Numbers can exaggerate bad news

Posted by

Oliver Po 778-898-5153

Century 21 In Town Realty

In Downtown Vancouver

Canada's housing market registered a couple of jolts Monday. One was quite serious; the other, not so much.

The bad one was the news that January's construction of new homes plunged to the lowest level since 2001. The implications for Canada's economy are ugly, at least for the next several months, since home construction is a key driver of employment and growth.

The other was the forecast by Canada's real estate brokers that prices will drop sharply this year. Don't worry so much about this one.

While it seems likely that prices will indeed drop, we need to be very cautious about the widely circulated figure from the Canadian Real Estate Association (CREA) for an eight per cent drop in the average 2009 price compared with the average 2008 price.

Why? Because the usual method of calculating average home prices greatly exaggerates price declines when markets go sour.

Here's an example of how much. Using the standard calculation from CREA, home prices were down by 9.8 per cent in November from 12 months earlier.

But using a more rigorous measure, embodied in the new Teranet-National Bank house price index, the average price for six of the biggest cities in Canada actually inched up a bit, by 0.6 per cent, in this period.

The numbers aren't an exact comparison, since the CREA calculation includes more than six cities, but since these big cities include a very large chunk of Canada's homes, they're a fair approximation of the national market.

And this new index does show how a real home has changed in value.

First, it gives the same weight to each city regardless of how many sales happen there in a particular month. Second, it tracks a huge database of specific homes, looks at those that have sold more than once, then compares actual price changes on the very same home.

The big gap between the two measures isn't because CREA's numbers are bad. It's simply because the association's averages are calculated in a way that isn't designed to demonstrate changes in the price of any particular house.

Instead, the CREA numbers add up the value of every home sale in Canada, then divide by the number of sales. If sales of high-priced homes fall the most, as tends to happen when the market turns sour, the average price falls because it's largely the highest-priced homes falling out of the calculation.

This happens in two ways. First, look at a city like Vancouver, where homes cost twice as much as in the cheapest big city, Montreal. Vancouver sales have collapsed in the past year, while those in Montreal have fallen only about half as much.

When lots of high-priced Vancouver sales drop out of the national average, but fewer low-priced Montreal sales, the reported price for all of Canada falls more than it should.

Second, within Vancouver (or Montreal or any other city) sales of the highest-priced homes tend to fall the most when the economy turns bad, notes CREA economist Gregory Klump. This skews the average for every city where markets are slowing, exaggerating the true price decline within that city.

The upshot is that the CREA numbers are useful for real estate brokers and economists who want to know about Canadian real estate conditions in general, but not a good guide to changes in the value of any individual home.

You can take away part of the distortion in CREA numbers by holding the weighting of each city or province steady, regardless of whether sales have fallen or grown.

By this measure, Klump calculates average prices this year will be down by 6.4 per cent, not the eight per cent raw average. But this still leaves the distortion among different kinds of homes in the same city, so you might reasonably assume that a typical home price will drop less.

And of course, even this is still just a national average. When it comes to housing, we each live in a specific city.

Sadly, Marc Pinsonneault, the National Bank economist who puts together Teranet-National Bank index, isn't in the business of forecasting home prices.

But based on his data through November, the weakest market in Canada was Calgary, with a drop of 7.7 per cent over 12 months, followed by Toronto, down 1.6 per cent, and Vancouver, down 1.3 per cent. The strongest was Halifax, with a gain of 5.8 per cent, followed by Montreal, up 5.1 per cent, and Ottawa, up 4.2 per cent.

That's not a guide to the coming year, since nearly all markets were weakening in November, but it could offer some idea where the biggest drops will come.

Source: Jay Bryan, Canwest News Service

Blog Archives