1111mm Calmlmims 000005 40.000 30.000 1 - 20.? 1- 40.033 - 4ft: 4's? '9 '9 Document released under the Access to Information Act. -6- Document divulgue en vertu de Loi sur l?aooes 2 [information Figure 3.- Sources of Annual Populqrion Grown: in Alberta individuals -205?) . .. 4? 43? a" all of" as? ?Net hter-hovlndalm'qation Sm: Statist Cad-.hdealculalims Scenariosfor Population Losses population 4,000,000 wanna) lacunae as ?f?g 5?s? 5 ?y Residents -Ponulltion ml Based on the 19803 and 2009 recession episodes of out-migration from Alberta we have developed two hypothetical scenarios for the current period of economic weakness: 0 Scenario 1, 2009-31er Out-migration: This scenario involves a one year out?ow equivalent to 5 per cent of the immigrants in the previous three years, followed by the resumption of previously forecasted population growth2 (Figure 4). 0 From July 2009 to June 2010, there was a net out?ow of 3,200 inter-provincial migrants that amounted to approximately 5 per cent of the net in?ow over the preceding three years. This inter-provincial out?ow was only temporary, with Alberta again becoming a destination for inter-provincial migrants in 2010. 0 Housing starts and completions in Alberta declined sharply ?cm 2007 to 2009 but stabilized and started to grow again in 20] l. 0 Scenario 2. 1980s-style Out-migration: This scenario involves a large and persistent out?ow of inter-provincial migrants over a three year period, with out?ows being equal to 50% of net?arrivals over the previous three years (Figure 5). This scenario mirrors the large inter-provincial out?ow ?om 1982 to 1985, which amounted to 50 per cent of the net- in?ow of inter-provincial migrants (or more than 64,000 people) over the preceding three years. Album The avenge of these two number-sums then used Io map growth into housing conmlaions. 000006 3 - A Document $6288-56 under the d" AVL- ,n 1? Access to mam/3?7 Act Document vertu .rrY't SUV 306635 8 mom-airs - -7- 0 Housing starts and completions fell markedly ?'om 1982 to 1984 and did not return to pre-recession levels for almost 20 years. Housing Market Response to Out?ow Scenarios The price pressures resulting from an oversupply of homes or timber slowdowns in the housing market would also be real but their magnitude would depend on the extent of economic and labour market weaknesses. 1n the most recent recessionary period, total housing market activity (de?ned as re-sales of existing units and completions of new homes) declined by approximately 29% while prices declined by approximately 12.5% their pre- recessionary peak to trough. Alberta?s 19805 recession featured a more severe decline in housing market activity, with sales and completions combined falling by over 45% while prices fell by more than a cumulative 22% over a four year period (Figure 8). The current housing market decline in Alberta has been relatively mild in comparison to these two episodes. with total housing market activity over the past year being 2% weaker than it was 000007 Document released under the Access to Information Act. 8 . Document divulgu? en vertu de la Loi sur l?acces a l?information in the year preceding, while prices have been recovering from their sudden decline in late- 2014. ,w trim ?Iii-{err-Pruvinciai Migration: 2009 Recession lntet-Prmrindal Migration: 1900: Recession individuals I indiviluils -- - n- - -- 50,000 .- 40,000 40-00" 301110,000+ -. 01- 40,000 - 3 40.000 4 . . -. ammo . .. 40.000 4.3 .9 .0 45? 49 .67 ?saggy? 40. meumm 50.000 - 50,000 45,000 . .. . 45m .. 40.000 u- 35,000 - 30,000 w_ 20,000 15,000 - 10.000 w- - - 5,000 - d9 ?9 0 0 it?? ?ffn? $66) #49 ?Compmions ?$Dlt? 'i'Ecmai Calculations. CMHC 000008 Document released under the Access to Information Act . Document divulgu? en vertu de la Loi sur l?acc?s a l?information Figure 8: Housirgg Mgr-kc! Activigr and Price Growth in Scenario 1 and Scenario 2 um Sneadoldoosnecesslon ?em mm WSW -- - aux 60,000- -- - - 15?s as 155"? mm 112,500 a 1595 50M 5% 100.000 59s as mm 37.500 - - - on 1 -59s 35.000 J?r- -. 75.? ?10? yumo-L-?1 - 40% Wva ?349" '5 3% feeffew "??im'?dmm .. . SWECMHC, cam, Imam] Calwlations 00001 0 Document released under the Access to Information Act Document divulgue en vertu de -11?acces a [information Annex: Methodology for the Housing Stock Model and Population Out?ow Scenarios The housing stock model consists of comparing the size of the housing stock with the number of households at any given time. Both these series do not exist in each year and so must be constructed using other data. This relationship can be described with the ratio HSJN., where HS. is the housing stock in any year and N. Housing Stock Up to the year 2000, HS. is Statistics Canada?s series on the number of dwelling units. Beyond 2000, _a new series for H8. must be constructed. This series de?ned as: CMP. CON. DEMQ, where: 0 is the housing stock in the previous year. This ?rst data point in this series is the end of Statistics Canada?s series on the number of dwelling units. All data points beyond then are determined within the model. 0 CMP. is the number of new housing completions in a year. This data is produced by CMHC. CON. is the number of dwellings added by converting other structures hotels, barns, etc.) into dwellings. This data is ?'om Statistics Canada. 0 DEMO. is the number of dwelling units demolished over a year. This data was acquired through a special request from Statistics Canada. Number of Households Statistics Canada?s Census presents the number of households in Canada every ?ve years with the latest estimate being in 201 1. From 2012 to 2014, Statistics Canada publishes annual estimates of the population by age group. Applying theheadship rates rate at which any person in an age group is likely to be the ?primary maintainer? of a household) in the 2011 Census to the current population breakdown by age yields the current number of households. 0 Census estimates of the population by age group differ from Statistics Canada?s annual estimates of population by age group. For example, annual estimates of the number of people aged 25 to 29 in 201 1 were approximately 10% larger than the 2011 Census estimates. Population estimates for each age group in 2012 onwards are consequently adjusted to be in line with Census estimates of population by age group. 00001 1 Document released under the Access to Information Act -12- Document divulgu? en vertu de la Loi sur I?acce?s 3 [information From 2015 to 2018, population estimates are based on the latest Statistics Canada forecasts. The highest and lowest population forecasts for Alberta by 2018 were used in this report to press!? the widest range of possible outcom. These scenarios are titled ?low growth? and hmeo?tiurlt?grth, 200] to 2006 trends? by Statistics Canada. 000012