Canadian Real Estate Prices Will Fall 28% By 2020 According To This Model

Posted by Ricky Dhadwal on Monday, July 31st, 2017 at 10:13am.

We get a few dozen emails weekly asking where we think Canadian real estate prices are going to be in the next few years. We can’t see the future (yet), but we can help you understand some of the financial models used to hypothesize these things. Today we’ll use the Organisation for Economic Co-operation and Development (OECD) House Price-To-Rent Index, and a linear regression model. This method predicts Canadian real estate prices will fall 28% by 2020.

OECD House Price-To-Rent Index

The House Price-To-Rent Index is a measure that compares the cost of ownership to the price of renting. If your rental yield is high, why would you sell your property for a discount? Likewise, if you can rent a property much cheaper, why would you pay more to buy it? Rent is a pretty good indicator of value, especially in large cities. If you’re thinking but “I’m not going to rent the property out!,” think of yourself as your own tenant.

To get the number, they take the cost of renting annually, and compare it to the cost of carrying a mortgage for a year. When the index shows a breakaway from the norm, home prices will come down or rents will come up. Rents are limited to income growth of the population however, so it’s hard to raise rent quickly. It happens, but it’s less likely unless the rest of the economy is in hyperdrive and pay is quickly rising.

Regression Modeling

Before we get to the numbers, let’s quickly run through what is a regression model. When a variable hits an extreme high or low, it will then move progressively closer to the mean line. A regression model is one that theorizes where that mean line will be. Like all forms of statistical analysis, it’s definitely not 100% accurate. It does provide a educated baseline to figure out where things are heading.

In matters of money however, it tends to be more accurate. Jason Zweig, an editor at the Wall Street Journal once famously said “regression to the mean is the most powerful law in financial physics. Periods of above-average performance are inevitably followed by below-average returns, and bad times inevitably set the stage for surprisingly good performance.” Whether this is a self-fulfilling prophecy is up for debate, but the point is it’s worth running the numbers.

Canadian Real Estate Prices To Drop 28%

There’s a few models we could be using, some are much more complicated than others. Today we’ll stick to your basic trendline using exponential linear regression. Using this model, and using the OECD index numbers, we should see prices drop by 28% by 2020.

Canadian House Price-To-Rent Index (w/ Exponential Regression Trendline)
QuarterIndex
Apr 1, 1970 76.01
Jul 1, 1970 78.22
Oct 1, 1970 77.59
Jan 1, 1971 76.99
Apr 1, 1971 75.83
Jul 1, 1971 75.74
Oct 1, 1971 73.89
Jan 1, 1972 74.16
Apr 1, 1972 72.41
Jul 1, 1972 71.22
Oct 1, 1972 71.38
Jan 1, 1973 70.49
Apr 1, 1973 70.84
Jul 1, 1973 71.14
Oct 1, 1973 75.07
Jan 1, 1974 77.16
Apr 1, 1974 80.92
Jul 1, 1974 84.52
Oct 1, 1974 79.88
Jan 1, 1975 75.83
Apr 1, 1975 76.94
Jul 1, 1975 77.68
Oct 1, 1975 77.18
Jan 1, 1976 77.4
Apr 1, 1976 78.88
Jul 1, 1976 77.03
Oct 1, 1976 80.01
Jan 1, 1977 78.33
Apr 1, 1977 77.56
Jul 1, 1977 75.24
Oct 1, 1977 75.61
Jan 1, 1978 74.07
Apr 1, 1978 69.67
Jul 1, 1978 71.42
Oct 1, 1978 71.16
Jan 1, 1979 70.92
Apr 1, 1979 72.67
Jul 1, 1979 71.49
Oct 1, 1979 71.03
Jan 1, 1980 70.4
Apr 1, 1980 70.19
Jul 1, 1980 71.76
Oct 1, 1980 73.7
Jan 1, 1981 74.2
Apr 1, 1981 74.51
Jul 1, 1981 73.09
Oct 1, 1981 73.72
Jan 1, 1982 72.31
Apr 1, 1982 66.18
Jul 1, 1982 62.69
Oct 1, 1982 61.73
Jan 1, 1983 63.04
Apr 1, 1983 66.59
Jul 1, 1983 66.62
Oct 1, 1983 62.12
Jan 1, 1984 62.43
Apr 1, 1984 61.29
Jul 1, 1984 59.53
Oct 1, 1984 58.82
Jan 1, 1985 57.9
Apr 1, 1985 57.7
Jul 1, 1985 57.11
Oct 1, 1985 59.09
Jan 1, 1986 59.95
Apr 1, 1986 61.11
Jul 1, 1986 64.1
Oct 1, 1986 66.71
Jan 1, 1987 69.16
Apr 1, 1987 71.64
Jul 1, 1987 71.81
Oct 1, 1987 72.22
Jan 1, 1988 73.41
Apr 1, 1988 77.89
Jul 1, 1988 78.55
Oct 1, 1988 78.7
Jan 1, 1989 81.13
Apr 1, 1989 86.05
Jul 1, 1989 79.45
Oct 1, 1989 83.66
Jan 1, 1990 85.76
Apr 1, 1990 80.06
Jul 1, 1990 78.39
Oct 1, 1990 78.55
Jan 1, 1991 76.18
Apr 1, 1991 78.09
Jul 1, 1991 83.58
Oct 1, 1991 79.1
Jan 1, 1992 78.54
Apr 1, 1992 80.06
Jul 1, 1992 80.14
Oct 1, 1992 79.64
Jan 1, 1993 81.95
Apr 1, 1993 81.34
Jul 1, 1993 78.47
Oct 1, 1993 80.58
Jan 1, 1994 81.12
Apr 1, 1994 83.66
Jul 1, 1994 82.34
Oct 1, 1994 82.6
Jan 1, 1995 81.84
Apr 1, 1995 78.95
Jul 1, 1995 75.89
Oct 1, 1995 77.9
Jan 1, 1996 77.37
Apr 1, 1996 76.48
Jul 1, 1996 77.3
Oct 1, 1996 77.21
Jan 1, 1997 76.7
Apr 1, 1997 77.46
Jul 1, 1997 77.79
Oct 1, 1997 76.12
Jan 1, 1998 75.74
Apr 1, 1998 74.59
Jul 1, 1998 73.9
Oct 1, 1998 72.33
Jan 1, 1999 72.19
Apr 1, 1999 72.59
Jul 1, 1999 72.04
Oct 1, 1999 71.65
Jan 1, 2000 71.71
Apr 1, 2000 71.46
Jul 1, 2000 71.36
Oct 1, 2000 70.76
Jan 1, 2001 70.26
Apr 1, 2001 70.16
Jul 1, 2001 72.61
Oct 1, 2001 72.28
Jan 1, 2002 72.79
Apr 1, 2002 73.02
Jul 1, 2002 75.5
Oct 1, 2002 76.27
Jan 1, 2003 76.93
Apr 1, 2003 78.1
Jul 1, 2003 78.51
Oct 1, 2003 80.37
Jan 1, 2004 81.7
Apr 1, 2004 82.38
Jul 1, 2004 82.31
Oct 1, 2004 83.05
Jan 1, 2005 83.69
Apr 1, 2005 86.11
Jul 1, 2005 86.27
Oct 1, 2005 86.03
Jan 1, 2006 87.36
Apr 1, 2006 87.72
Jul 1, 2006 91.27
Oct 1, 2006 92.74
Jan 1, 2007 93.59
Apr 1, 2007 92.82
Jul 1, 2007 97.15
Oct 1, 2007 99.71
Jan 1, 2008 99.96
Apr 1, 2008 99.34
Jul 1, 2008 99.13
Oct 1, 2008 98.09
Jan 1, 2009 96.76
Apr 1, 2009 95.13
Jul 1, 2009 93.02
Oct 1, 2009 94.42
Jan 1, 2010 97.73
Apr 1, 2010 98.95
Jul 1, 2010 101.68
Oct 1, 2010 100.74
Jan 1, 2011 98.64
Apr 1, 2011 99.82
Jul 1, 2011 101.11
Oct 1, 2011 102.48
Jan 1, 2012 102.56
Apr 1, 2012 102.95
Jul 1, 2012 104.24
Oct 1, 2012 103.66
Jan 1, 2013 103.48
Apr 1, 2013 102.52
Jul 1, 2013 103.28
Oct 1, 2013 103.31
Jan 1, 2014 103.97
Apr 1, 2014 105.95
Jul 1, 2014 105.48
Oct 1, 2014 105.82
Jan 1, 2015 106.73
Apr 1, 2015 107.61
Jul 1, 2015 106.71
Oct 1, 2015 107.87
Jan 1, 2016 108.75
Apr 1, 2016 111.25
Jul 1, 2016 114.69
Oct 1, 2016 117.64
Jan 1, 2017 119.27
Apr 1, 2017 136.4
Jul 1, 2017 141.3
 

Source: OECD, Author’s Calculations.

You won’t necessarily see a 28% reduction in sticker value. Inflation will likely decrease the value of money, and reduce the drop by 2% per year if the Bank of Canada can consistently hit their target going forward. It is a pretty big drop, but not a lot of people could have predicted the big climb either.

Using This Model For The Toronto’s Last Crash

Backtesting these numbers, we can see this method isn’t perfect. Using Toronto as the example, at the peak hit in 1989, the model predicted a 26.6% decline of average prices by 1993. By 1993, average sale prices were only down by 24.55%. It wasn’t until 1996 that prices declined 27% from the 1989 peak, before reversing. Not every model is perfect, but for such a relatively simple model – I think it did a fantastic job.

It Doesn’t Have To Drop 28%

There is another option rather than prices dropping 28%, rents can soar. If rents increase by 39% by 2020, we once again have a balanced market. This is less likely due to the fact that rent increases are tied to income increases, but it could happen. More likely a combination of the two will occur to help balance things out.

No one can tell the future with 100% certainty. However, big money investors use well-established models like these to make decisions. The fact that decisions are made on these numbers by such a large group of people, likely does make it a self-fullfilling prophecy in my humble opinion. We’ll be running more complicated models in the future. If you’re in Toronto, you might want to brush up on the inverse relationship between rent and home prices in the meantime.

Ricky Dhadwal
604.418.6600

Leave a Comment

Have a Question?

Contact Us