I've recently come across some fascinating reading by York University professor Moshe Milevsky on the subject of residential real estate and personal financial risk. One of his many papers, titled Houset Allocation (yes, it is spelled with the "t"), which is available at IFID website newsletters, shows rates of return on Canadian residential real estate in 19 housing markets across the country for the 25 years ending in 2004 when the paper was written. The annual compound rates of return vary considerably - from a low of 2.74% in Edmonton to a high of 6.23% in Peterborough. Along with these are the measure of volatility, or risk, in each market, again with a wide range from 2.81% in Regina to 11.14% in Vancouver. These figures alone remind one that owning a house as an investment isn't necessarily the best, considering options which include stock markets like the TSX and the S&P 500. Would one have been better (if one had perfect knowledge of the future, of course, which we all do in retrospect ;-) having a house in Vancouver that returned 3.68% per year compounded for 25 years at 11.14% volatility, or 13.85% in the S&P 500 with 15.71% volatility or a higher 5.7% return with lower 6.23% volatility in Ottawa, where my house is?
How about the homeowner in Edmonton, facing the laggard market there for 25 years? He/she might even have thought that the tar sands would one day have to take off and cause the local house price boom that is now underway. Sitting there in 2004, how much longer should he/she wait? Maybe a job change or retirement would force a sale, preventing the gain from ever being achieved. Years ago, I bought a condo in Ottawa at the same price the original owner had paid five years earlier, then there was a bump in prices and I sold it for a 250% gain six years later, pure luck because I needed to move with my growing family into my larger house. That house bought 22 years ago, well maintained and in a good neighbourhood, has only appreciated 4.8%, i.e. below the Ottawa average. The long term can be a very long, one just doesn't know how long.
One of my relatives lost their whole investment in their house when the mill in their small town closed down and there were no buyers so they had to walk away from it. An individual asset risk of a specific house in a specific market is not a sure investment gain. That is a key point made by Mr. Milovesky in the above paper, namely that a house is a non-diversified investment. It is also likely a huge proportion of a family's total investment assets. It certainly is a big part of my net worth. If one were able to own several houses in different cities across Canada, the low correlation of price gains between the cities would allow a considerable reduction in risk. He even proposes that housing REITs be set up along the lines of commercial property REITs so that each of us could sell a part stake in our houses and buy someone else's. He says it's been done in the UK and Australia but not yet here in Canada.
As an asset class on the other hand, a house is quite un-correlated to stock markets. Ottawa residences were negatively correlated with both the TSX and the S&P 500 over his study period and so are good to hold in a portfolio. In any individual period, as opposed to the long term during which they both go up, when one goes down the other asset class goes up and this reduces risk.
Among the options Mr. Milevsky reviews for reducing housing risk for individual homeowners, only the reverse mortgage is currently available in Canada. Given the fixation of many people on the tax-free capital gain on a principal residence as being a kind of be-all, end-all justification for owning a house in Canada, the above certainly warrants some thought. I'm keeping my house but whenever the housing REIT comes along I'll be having a close look, and the reverse mortgage is a good option to know for the future too.