Here is a recent report by BMO regarding housing prices nationally.
The media and some analysts remain glued to the idea that the Canadian housing market is a bubble ready to implode à la the U.S., Ireland, Britain, and Spain, where prices have dropped 22% on average. The latest concerns were fanned by an article in The Economist that suggests the Canadian market is 24% overvalued based on a comparison of prices and rent. Rent captures the stream of income derived from housing, and thus, in theory, is an appropriate valuation metric. However, in practice, it is important to note that statistical agencies often have difficulty measuring effective rents. Moreover, the downward trend in interest rates in the past three decades would normally favour ownership over renting and, hence, encourage a higher price-to-rent ratio today than in the past (Chart 1). Perhaps for these reasons, of the 20 countries reviewed by The Economist, the median overvaluation was 20%, and only four countries were overvalued by less than 10%, casting some doubt on the appropriateness of this metric.
In our view, comparing house prices with personal income, rather than rent, provides a superior methodology. As shown in Chart 2, the historical trend for this ratio has a zero slope, meaning prices tend to rise alongside incomes over time. This makes sense, since a family’s income largely determines how much it can pay for a house. For ease of comparison, we rebased the ratio such that the long-term trend is 100. One can interpret this trend as the ‘fair’ value of housing relative to income. Comparing the current ratio with the long-term trend suggests Canadian house prices hit a peak overvaluation of 18% in late 2009. This raises some concern because it suggests the market is only slightly less overpriced than in the late 1980s, or compared with the U.S. in 2006. However, a 3% decline in (seasonally-adjusted) prices so far this year, coupled with continued moderate income growth, has reduced the estimated degree of overvaluation to a less worrisome 11% in 2010Q3.
Besides income, interest rates determine housing affordability and thus influence demand and prices. Chart 2 shows that over long cycles, the price-to-income ratio has deviated significantly from its long-run trend. An important reason for this is that the fair value derived from the long-run price-to-income ratio for housing can no more be considered in isolation of interest rates than can the fair value derived from the trend price-toearnings ratio for equities. When real interest rates a revery low, as they are now and have been since late 2008, house prices would be expected to rise relative to personal income, to levels above trend. And, similarly, when real interest rates are very high, as they were in the early 1980s and early 1990s, house prices would be expected to fall to levels below trend.
Because of historically low mortgage rates, affordability remains reasonable today. Even with the notable rise in house prices during the past few years, mortgage service costs (including principal) for typical buyers of an average-priced house, are running close to long-term norms of 34%. True, if mortgage rates shot up three percentage points overnight, affordability would weaken substantially, and demand and prices would likely fall too. However, given the importance of fixedterm loans in the financing of mortgages, average borrowing costs do not fluctuate as much as market rates (Chart 1). Even when market rates normalize, average borrowing costs are unlikely to escalate materially from current levels as more customers lock into fixed-term contracts at relatively low rates. In addition, given substantial excess capacity, mild inflation and weak economic recoveries in the advanced nations, the normalization of interest rates could take several years, with Canadian interest rates rising a moderate 2-to-3 percentage points. By then, incomes should catch up to prices.
In assessing any potential correction in house prices from current levels, it is instructive to examine the severe price declines in the early 1980s and early 1990s, aggravated by very high interest rates (real prime rates in double-digits) and high unemployment rates (averaging 11.2% in 1982-1985 and 10.8% in 1991-1994). In both periods, the peak-to-trough decline in prices was 13% nationwide, although there were larger adjustments in some regions that previously experienced boomlike conditions. By comparison, since the beginning of 2009, real prime rates have averaged just 0.6% and the unemployment rate 8.2%. Given our view that economic growth will continue at a moderate pace in coming years, with gradual declines in the unemployment rate and only moderate increases in interest rates, the adjustment in the ratio of Canadian house prices to personal income toward its long-term trend will likely involve just a modest decline in prices coupled with a continuing increase in personal incomes. The price decline will likely be limited to a further 5%. The fact that housing demand and supply are currently balanced also weighs against a sharp drop in prices (Chart 3).
Of course, all housing markets are local. Of the three provinces that we examined (using available annual data and estimates for personal income in 2009 and 2010), British Columbia is the most overvalued, with the price-to-income ratio 17% above the trend line in 2010 (Chart 4). Note the price-to-income ratio has trended up since the early 1980s, reflecting the influence of pricey Vancouver, where a limited amount of space has put upward pressure on land prices, and where wealthy foreigners are supporting the market. Unlike in other provinces, B.C. house prices have continued to trend higher in 2010, reaching new peaks in September. Consequently, it likely faces a greater downside risk than other provinces. Valuations in Alberta and Ontario are closer to the current national estimate. In Alberta, falling prices and rising incomes have reversed nearly half of the 20% overvaluation in 2007. In Ontario, current overvaluation is about half that of the late 1980s—now THAT was a bubble!
While metrics such as the ratio of house prices to personal income or rent provide useful guidelines, they must be carefully interpreted. Such ratios are not independent of the level of interest rates and can be subject to measurement error. All things considered, the Canadian housing market does not appear to be in a bubble, and is unlikely to suffer a U.S.-style collapse. A key and over-riding difference is the quality of loan origination in the past decade, as well as other institutional factors such as mortgage insurance and recourse against defaulters. U.S. house prices were driven by a major deterioration in lending standards that flooded the market with buyers who simply could not afford their homes, leading to a vicious spiral of rising foreclosures and falling prices. This is simply not the case in Canada, where mortgage delinquency rates have remained low even during the recession.
The Bottom Line: A comparison of house prices with personal income suggests that Canada’s housing market is moderately overvalued. After doubling in the past decade, prices are now adjusting lower in response to less pent-up demand and relatively high household debt. The adjustment means that, unlike in the past decade, housing will probably not act as a tailwind for the economy, and investors seeking continued strong returns might want to look elsewhere. However, prices are not so out of line with income that the market cannot adjust through further modest declines in prices and increases in income. Though overpriced, the absence of widespread speculation and egregiously loose credit standards suggests the market is not in a bubble. Instead, Canada’s housing market remains reasonably affordable because of exceptionally low interest rates. Barring a sharp spike in mortgage rates or a relapse into recession, a substantial price correction is unlikely to occur. The greater risk could be that sustained low interest rates might recharge the housing market and inflate a true bubble that ultimately bursts when rates normalize.