Credit risk across the nation: Who’s most at risk from an interest rate shock?

TD has come out with a very fascinating report in which they’ve assessed the financial vulnerability of households across the country.  It’s full of interesting data, though some of them certainly questionable.  Perhaps most interestingly, they have made easily available a range of financial data going back over a decade.  The trends certainly speak for themselves, and we’ll look at those in more detail in a moment.

In addition, they have created a new index designed to gauge financial vulnerability.  The report reveals some fascinating trend, but also raises a few questions about the raw data itself.

Let’s dive in!

Household Financial Vulnerability Index

I applaud TD for attempting to quantify financial vulnerability by examining a number of related metrics.  The specific metrics used in their new index can be found on page 2 of the report.

I do question why they would weight debt serviceability at 20% of the total index, particularly when the authors actually have calculated a much more meaningful adjusted debt service ratio that measures the effects of future Bank of Canada rate hikes.  If the goal is to measure credit risk, it would make sense to look at household finances once interest rates normalize.  Debt serviceability is simply the percentage of household income being used for principal and interest payments.  It should be obvious that living through an expansion of debt during a period of unprecedented low interest rates will mask some of the underlying weaknesses.

For example, debt to income ratios have ballooned at the same time that the percentage of income directed to debt repayment has fallen.  If we look simply at the amount of income directed towards debt repayment, things look relatively safe.  But to use that as a measure of current credit risk, I would argue, is misleading.

I would also be inclined to weight the debt-to-asset ratio lower than 15% of the total index.  If there’s one thing we’ve learned from watching credit bubbles implode in different parts of the world, it’s that asset prices (particularly leveraged assets like real estate…the bulk of the average Canadian’s net worth) always look great just before the party ends.  I’d suggest that it is a ratio that masks the appearance of credit excesses, the very thing the index aims to measure.

All that being said, it is a step in the right direction.  Here is the current index by province, with the Canadian average set at 1.  Hence, a province with a number larger than 1 is more financially vulnerable than the average:

Note that BC leads the list….no shock….with Alberta and Ontario rounding out the ‘over achievers’.

The raw data

I will come back to the report itself in a later post, but for now I think it’s worthwhile to dig into the raw data and see some of the fascinating trends.

1)  Debt income ratios ballooning

You’ll note that the debt-income ratio has expanded in every area of the country since the start of the new millennium.

Interestingly, TD used Ipsos Reid data in calculating their debt/income ratios rather than Stats Canada data.  Stats Canada has calculated the average debt/income ratio at 148%, much higher than Ipsos Reid’s estimate of 127%.  While the Stats Canada data is aggregate data and the IR data is broken down a little more, I still have a difficult time reconciling the differences in these two averages.

Furthermore, in discussing the limitations of the IR data, TD notes that, “There might be, for example, a tendency for households to over/under estimate the value of their homes”.  Add to that the tendency to over/under state their incomes.  In the name of being neutral I suppose we need to aknowledge that survey respondents could either overstate or understate their incomes and house value…..but suffice to say that I’d suggest it’s highly likely they would do one over the other.

On the national level, debt has really ballooned since 2005, consistent with my observations in my post on Canada’s credit bubble.  Ontario has seen a particularly large increase in debt/income levels since 2008, with BC and Alberta seeing massive jumps since 2009 (though both were significantly stretched well before that).

2)  Debt Service Ratio (DSR) artificially low, but still alarmingly high compared to US consumers

The debt service ratio (the percentage of income diverted towards interest and principal payments) is currently hovering around the lowest point point contained in the data set.  The grey column on the right indicates what the debt service ratio would be if the Bank of Canada raises its overnight interest rate from the current 1% to 3%.

Despite the low current readings, we know that debt as a percentage of income has increased markedly.  What’s up with that?  Two factors actually:

  • 2)  A broad loosening of mortgage insurance standards since 2005.  Increased amortization length has decreased the total monthly debt burden, but greatly increased the total mortgage paid over the life of the loan.

Although I don’t foresee a rapid rise in interest rates, it is nonetheless safe to say that the two factors that have enabled the DSR to remain deceivingly low are now working in reverse now that new mortgage rules are on the way.

Of interest, the DSR in the US peaked around 14% of disposable personal income in 2007 compared to a  current 18.6% reading in Canada.  I have a difficult time accounting for such a massive disparity in the levels between the two countries given our relatively comparable debt/income ratios, but the specific data source used by TD in their calculations is not given.  It could be that mortgage interest deductability in the US is considered in their calculations, but the Fed data suggests that they have used personal disposable income, which implies after taxes are considered.  I’m somewhat stumped on this one.

3) Home price to income ratios……bubbly no matter how you slice it.

Behold!  Note that Canadian home prices were last near their long-term average of 3-3.5 times income back in 2001.

From the report:

“The ratio of existing home prices to income – average resale prices from Multiple Listing Service (MLS) as a per cent of disposable income. This captures relative overvaluation in the housing market as well as the susceptibility of household balance sheets to a housing price correction.”

Based solely on long-term averages, the Atlantic provinces and Manitoba are closest to fair value, while BC, Ontario, Quebec, and Alberta are getting ridiculous (especially BC).

Not much more to comment on here except to point out that house prices in Canada continue to defy virtually all long-term measures of fundamental value.  All the anecdotes in the world don’t stack up against the reality that when the average family can no longer afford the average home without sacrificing their long-term financial health, something will give.

It also bears noting that much of the significant overvaluation is centred in the larger cities.  This was noted in the most recent Demographia survey which found that while the six largest cities in Canada were seriously unaffordable, when the 35 largest markets were considered together, the aggregate number indicated only moderate unaffordability.

4)  Personal savings rate

The personal savings rate in Canada averaged 3.9% in 2010, low by historical comparisons, but arguably not shocking based on current interest rates.

Since 2000, the savings rate has eroded most significantly in Ontario (7.6% to 2.9%).   Most surprisingly is BC’s large and sustained negative savings rate for over 10 years running!

A trusted reader made the following observations in an email exchange:  “The ONLY way the denizens of BC can maintain a negative savings rate for that long is if they are spending the equity in their homes.  That is, they are getting HELOCS and paying bills from the HELOC.  Understand, eventually you can get to the point where you pay the credit card, the electric bill, the mortgage, and the car payment from the HELOC.  That is sustainable until house prices stabilize (and in Vancouver they have and are headed down a bit) and then you can’t get an increased HELOC.  Shortly after that, you have a personal credit crisis.”

“BC is headed for a credit crisis.  A negative savings rate means more money is going out (of the household account) than is coming in from income.”

This is particularly troublesome for BC when one considers that new rules will be restricting HELOCs to no more than 85% of the total home value.

I hope to delve more into this data if I can find the time, but I’ll have to leave it here for now.

Cheers,

Ben

Advertisements
This entry was posted in Economy, Real Estate and tagged , , , , , , , , , , , , , , , , , , , . Bookmark the permalink.

21 Responses to Credit risk across the nation: Who’s most at risk from an interest rate shock?

  1. John in Ottawa says:

    Nice catch Ben!

  2. LS says:

    Do you know how the savings rate is calculated? This article is pretty scathing as to the savings rate being an indicator of anything. He raises some good points but I’m not sure how it is calculated in Canada: http://realclearpolitics.blogs.time.com/2007/02/22/personal-saving-rate-is-a-misleading-indicator/

    • John in Ottawa says:

      Did you notice the date of the article? Coincidence? Check US net worth these days.

      • LS says:

        Right, but his points that cap gains and such aren’t included, and that a pension might not be counted are good ones. If he’s right then it would be a pretty terrible metric.

      • John in Ottawa says:

        It is true that BC may have a negative savings rate because it is over represented by retired people drawing down RRSP savings, but that only lasts until, what age 70. After that, the RRSP is converted to an RRIF, and that is counted as income.

        Maybe people in BC have unusually high savings and are drawing those down as opposed to living on ever increasing credit. Mortgage debt to income is moot if I have almost no income but massive savings, so RBC’s 60+ percent becomes a meaningless metric too. BC remains a mystical lotus land that defies every metric affecting the rest of Canada. Nothing new there.

        As for this author’s point that the savings rate is useless, I would have liked him to walk me through his numbers and provide more robust proof. Instead, I don’t now whether or not he pulled his numbers out of his hat.

        It is hard to know, sometimes, whether statistics are contributing or coincident. In this case, he wrote the article soon after savings rates went negative in the US, and just before all hell broke loose.

      • patriotz says:

        “It is true that BC may have a negative savings rate because it is over represented by retired people”

        That is a fallacy. BC has 15% over 65, compared with 14% Canada-wide.

        It’s the East Coast which has the greatest % of retirees.

        http://www40.statcan.gc.ca/l01/cst01/demo31a-eng.htm

    • Jungernaut says:

      Not that it’s relevant to the author’s point, but I like how the post is from Feb 2007 and it suggests other yard sticks show “the American consumer is in excellent financial health.”

  3. I think there is some merit to what the author was saying, but it is still worth monitoring trends over time. It also makes sense that they wouldn’t count cap gains (losses) in the savings rate or they would be too volatile to be meaningful. You would have had a -30% savings rate in 2008 followed by a 100% savings rate in 2009-2010. See what I mean? It is just one more (imperfect) metric to watch. No one metric gives the full picture, but collectively they start to make a landscape.

  4. Hannibal says:

    Any insight into the reason Alberta’s savings rate has increased over the ten year time period? More specifically, how does that relate to housing prices?

  5. jesse says:

    Including cap gains does not explain the discrepancy between provinces. Do you really think BCers have more growth stocks in their portfolios, or do their cap gains come from real estate?!?

  6. Mango says:

    Your last point on BC is thinking to local. Locals earn local income, but are taking on debt at international prices. See the problem with that? International income/savings are left out, so looking at local savings rate makes no sense.

    • John in Ottawa says:

      Hi Mango,

      I didn’t understand your comment. Perhaps you could explain further. You say that people in BC are taking on “debt at international prices.” What do you mean by an international price and how does that fit into BC’s savings rate?

      Does the rest of Canada take on debt at “local prices?” Why would only BC take on debt at international prices? Is the price of international debt higher or lower than local debt.

    • jesse says:

      I think you overestimate how many people in BC derive their incomes from offshore. Occam’s razor tells me the negative savings rate and high debt levels are related; they sure were in the US.

  7. Stats says:

    Every province (except Saskatchewan) saw debt/income go up starting in 2006-2007…. what does this tell us?

    1) CMHC rules changed in 2006 — and this is the #1 enabled of debt that fuels this bubble.

    2) Potash created its own unique wealth effect. Based on these numbers, Saskatchewan wasn’t that closely linked to the $140/barrel Oil run-up or the Stock Market run-up and crash… but they’ve caught up thanks to the huge run that fertilizer has had.

  8. John in Ottawa says:

    @ Patriotz: Emphasis on “may.” I didn’t believe it. However, next we will be asked to believe west coast retirees are bazillionairs and east coast retirees are destitute.

    I’m doing some demographic research on the US and I was joking the other day with a correspondent that the US breaks down their household financial data by length of hair and breed of pet. An exaggeration, of course, but I can pull up in an instant the net worth, debt, savings, income, etc by age, education, race, sex and region, etc, etc, etc in the US. For free. It takes away a lot of the guess work.

    The real problem and I’ve said it on this blog before and the TD report reiterated it, is the absolute paucity of decent statistics in Canada. I don’t know how the government can run this country with the vague, all encompassing aggregates that pass for statistics here. It is a recipe for making the wrong decision. It is why the TD Bank paid Ipsos Reid to go out and get them some data they can use to make financial decisions! It’s no wonder it was the President of the TD Bank that led the call for tightening of mortgage credit standards in Canada. He got first access to this report.

    • Great points John. I’m wondering if you have any insight into why the Ipsos data on debt/income is so different from the Stats Can data when both are expressing it as a an average percentage of personal disposable income. I’m stumped on this one. The 127% to 148% readings are far too large to be explained by anything other than methodology….but I can’t get my hands on exactly how the data was collected.

      While 127% debt/income ratio is certainly concerning, it is a far cry from the oft-cited 148% stats can estimate.

      • John in Ottawa says:

        I’m going to go with Stats Canada at closer to 148 than 127. StatsCan does an “establishment” survey, that is financial institutions of all types are asked to report credit outstanding, and income if they know it.

        Ipsos Reid was commissioned to do a “household” survey. That is, call people like you and me around dinner time and try to get us to answer some questions.

        They talked to enough people into answering to get a statistically meaningful survey, probably +/- 5%. But think about this. It is dinner time; do you know exactly what your Visa balance is, the Bay, the Brick, your line of credit, your mortgage outstanding?

        I know my wife’s income within 5%, but I haven’t got a clue how much she owes on her Visa or her line, or for that matter what her bank account balance is.

        So, household surveys are prone to a pretty significant error simply because people don’t have the facts necessary to answer the questions accurately right in front of them. Establishment surveys tend to be more accurate when it comes to household finances.

        On the other hand, household surveys can be more accurate when it comes to calculating things like unemployment. Chances are the respondent knows their employment status right off the top of their head.

  9. Probably a pretty good guess. Makes sense. Thanks for the insight.

  10. Pingback: Primer 7: Housing envy and the ‘Joneses’…How a realignment in consumer expectations will pull at aggregate house prices | Financial Insights

  11. Pingback: Canada: A nation of marshmallow eaters? | Financial Insights

  12. Pingback: So much for the ‘conservative’ Canadian consumer: Another look at Canada’s credit bubble | Financial Insights

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s