MAY 11, 2011
Today we turn our eyes to the change in house prices relative to rents. The data sources are the Canadian Real Estate Association (CREA) who kindly provided me with quarterly house price data going back to Q1 1980, and the Rented Accommodation Subindex of the Consumer Price Index.
For today, we will examine changes in rents and house prices in Vancouver, Calgary, Edmonton, Winnipeg, Ottawa, Toronto, and Halifax.
Before examining the data, let's review why rents are an important determinant of the value of housing (both investment housing and residential housing).
Why rents matter
Interestingly, it was the house price to rent ratio and the house price to income ratio that alerted some economists to the severe overvaluation in the US housing market (precious few as those economists were). It is worth examining why this ratio is an indicator of potential house price overvaluation and why it applies to both residential and investment housing. With permission, I am going to include some sections from a great post made by Jesse over at his blog:
To understand the price-rent ratio, we can first go back to basic finance math and look at the “net present value” (NPV) of a stream of future cash flows. Cash in the future is normally worth less than today and such is discounted at the “discount rate”, sometimes referred to as the “cost of capital”. For example if I promise to pay you $1000 today or $1000 a year from now, you will put less value on the future $1000 because you can invest today’s $1000 and receive, say, 2% interest risk-free, so my future $1000 is only worth around $980 (1000/1.02). If you are skeptical I am capable of paying you back on time (or at all), you will further discount my $1000 by some amount. If I’m the Canadian government you probably won’t discount much beyond expected inflation, if I’m the Greek government or some Joe of the street you are probably discounting a lot. If you can recoup the money by liquidating collateral put up by the debtor, the discount rate decreases. We can sum multiple discounted cash flows from different payment periods together and get its NPV. NPV is the price an investor would be willing to pay for an investment given the risks involved.
For a property, this series of discounted cash flows is simply revenue (rents) minus expenses (maintenance, taxes, management fees, and capital replenishment) from future payment periods. The revenue and expenses have some variability and risk inherent in property – there is no real way of avoiding it. As a result property will always have a higher discount rate than risk-free. However with property, as a bonus of sorts, revenue and expenses tend to increase roughly with inflation. (Well not quite; as a building ages it depreciates and has higher ongoing maintenance expenses, while the rent will slightly lag inflation – older buildings rent for less than newer ones.) As with any investment we sum the cash flows to get NPV:
NPV = (R-E) + (R-E)(1+i)/(1+d) + (R-E)(1+i)^2/(1+d)^2 …. + (R-E)(1+i)^N/(1+d)^N
where R is the rent, E is expenses, i is the rental inflation rate (which normally but not necessarily tracks CPI), d is the discount rate, and N is the final payment period. Assuming N is large, using simple math this reduces to:
NPV = (R-E)/(d-i)
Notice the denominator d-i for a moment. Remember d is the discount rate, which includes future inflation expectations and risk. In other words, d-i is what we can call an inflation-independent measure of risk. (Again this is a bit of a lie but not much of one.) This number is also referred to as the “cap rate”.
(You may also note that E does not include financing expenses. For the purposes of this analysis, assume we use our own money. Besides, anyone we borrow from or invests with us should be doing the same calculations we are.)
The analog for cap rate in other investments is the price-earnings ratio and the principle is the same (P/E => 1/caprate). With so-called growth investments, the expenses are front-loaded and a potential boon from revenues is pushed out and often heavily discounted, meaning their price-earnings are higher. Property investment will tend to more closely match the price-earnings of utilities like gas, electricity, or water. (Property is, after all, a utility in its own right.)
Misuse of price-rent
Price-rent ratio metrics are an approximation. There are some deviations that can and do exist, some substantial. We hear stories of properties with 1000:1 price-(monthly) rent in parts of Greater Vancouver. When using price-rent to value a property it’s important to take into account its highest and best use. For example, a small house surrounded by larger more affluent houses (or condominiums) is likely going to be re-developed. Its value will be based on comparing various DCF scenarios and choosing the highest one. An underdeveloped piece of land is akin to a call option. In this scenario there are definite grounds for a higher price-rent (though maybe not 1000:1!) because an owner can sell today for today’s best use.
A piece of land expected to be redeveloped (or simply to have its revenue outpace inflation due to significant income growth) at some point in the future, but is not occurring today, is akin to a growth investment, where the market is assuming future cash flows will increase significantly some point in the future. This may or may not be rational and the premiums placed on these properties are speculative.
An important consideration for residential real estate, unlike other investments, is that it is not just investors who are active in the market. Owner-occupiers who prefer to own will be willing and able to pay a premium over rental value for a property, a so-called consumer surplus. In the case where owner-occupiers are heavily active in a market an investor either tags along for the ride, hoping to sell at a higher price in lieu of a low cap rate, or simply sits on the sidelines or invests in better returns elsewhere. Canada’s ownership rate has increased from 64% at the turn of the century to close to 70% today. If one believes the newly-minted Canadian owners place a premium on ownership, this will tend to have driven marginal prices higher. At some point, however, there are no more marginal owner-occupiers buying and we’re left with the “sidelines” investors requiring rental value without the premium.
Where Canada stands and what it implies:
It's first worth noting that the aggregate price/rent ratio in Canada is at an all-time high.
In fact, back in 2008, the OECD released house price and rental data for a number of countries. When compiled, it gave us the following graph. Note that Canada is the light blue line and at that time was second only to Spain in terms of our house prices versus the rent they would fetch.
None of this is particularly new. In fact, in a 2005 paper by the OECD, they recognized that massive, anomalous rise in price/rent ratio across Canada. With the exception of a brief dip in 2008-2009, we know that in the time since 2005, prices have far exceeded rental growth.
What does this imply for future price movements? To answer this, we'll turn our attention to a remarkably prescient paper written by John Krainer and Chishen Wei of the Federal Reserve Bank of San Francisco back in 2004. For reference, the US market peaked barely a year after this paper was published. Krainer and Wei argued that the price to rent ratio is important in determining future movement of house prices. Interestingly, they argued that when the price to rent ratio is too far out of whack, the bulk of the movement to realign the ratio occurs via a change in house prices rather than a change in rent. In other words, when the price/rent ratio is at historic highs, as it is in Canada, it is likely that the realignment will come by falling or stagnant house prices rather than rapidly rising rents.
The majority of the movement of the price-rent ratio comes from future returns, not rental growth rates. This will not comfort everyone, as it implies that price-rent ratios change because prices are expected to change in the future, and seemingly out of proportion to changes in rental values.
We found that most of the variance in the price-rent ratio is due to changes in future returns and not to changes in rents. This is relevant because it suggests the likely future path of the ratio. If the ratio is to return to its average level, it will probably do so through slower house price appreciation.
How true that wound up being. As the chart below shows, the US is closing back in on its long-term price-to-rent ratio trend line. We know that it has done this largely through falling house prices.
What about rent control?
Some Canadian provinces, most notably Ontario, Quebec, and Manitoba, have rent controls. Others have less strict rent controls such as BC, while others are largely hands-off when it comes to setting rent increases, like Alberta. I will say that I am not a fan of rent control. For some interesting reads on how rent control is poor economic policy, check out these papers and editorials:
Rent Controls: Do Economists Agree- Econ Journal Watch
Reckonings; A Rent Affair- Paul Krugman in the NY Times (Hey....the guy doesn't always get it wrong...)
How Rent Control Drives Out Affordable Housing- CATO Institute
While rent controls have arguably suppressed rents relative to house prices, there are two things worth noting:
1) Rent controls have existed in Ontario since the mid 70s and in Manitoba since the 80s. BC also has a measure of rent control while Alberta has no rent control laws. As we will see in a minute, the major cities in all of these provinces, those with and those without rent controls, have seen house prices far outpace rents, particularly since the mid 2000s. In the case of Winnipeg, a city with strict rent controls, house prices and rents tracked each other nearly perfectly until the early 2000s, then experienced a massive divergence thereafter. Clearly there is another macro factor at play. It is much more likely that house prices across the country have been buoyed by policy at the national level rather than at the provincial level. Loosening of CMHC mortgage requirements, the removal of the CMHC insurance ceiling, and a falling interest rate environment are much more likely responsible for the divergence between house prices and rents.
Along this same line, it's worth noting that in provinces where strict rent controls are in place, they prevent the realignment of the price/rent ratio through rapidly rising rents, leaving stagnant or falling prices as the only possible means of realignment.
2) In the US, there are fewer examples of strict rent control states, but one glaring example is in California. Despite this, house prices have fallen significantly in that state as well.
Rents and price in major cities:
With all of that as a background, let's see how some of the big cities stack up. What is depicted in each graph is the change in house prices and rents with Q1 1980 set at 100 for each variable.
What are the implications?