I feel bad about writing too many abstruse theory posts. How many angels can dance on the head of a monetary pin?
Here's something more practical. Think of it as a companion to Livio's post about Canadian house prices. Why would we worry if Canadian house prices are overvalued? One important reason (though not the only one) is that if house prices are overvalued that is one reason (again, not the only one) to expect they might fall. And if house prices fall, people with large mortgages might have negative equity, and this might lead to defaults, and problems in financial markets.
For American readers especially: let's remind ourselves that Canadian mortgages are mostly recourse mortgages, so you can't just hand in the keys to the bank and walk away from your mortgage to escape negative equity. But nevertheless, if you can't make your monthly mortgage payments (because say you lose your job), the ability to sell your house for more than the mortgage (or for the bank to sell it for more than the mortgage) still matters.
Most headlines about household debt just talk about average debt. As I have argued in the past, average debt is a next to useless measure. This is more than just saying that averages don't tell the whole story ("my hands are burning but my feet are freezing so on average I'm just right"). When it comes to debt, the average tells us even less than it normally does.
For every dollar borrowed there's a dollar lent. Credits=debits. Average net debt (debits minus credits) must be zero by definition (OK, for a closed economy). The representative agent, who personifies the average, does not borrow from himself.
Debt must be disaggregated across the population. Some people borrow from other people. Some have positive debt; others have negative debt. What we need to do is look at the whole distribution of debt across the population. Or, more precisely, we need to concentrate on one tail of that distribution. Concentrate on the people most at risk of default. Those will presumably be the people with the highest debt to asset ratios, or the highest debt service to income ratios. Or, better still, those who have both a high debt asset ratio and a high debt service ratio. Because if things go wrong, they won't be able to make the monthly payments, and won't be able to pay off their mortgage by selling the house.
And I'm very glad to see that Will Dunning, chief economist of the Canadian Association of Accredited Mortgage Professionals, has done just that (pdf).
This to my mind is the most important paragraph in his survey (page 32):
"Thirdly, most of the home owners with mortgages have substantial amounts of
housing equity. Most of those who might face unaffordable increases in mortgage
costs could solve their problems by selling their properties. Among those for whom
the estimated tolerable interest rates are below 5%, an estimated 11% have less than 10% equity. This very small group – representing about 25,000 households – would have the least ease of selling out of a problem. 6% (about 10,000 households) have 10-14.9% equity and 13% (about 25,000) have 15-19.9% equity. A substantial
majority (69%) has 20% or more equity."
In other words, he asks 'how many households are most at risk because they have both a high debt service ratio and a high debt asset ratio?'. And he answers: '25,000 households'. He could be wrong of course. But at least he's asking the right sort of question.
I have no special expertise whatsoever in evaluating Will Dunning's Report. I worry about the response rate to the survey data he relies on, and whether there might be any response bias. But apart from that it looks good to me.
Someone reading this must know more than me. What do you think?