TD Economics: Canadian Household Debt A Cause For Concern

8/8/2019 TD Economics: Canadian Household Debt A Cause For Concern

1/13

Special Report TD Economicswww.td.com/economics

October 20, 2010

CANADIAN HOUSEHOLD DEBTA CAUSE FOR CONCERN

HIGHLIGHTS

Canadian personal indebtednesshas become excessive. Low in-come families seem particularilyvulnerable

Economic and financial funda-mentals suggest that the personaldebt-to-income ratio should be inthe 138% to 140% range over thecoming five years. The currentratio is at 146%.

A U.S.-style crisis is not in the

making, but Canadian personaldebt growth must slow relative toits past rapid pace of increase.

Various factors point to a modera-tion of household borrowing, buta sustained low rate environmentwith short-term rates only return-ing to 3.50% by 2013 may stillsupport personal liability growthof 5% annually. With personalincome growth likely to advanceat 4% per annum, personal debt-

to-income could rise to 151% by2013. This suggests that further pruden-

tial actions might be warranted,but should not occur until thecurrent housing cooling has runits course and the economy is ona firmer footing.

The relentless rise in household debt in Canada, both in absolute terms andrelative to personal disposable income (PDI), is a growing cause for concern.Since the mid-1980s, total household debt as a share of PDI in Canada has almosttripled from 50% to 146% and a visible acceleration in the long-term trend of debt accumulation has taken root since 2007. With debt-loads mounting in Canadaand U.S. personal debt in decline (re ecting deleveraging and home foreclosures)over the past couple of years, there has been a rapid convergence in the Canadianhousehold debt-to-income ratio vis--vis that of the United States.

In this report, we provide answers to some of the most pressing questionson the topic of Cana-dian household debt.In particular, is Canadaheaded for a U.S.-stylehousehold debt crisis?And, is there an optimalor sustainable level of household debt? Theanswer to the rst ques-tion is no. The Cana-dian debt imbalance iscurrently not as greatas that experienced inthe U.S. and does notrequire a major dele-veraging. However, theanswer to the secondquestion is that Canadian personal indebtedness has become excessive relative towhat economic models would predict as appropriate. In other words, growth inpersonal debt must slow relative to income growth over the coming years or elsethe risks of a future deleveraging will increase.

What demand factors account for the upward trend in householdindebtedness?

The long-term upward trend in personal debt cannot be pinned on just one ortwo factors, although a signi cant portion can be tied to structural shifts in themacroeconomic environment particularly during the 1990s. The introductionof in ation targeting by the Bank of Canada in the early 1990s set the stage for asecular decline in interest rates that improved debt affordability. At the same time,these macroeconomic trends created a heightened sense of nancial security amonghouseholds. Low and stable in ation reduced the likelihood of future interest-ratevolatility, while relatively stable growth in the economy and job market loweredthe probability of layoffs and an interruption in household income all of which

made households more comfortable carrying greater debt loads.

HOUSEHOLD INDEBTEDNESS - CANADA AND U.S.CONVERGE

60

80

100

120

140

160

180

1990 1995 2000 2005 2010

Canada

U.S.

debt as a % of pdi

Source: Statistics Canada, Federal Reserve Board, Haver Analytics *Note: Both measuresare caculated in a way that makes them comparable. The Canadian measure totals 146%when you include accounts payable.

Craig Alexander, SVP and ChiefEconomist416-982-8064

mailto:[email protected]

Derek Burleton, Vice President &Deputy Chief Economist (Canada)416-982-2514mailto:[email protected]

Diana Petramala,Economist (Canada)416-982-6420mailto:[email protected]


2/13

Special ReportOctober 20, 2010

TD Economicswww.td.com/economics

2

Nowhere was the impact of lower borrowing costs andgreater household con dence more clearly observed thanin the housing market, where ownership rates increasedsteadily over the past two decades. A self-perpetuatingcycle occurred. Strong increases in demand bid up hous-ing prices, which together with equity market gains priorto the 2008/2009 recession, raised net wealth. This positivewealth effect encouraged households to increase their rate of investment and consumption, further driving up borrowingand debt levels.

Demographics also helped drive demand for credit.Baby boomers (individuals born between 1946 and 1964)shaped debt trends just as they shaped product markets.They bought homes and then moved up the property ladderover time, using real estate as a source of wealth creation.

Another key macroeconomic trend that boosted demandfor credit was increased labour market participation bywomen. Experience shows that households with two incomeearners tend to carry more debt per person relative to theirincome. Having two incomes creates a sense of incomesecurity, as the probability of losing both income streams ismuch reduced. This can be a false sense of security if bothincomes are needed to service debt.

During the 2000s, the echo generation has been provid-ing a boost to home purchases, helped by favourable housingaffordability created by low interest rates that allowed theseyoung workers to borrow sizeable amounts.

Demand for credit has received a signi cant boost froma cultural shift from thrift towards consumerism. This hasbeen an international trend, in which consumers have agreater desire to consume larger quantities of goods andservices than they have in the past particularly discre-

tionary items. Households have also become impatient,meaning that when they want something, they have becomemore inclined to nance purchases through credit to enjoyconsumption sooner rather than later. This has altered thelifepath of spending. The traditional lifepath model is thatindividuals wish to smooth consumption over their lifetime.They borrow when young, pay down debt and save for re-tirement when more mature, and then run down savings andassets when older. However, individuals are now taking ondebt earlier, and maintaining debt longer. For example, anincreasing number of retirees are carrying debt after leavingthe labour market.

As one might expect, carrying a higher debt burdenmeans that more Canadians are at risk of running into dif- culty meeting their nancial obligations in the event of an unforeseen economic or nancial shock. This wouldnormally act as a check on growth in demand for credit.However, the social stigma associated with declaringpersonal insolvency has declined. Indeed, whereas in the1950s or 1960s individuals would nd it dif cult to admitbankruptcy, today such an occurrence is generally met withunderstanding and support a desirable and positive devel-opment but one that is still supportive to increased leverageof personal balance sheets. Moreover, individuals whogo into bankruptcy are no longer credit market outcasts.Seven years after declaring insolvency, individuals becomeeligible for credit once again from most institutions, and inthe interim most are able to get access to credit, albeit atlikely punitive interest rates.

How important have supply side factors been indriving credit?

As demand for credit rose in recent decades, a number

-15

-10

-5

0

5

10

15

20

25

1980 1989 1998 2007

Real household credit(y/y%)3-month t-bill yield (%)

Source: Statistics Canada, Bank of Canada, Haver Analytics

HOUSEHOLD CREDIT SHOWS DELAYED RESPONSE TO INFLATIONTARGETING

Inflation targetintroduced

%

CANADIAN HOMEOWNERSHIP RATES

56

58

60

62

64

66

68

70

1971 1976 1981 1986 1991 1996 2001 2006

% of occupied private dwellings owned

Source: Statistics Canada


3/13



3

of supply-side developments helped raise credit availabil-ity to households. These supply-side developments haveincluded increased competition within the nancial indus-try, the growing use of securitization, product innovation,deregulation in the banking sector and the relaxing of somecredit constraints, which particularly bene ted rst-timehome buyers.

A number of reforms to the Bank Act occurred in the1980s and early 1990s that increased competition in theCanadian financial sector. Domestic competition washeightened, and while foreign banks have been operatingin Canada since the early 1980s, changes to the Bank Act inthe late 1990s removed some restrictions on foreign banks.Non-bank lenders also became more active in credit markets.The impact on pricing was more limited than the impact onthe supply of credit as institutions fought over market share.

Furthermore, nancial innovations like the automation of credit approval, and widespread use of standardized creditscoring helped to make the loan application process movemore quickly and ef ciently. But even more importantly,increased competition helped to spur signi cant nancialinnovation that made credit more attractive. Credit cardsthat provided bene ts to card-holders for travel and the likewas one innovation of note in the 1980s that encouragedindividuals to carry larger monthly balances. The introduc-tion of home equity lines of credit (HELOCs) was the mostsigni cant innovation of the 1990s and 2000s. Prior toHELOCs, the ability of households to borrow was largelyconstrained by their current and future income. HELOCshave allowed households to borrow more against the valueof their homes or extract equity from their home for con-sumption or investment purposes, while simultaneously of-

fering more exible repayment terms than with a traditionalmortgage. The result has been greater access to credit andlower monthly payments in a low interest rate environment.As shown in the accompanying chart, the popularity of HELOCs has risen over the past 10 years.

Innovations in the ways nancial institutions fund mort-gages and other loans have also played a supportive role.Securitization of mortgages and other loans lowered fundingcosts for nancial institutions, which in turn increased thesupply of credit.

Another contributor to rising household indebtednessover the last 20 years has been adjustments to mortgageinsurance rules. Three major changes to mortgage insur-ance rules helped to make mortgage credit more availableand attractive.

First, the required down payment was reduced. In theearly 1990s, a homebuyer required a 10% down payment toqualify for mortgage insurance. Through a series of regula-tion changes over the 1990s and early 2000s, the minimumdown payment was reduced to 5%. The down paymentwas temporarily lowered to zero by the end of 2006, butwas then taken back up to the current requirement of 5%in October 2008.

Second, the quali cation requirement for mortgage in-surance was eased in April 2007. A homebuyer is currentlyonly required to purchase mortgage insurance if the downpayment is less than 20%; previously that threshold was 25%

Third, the maximum amortization was increased in stepsfrom 25 years at the start of 2006 to 40 years by the fall of that year. As we discuss later, this has since been reducedto 35 years in October 2008. The extension to 40 yearsamortization provided a sizeable boost to affordability. For

CANADIAN HOUSEHOLD CREDIT

0

200

400

600

800

1,000

1,200

1,400

1,600

1971 1976 1981 1986 1991 1996 2001 2006

Economy-wide

Chartered Banks

Bil. C$

Source: Statistics Canada/Haver Analytics

Impact of securitization, non-banklenders and foreign lenders

CANADIAN HOUSEHOLD CREDIT BY TYPE OFCREDIT

0

200

400

600

800

1,000

1,200

1,400

1,600

2006 2007 2008 2009 2010

Other Credi t HELOCs

Source: Bank of Canada, Equifax

Bil. C$


4/13



4

example, an individual with an income equal to the nationalaverage in 2007 that purchased a home equal to the nationalaverage at that time and opted to nance at the 5-year postedrate could carry a mortgage $40,000 larger with a 40-yearamortization period rather than a 25-year one with the samemonthly payments.

The bottom line is that all of these changes boosted hous-ing affordability and encouraged more rapid growth in realestate associated debt.

Up to this point, we have itemized the various demandand supply factors, but one should also acknowledge thatthere is a signi cant interplay between the two. Typically,this dynamic is pro-cyclical. During periods of robusteconomic and housing activity, demand for credit rises and nancial institutions accommodate the growth by increasingcredit availability and innovating in order to boost supplyof credit. However, the last couple of years have beenatypical. Household debt accelerated relative to incomeduring the most recent recession, bucking the experience of the past two recessions in the 1980s and 1990s when risingunemployment led to slower personal debt accumulation.In both of the last two economic recessions in Canada, atightening in monetary policy was a leading contributor tothe downturn. Leading up to this past downturn, interestrates were not as high as they were heading into the 1980sand 1990s recession, and as a result, monetary policy hasbeen far more accomodative this time around. As we willargue later, this countercyclical behaviour may have helpedget the Canadian economy out of recession, but it has alsomeant that the economic downturn failed to unwind theperiod of excessive debt growth relative to income that took place in the 2000s.

Is Canada alone in experiencing an upward trend inthe debt-to-income ratio?

With similar supply and demand dynamics evident

across the advanced economies, the upward trend inhousehold indebtedness over time has been an internationalphenomenon. In countries that tend to be more conserva-tive towards household borrowing and have lower homeownership rates like Germany, Italy, and France therise in indebtedness has been more shallow and gradual.Nonetheless, in ation targeting by the European CentralBank and credit innovations that improved debt affordabilitystill boosted credit growth. In the historically Anglo-Saxoncountries of the U.S., U.K., Canada and Australia whichtend to have a culture more tied to consumption and homeownership household debt growth has been the strongest.The rise in indebtedness in the U.S. and U.K. was remark-able and clearly excessive. These countries experienced alarge housing bubble that was subsequently followed by areal estate bust in 2007 and 2008, which resulted in a sharpdecline in household debt-to-income ratios. Australiaseconomic performance over the last two years has been themost like the Canadian experience, but nothing like that inthe U.S. and U.K.. The comparability of the Canadian andAustralian experience is not surprising, as both are smallopen commodity-driven economies where domestic demandremained strong and the housing market and labour marketrecovered quickly after a short-lived contraction duringthe 2009 recession. Nevertheless, monetary policy has notbeen as accomodative down under, nor was there the samedegree of relaxation in mortgage rules in Australia. Accord-ingly, the rise in household indebtedness in Australia overthe last three years has been more subdued than in Canada.

INTERNATIONAL PERSPECTIVE ON HOUSEHOLDINDEBTEDNESS

0

20

40

60

80

100

120

140

160

180

1993 1995 1997 1999 2001 2003 2005 2007 2009

Australia FranceGermany Sweden

U.K.

debt as a % of pdi

Sources: ABS, BBK, ONS, INSEE. *Note: Methodology differences makethese meaures not comparable to Canada, but are still indicative of trend

EXISTING HOME AFFORDABILITY

20

25

30

35

40

45

90 92 94 96 98 00 02 04 06 08 10

40-year amortization

25 year amortization

Mortgage payments* as percentage of income

* For the average price home using a 75% LTV ratio, 5-year fixed rate.Source: CREA, Statistics Canada, Bank of Canada


5/13



5

Is there a limit to how high the debt-to-income ratiocan rise?

There is no constant or optimal debt-to-income ratio.As mentioned earlier, the demand for debt is in uencedby trends in personal nance and personal preference. Forexample, if debt is being accumulated to purchase assets,and asset prices are likely to rise at a considerable rate,then borrowing to accumulate wealth makes perfect senseand can lead to a higher debt-to-income ratio. Moreover,

nancial innovation can lead to a structural rise in debtrelative to income. Imagine a situation where mortgagescould be amortized over 100 years. The resulting decline indebt service costs would mean that households could carrymuch higher debt relative to income. What truly matters iswhether the prevailing debt-to-income ratio makes sensebased on the current structure of nancial services and theprospects for personal nances.

The recent U.S. and U.K. experience shows what canhappen when the ratio does become excessive relative toeconomic and nancial fundamentals. Coincidently, debt-

to-PDI ratios in both countries peaked at close to 160% of PDI some 14 percentage points above Canadas currentlevel.

Can a 160% threshold be used as an appropriate guide-line for determining when a particular risky level has beenreached? The challenge with merely applying the 160%threshold to Canada is that it fails to account for the fact thatdebt ratios in the U.S. and U.K. likely overshot their sustain-able levels. The issue is by how much, and that is dif cultto tell. Furthermore, the appropriate level of debt relativeto income is likely higher given the different debt structure

in these countries. For instance, in the U.S., households

can deduct mortgage interest payments from their incometaxes payable. This would have the effect of allowing U.S.households to carry more debt relative to income than Ca-nadian households.

Moreover, the debt-to-income ratio has its own inherentlimitations. Since households often use debt to accumulateassets, which in turn can be drawn on in case of nancialstress, or used to smooth out their consumption and/or toprovide income during retirement, it becomes critical tolook at a broader array of household ratios to assess thevulnerability of households to economic shocks. These in-clude the debt-to-net worth ratio, the asset-to-liability ratioand the share of homeowners equity within total assets.Most importantly, the total debt-to-income ratio falls shortin providing a clear gauge on the ability of households tomeet their debt obligations. This is because the incomeused is annual income, and Canadian households dont payoff all their debt in one year. So, affordability of debt asmeasured by an estimated debt service ratio needs to be akey part of the analysis.

What are these other indebtedness metrics saying?

Consistent with the debt-to-income ratio, all of the Ca-nadian debt metrics seem to line up on the side of growingvulnerability, but not a looming crisis.

First, one needs to consider the evolution of householdbalance sheets. Over the 1980s and 1990s, the rise in thehousehold debt-to-income ratio was accompanied by sta-bility in the debt-to-assets ratio and the debt-to-net worthratio. In the 2000s, however, there was a break in the trendwhere accumulation of household debt was encouraged bycapital gains. Put another way, households saved less and

HOUSEHOLD INDEBTEDNESS GROWS OVERTIME

60

70

80

90

100

110

120

130

140

150

160

1990 1995 2000 2005 2010

debt as a % of pdi

Source: Statistics Canada, Haver Analytics

HOUSEHOLD INDEBTEDNESS AND ASSETACCUMULATION

0

100

200

300

400

500

600

700

800

900

1990 1994 1998 2002 2006 20100

20

40

60

80

100

120

140

160

assetsindebtedness

debt as a % of pdiassets (book value) as a % of pdi

Source: Statistics Canada, Haver Analytics


6/13



6

relied too heavily on asset price gains to do the saving forthem, while at the same time they were comfortable addingto their debt obligations. The trouble is that at least some of the asset price gains are likely unsustainable, as equities arestruggling to regain the ground they lost during the recent nancial turmoil. Moreover, TD Economics believes thatnational average home prices have risen roughly 10-15%above the levels supported by economic fundamentals.Since 2006, the rise in the debt-to-income ratio has beenassociated with a large increase in the ratio of debt-to-assets

and debt-to-net worth for the rst time on record. Over 2009,the rise in these measures largely re ected a sharp decline inasset values related to the short-lived correction in housingand equity markets. However, even as asset values havereturned to pre-recession levels, these measures remain athistorically high levels. This is particularily concerninggiven that we believe that the rebound in asset values maybe a bit overdone.

Estimates of the aggregate household debt service ratiohave also been ashing some warning lights. What is typi-cally used to assess debt affordability is the ratio of interest

payments on debt as a share of PDI. This measure, whichdoes not include principal payments, is popular in largepart because it is computed and made readily available byStatistics Canada. Interest costs also pose the biggest risk to household nances, whereas principal payments aregenerally stable.

Although interest costs absorb a relatively small shareof PDI, the fact that they are not at record lows given thenear record low level of interest rates is striking. In fact,the last time the debt-service ratio was at its current valueof 7.2%, the overnight rate was at 4.25% rather than its

current level of 1.00%.

This perspective is somewhat backward looking, how-ever, since interest rates will not remain at these emergencylevels over the medium-to-longer term. An importantexercise is to calculate what this ratio would rise to if theovernight rate were at a more normal level, say 3.5%, givena debt-to-income ratio of 146% and a moderate pace of PDIgrowth of 3.0-4.0%. Under this scenario, the debt-interestcost ratio would climb to 8.5% a level not experiencedsince mid 1990s when interest rates were at double digits.

But, once again, this does not provide the full story, sincehouseholds must shell out to meet both principal and interestpayments. In Canada, data on principal repayments is dif- cult to come by. The Bank of Canada uses data availablefrom the Ipsos Reid Canadian Financial Monitor (CFM),which provides detailed nancial data on households acrossthe country 1.

Similar to the interest-only debt service ratio, these gures reveal that the affordability of debt remains withina comfortable range but only because interest rates remainextremely low. However, if short-term rates were to risetowards 3.5% and 5-year rates were to rise towards 5%, thecombined principal and interest payments would reach 23%of PDI the highest level since 1999, which is the start of the series. Under the more likely scenario in which the debt-to-income ratio continues to grow albeit at a much moremuted pace than experienced since 2007 debt servicingcosts will reach 24% of PDI as interest rates return to morenormal levels. Put another way, if interest rates rise 3 per-centage points, the debt-to-income ratio would have to fallback to levels seen in 2006 (125%-130%) to have the samedebt service costs as today.

INTEREST DEBT SERVICE RATIO

-6.0

-1.0

4.0

9.0

14.0

1990 1993 1996 1999 2002 2005 2008 2011 20145.0

6.0

7.0

8.0

9.0

10.0

11.0

3-month t-bill

debt-to-income ratio

3-month tbill (%) DSR (%)

Souce: Statistics Canada, Haver Analytics, Forecast by TD Economics as ofOctober 2010

Forecast

CANADIAN HOUSEHOLD METRICS OFINDEBTEDNESS

10

12

14

16

18

20

22

24

26

Q1-1990 Q1-1993 Q1-1996 Q1-1999 Q1-2002 Q1-2005 Q1-2008

Debt-to-assetsDebt-to-net worth

Ratio, %

Source: Statistics Canada/Haver Analytics


7/13



7

What share of the Canadian population is particularlyvulnerable?

Average statistics can conceal the real story. In this case,the moderate increase in the aggregate debt affordabilitynumbers hides the true extent of Canadians that are in aposition of nancial stress.

Researchers at the Bank of Canada (BoC) use the IpsosReid data to estimate the number of households that wouldbecome nancially vulnerable given the current level of

indebtedness, and under various interest rate outcomes. Thestandard the BoC applies to determine nancial vulnerability(or stress) is households whose debt-service ratio exceeds40%. The BoC uses a 40% threshold to determine vulner-ability, because households with debt service ratio above thismark have a greater probability of defaulting on their loans.

The analysis was last published in early 2010, using2009 data. The BoC found that just over 6% of householdswere in a position of nancial stress at that time. Under ascenario where the overnight rate rises to 3.5%, debt contin-ues to grow at its current rapid pace, and PDI growth runs

at a healthy 5% annualized growth, approximately 7.5% of households would become nancially vulnerable.Since the BoC conducted its analysis, more up-to-date

gures from Ipsos Reid have been released for the rsthalf of 2010. We have used similar methodology to theBoCs to rerun the analysis. Based on the new gures, aslightly higher 6.5% of households are currently nanciallyvulnerable (or have a debt-service ratio of 40% or above).More striking, the share of those on the verge of becomingvulnerable (those with a debt-service ratio of 30-40%) hadrisen from 7.2% in 2009 to 9.3% - up almost two percent-

age points. Given the change in the distribution of debt,

we have estimated that as much as 10-11% of householdsmay become nancially vulnerable if the overnight raterose to 3.5% under similar assumptions used by the Bank of Canada 2. The ratios do not suggest that a major personal nancial crisis is brewing. For example, the vulnerabilityratio reached 15% before U.S. households became signi -cantly distressed. Furthermore, nancial stress in the U.S.was compounded by a massive spike in the unemploymentrate. But, the analysis does highlight that more Canadiansare vulnerable to higher interest rates that must ultimatelycome when the economy is stronger.

What segment of the population is particularly atrisk?

Digging even deeper into the Ipsos Reid database, it be-

comes apparent that the concentration of those householdsin nancial stress are at the low end of the income spectrum.While about three quarters of overall debt is still held bymiddle-to-high income families, low-income families havethe highest debt-to-income ratio (180%), and the highestdebt-service ratio (25%). Households in the lowest incomebracket are also more vulnerable to rising interest rates andwould face debt service costs exceeding 30%, on average, if interest rates were to normalize, thus pushing them close tothe nancial stress threshold. These statistics are importantbecause low-income families are more susceptible to adverse

economic shocks (more likely to lose their jobs), and theydo not have a strong asset base that they can liquidate intimes of nancial stress.

The Ipsos Reid data also shows that the older popula-tion (65+ years) is holding more debt than they have in thepast, and those that should be preparing for retirement (ages

HOUSEHOLDS ARE HOLDING MORE DEBT

0

5

10

15

20

25

30

35

40

0-10% 10-20% 20-30% 30-40% greater than40%

2005 2009 2010Q1

Debt Service Ratio (%)

% of debt-holding households with given debt-service ratio

Source: I sos Reid

TOTAL HOUSEHOLD DEBT-SERVICE RATIO(for debt-holding households)

0

5

10

15

20

25

30

1999 2002 2005 2008 t=1

Source: Ipsos Reid, Forecast by TD Economics as of October 2010

Debt service costs as a % of PDI

Impact as interest rates normalize,and debt-to-income remains constant


8/13



8

55-64 years) are carrying heavier debt burdens than in thepast. The data indicate that more Canadians are choosing tostay in the labour force longer perhaps due to higher debtloads or lower nancial security. However, interpreting thedata is dif cult. An alternative perspective could be thatthe higher debt load is the result of households choosing towork longer simply because they want to, or are physicallyable to. The causation is unclear, but the higher indebted-ness of older Canadians is concerning.

Are we headed for a U.S.-style deleveraging?

In spite of growing vulnerability of a signi cant minor-ity share of households, the level of risk associated withhousehold indebtedness appears signi cantly lower thanthat in the United States. As we have already noted, theU.S. debt ratio at its peak in 2007 was signi cantly higherthan in Canada today. In the U.S., bankruptcies and loan

delinquencies were also a larger concern than they are inCanada. For instance, the share of mortgage in arrears inthe lead-up to the recession (i.e., those related to the levelof debt rather than unemployment) jumped to 1.5% in theU.S. a level three times higher than that currently recordedin Canada.

The key explanation behind the differing householdcredit conditions are well known by now. In the U.S., -nancial institutions undertook much riskier lending practicesduring the sub-prime boom. As a result, a much higher shareof U.S. households became over-extended. In 2007, 15%of U.S. indebted households had a debt service ratio above40% twice the level of that in Canada. Even under thescenario where household debt continued to grow signi -cantly faster than income over the next couple of years, theshare of Canadian households that would become nanciallyvulnerable would not reach the levels experienced southof the border in 2007. As mentioned above, the ability of households to deduct mortgage interest costs from taxes pay-able would have the effect of allowing U.S. households tocarry more debt relative to their income than their Canadiancounterparts. However, that being said, research still showsthat households with a debt-service ratio of 40% and aboveare more likely to become delinquent on loan payments.

Another differentiating factor is the stronger balancesheets enjoyed by Canadians, on average. In the UnitedStates, there was a more pronounced deterioration in othermeasures of indebtedness such as the debt-to-assets anddebt-to-net worth ratios leading up to the crisis comparedto the recent experience in Canada. The average Canadianhas a signi cantly higher amount of equity built up in theirhomes, relative to their U.S. counterparts, where many hadnegative equity positions. In the U.S. (and U.K.) a large

MORTGAGE IN ARREARS

0

1

2

3

4

5

6

1990 1995 2000 2005 2010

Canada

U.S.

% of outstanding mortgages

Souce: CBA, Moody's

U.S. HOUSEHOLDS MORE VULNERABLE THANCANADIAN HOUSEHOLDS

0

5

10

15

20

25

30

35

0-10% 10-20% 20-30% 30-40% greater than40%

U.S. (2007)

Canada (2010 Q1)

Debt Service Ratio (%)

% of debt holders within given range of debt-service-ratio

Source: Ipsos Reid, Federal Reserve Board

DEBT SERVICE RATIO BY INCOME GROUPExcluding credit cards

0

5

10

15

20

25

30

35


9/13



9

Implications for Personal InsolvenciesThe growing nancial stress among low income

Canadians and the likelihood of a relatively high un-

employment rate projected (7.5-8%) over the mediumterm suggests that the rate of credit delinquencies andbankruptcies will remain elevated at close to their cur-rent level over the next few years. At the same time,however, the share of debt held by households whodeclare insolvency (or that write-off debt) is likely toremain low at 0.6%. Ditto for mortgages in arrears,at 0.5%.

There are risks to this outlook. Historically, thenumber of bankruptcies have been tightly tied to la-bour market conditions, but as of late they have beenmuch higher than would be suggested by the rise in theunemployment rate. The number of bankruptcies percapita during this recession was 50% higher than the1990s recession despite the stronger performanceof the domestic economy and labour markets this timearound. And, despite a stunning recovery in Canadianemployment, the level of bankruptcies and insolvencieshave remained elevated likely a consequence of thelevel of debt. The implication is that one needs to bemore cautious when looking at the unemployment rateas the traditional driver of delinquencies, as greateremphasis is likely required on the level of indebtedness.

share of households had taken advantage of a quicker ap-preciation in home prices relative to that in Canada in orderto increase their borrowing further. At one point, 75% of mortgage renewals in the U.S. were taking on larger out-standing balances, as Americans were rapidly extractingequity from their homes for consumption purposes. WhileCanadians have also been extracting equity from their homesfor consumption, the trend has been far less pronounced andthe bulk of the debt accumulation has been largely associ-

ated with the purchase of a home and in particular

rsttime homebuyers jumping into the market.

What is the appropriate level of the Canadianpersonal debt-income ratio?

Although estimating the appropriate level of the debt-to-income ratio is not an exact science, we have developeda model that appears to provide good predictive power.Variables in the model include: assets as a per cent of PDI,the unemployment rate, core in ation, housing affordability(which would include changes in rules that increase amor-tization), the 5-year government bond yield (a proxy forthe 5-year mortgage rate) and home prices. According tothis model, household indebtedness can sustainably growin direct relation with a rising asset base, improving afford-ability and a lower jobless rate. The model suggests thatafter running more or less in line with its equilibrium leveluntil 2007, the debt ratio has since exceeded it. Applyingthe TD Economics base case economic forecasts for theseinputs over the next few years, a sustainable level of debt-to-PDI is estimated in the 138-140% range over the next 5years some 6-8 percentage points below its current level.

APPROPRIATE PATH OF CANADIAN HOUSEHOLINDEBTEDNESS

0

20

40

60

80

100

120

140

160

1980 1984 1988 1992 1996 2000 2004 2008 2012

debt as a % of pdi

Actual + TDEconomics

September forecast

Model estimationof optimal debt-to-

income ratio

Source: Statistics Canada/Haver Analytics, est. and forecast by TDEconomics as of October 2010

FCST

What does this sustainable range imply about thefuture path of borrowing?

Taken at face value, the current excess of household debtrelative to income implies that a considerable and protractedadjustment is required in order to bring the ratio back to a

HOMEOWNER'S EQUITY

35

45

55

65

75

1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010

U.S.

Canada

%

Source: Statistics Canada, Federal Reserve Board, Haver Analytics

Start of U.S.recession


10/13



10

an appropriate level. Consider that average annual growthin household debt has run at 8% per year over the past de-cade. In a world of moderate 4% annual average growth inPDI, average growth in household credit would need to beconstrained to about 2% per year in order to return the debtratio back to 140% within three years. Average annual debtgrowth of 3% would get you there in ve years. These slightrates of debt growth would be unprecedented in Canada ina non-recessionary period, and they are unlikely to occur inan abnormally low interest rate environment. However, wedo believe that debt growth might slow from the 8% aver-age annual gain in the last decade to 5% per annum overthe coming ve years.

While not an exhaustive list, there are a number of in- uences that will provide a natural brake to credit growthover the next few years: Housing activity is likely to remain relatively subdued

in recent years, rst-time home buyers have accountedfor as much as half of purchases, up from their long-term average of about one-third. With many rushing

to get ahead of higher rates, and the pool of

rst timebuyers largely exhausted in our view, housing activityis unlikely to return back to its recent peak over thenext several years. The slowing in the housing marketwill feed through to other types of big-ticket consumerpurchases and overall demand for credit.

Capacity to borrow is likely to be more constrained the consequence of being above a sustainable level of indebtedness is that the capacity to take on more debtis constrained. First, with increased usage during therecession, the available credit on home-equity lines

has fallen relative to pre-recession levels. Second, aswe have discussed, as interest rates head up gradually,households will have to devote a greater share of theirincome to paying their monthly debt obligations.Higher interest rates will also diminish the numbersof individuals qualifying for credit.

Appreciation of asset values will be more moderate going forward, we expect that households will notbe able to leverage rapidly growing asset values tothe same extent as over the past decade. In viewof the widespread belief that economic growth willbe only gradual, the pace of corporate pro t growthin the coming years is likely only to support equityreturns of 6-8% over the long haul, almost half the rate

experienced over the last decade. Meanwhile, homeprices are expected to grow at their long-run averageof about 4% in the coming decade, which is also closeto half of its trend rate before the recession.

Structural supply in uences likely to provide less of aboost . Some of the changes to the mortgage insurancerules were reversed. The maximum amortization periodwent from 40 years to 35 years, and the required downpayment went from 0% to 5% in October 2008. TheIn early 2010, the Federal Government also hardenedmortgage insurance quali cation rules. Banks are now

required to income test borrowers against the 5-yearposted mortgage rate for all mortgage vehicles of lessthan 5-year term and the 5-year contracted rated on all5-year mortgages, whereas banks had been using the3-year posted mortgage rate in the past. These changeshave eroded a quarter of the improvement in housingaffordability that occurred in 2007 with the looseningof the mortgage insurance rules. The minimum downpayment on non-owner occupied properties was alsoincreased to 20%. Finally, some of the kick to creditgrowth provided by the decade-long shift to exible

lines of credit may have run its course. Demographics might also temper credit growth, as

more baby boomers enter retirement. This wouldbe the typical conclusion from lifecycle modeling.However, a case could be made that the effect mightbe constrained by financial innovation. Given thehigher debt loads among individuals nearing, or in,retirement coupled with less retirement income anda low personal savings rate there might be greaterdemand for nancial vehicles that allow retirees towithdraw equity from their homes. For instance,

HOUSEHOLD ASSET GROWTH

-6

-4

-2

0

2

4

6

8

1012

14

2000 2002 2004 2006 2008 2010 2012

Forecast

year-over-year % change

Source: Statistics Canada, Haver Analytics, Forecast by TD Economicsas of October 2010

2000-2008 annual average


11/13



11

reverse mortgages may become more popular in thefuture. Financial institutions are likely to create newlending products to accommodate this demand, whichwould act as a new source of credit. Clearly, if such atrend took place, policy makers would need to considerprudent regulatory guidelines. However, this innovationis not likely in the next couple of years, but is plausibleover a longer time horizon.

If debt growth does slow to the 5% annual pace that TDEconomics anticipates over the next ve years, not onlywould it not fail to unwind the excess in personal indebted-ness present at the moment, but would aggravate it. The basecase forecast is for real economic growth of roughly 2% peryear, supporting personal disposable income growth of 4%annually. This mix of debt and income growth would seethe personal debt-to-income ratio climb to 151% by 2013 roughly 11 to 13 percentage points above our estimationof the sustainable level.

In our forecast, the moderate economic growth and sus-tained low in ation environment means that interest ratesrise slowly, with the overnight rate only returning to 3.50%in 2013 and holding at that level for some time. So eventhough Canadians facing debt service charges in excess of 30% will face a challenging period ahead and debt will beincreasingly excessive relative to income, debt will remainmanageable for the majority of Canadian households.

Incidentally, in the late 1970s and early 1980s, thehousehold debt-to-income ratio remained above its sustain-able level for about ve years. This period of unsustainableborrowing was followed by a drop in the debt-to-incomeratio, which remained below its long-run trend value fora subsequent three years. The catalyst for the adjustment

was a deep recession and a large spike in the unemploy-ment rate. The challenge this time around is that, short of a double-dip recession that we dont expect, the trigger toscale back household borrowing by more than in the TDEconomics base case forecast must be higher interest ratesthan currently projected or prudential actions.

What are the key risks that might lead to a harderlanding?

Due to our expectation of continued low interest rates,our base case outlook pushes the adjustment period out tothe second half of the decade.

In the near term, we are concerned about two negativerisks in particular: (i) a negative shock to income growth or(ii) a renewed wave of borrowing that could lead to a more

painful consumer nances adjustment down the road. Theodds of either event happening is material, but not highenough to be the most likely scenario. We would put theodds of either outcome at perhaps 1-in-3.

In terms of the rst risk, a major disruption to householdincome resulting from a double-dip U.S. recession or an un-anticipated nancial shock that would impact the Canadianeconomy and impact household nances. In contrast to the2008-09 global recession, the ability of Canadian consumersand governments to spend their way through the downturnwould be much more constrained, leading to a material

recession and a sharp increase in the unemployment rateto above 10%. Since many households do not have much nancial wiggle room, any signi cant disruption in incomecould cascade into larger delinquencies and a deleveragingby households. Under this scenario, we project that theappropriate debt-to-income ratio would fall to 127-128%by 2013 (see chart) and would require a greater degree of

HOUSEHOLD DEBT SERVICE RATIO UNDERVARIOUS ECONOMIC SCENARIOS

0

5

10

15

20

25

30

1999 2001 2003 2005 2007 2009 2011 2013

Base caseDouble-dip recessionIndebtedness continues current trend

debt payments as a % of income

Source: Ipsos Reid, Forecast by TD Economics as of October 2010

fcst

0

20

40

60

80

100

120

140

160180

200

1980 1985 1990 1995 2000 2005 2010

TD Economics September 2010 ForecastDouble-dip Scenario Sustainable LevelBase Case Sustainable LevelCurrent Trend Continues

Source: Statistics Canada/Haver Analytics, est. and forecast by TD

Economics as of October 2010

FCST

debt as a % of pdi

APPROPRIATE LEVEL OF HOUSEHOLD INDEBTEDNESSUNDER VARIOUS ECONOMIC SCENARIOS


12/13


13/13



13

This report is provided by TD Economics for customers of TD Bank Financial Group. It is for information purposes only and may not beappropriate for other purposes. The report does not provide material information about the business and affairs of TD Bank FinancialGroup and the members of TD Economics are not spokespersons for TD Bank Financial Group with respect to its business and affairs.The information contained in this report has been drawn from sources believed to be reliable, but is not guaranteed to be accurate orcomplete. The report contains economic analysis and views, including about future economic and nancial markets performance. Theseare based on certain assumptions and other factors, and are subject to inherent risks and uncertainties. The actual outcome may bematerially different. The Toronto-Dominion Bank and its af liates and related entities that comprise TD Bank Financial Group are not liablefor any errors or omissions in the information, analysis or views contained in this report, or for any loss or damage suffered.

Endnotes and References

1. The Ipsos Reid data are not widely available, and thus, are not frequently cited. Note, we use the Ipsos Reid data to look athouseholds who currently hold debt. The results from the CFM are not comparable to the Statistics Canada interest only debt-service ratio, because that estimate looks at all households those who hold and do not hold debt. According to the CFM, 30% of Canadian households are debt-free. Despite the methodological difference, the Ipsos Reid data tell the same story.

2. Using Ipsos Reids CFM data at less than at an annual frequency may lead to biases in the results due to small sample issues,as the sample size in a quarter is roughly a quarter the size of the annual sample. However, quarterly data provides a good leadingindicator of the direction of the data.

3. Faruqui, Umar, Indebtedness and the Household Financial Health: An Examination of the Canadian Debt Service RatioDistribution, Bank of Canada working paper, December 2008. http://www.bankofcanada.ca/en/res/wp/2008/wp08-46.html

4. Djoudad, The Bank of Canadas Analytic Framework for Assessing the Vulnerability of the Household Sector, Bank of CanadaFinancial System Review, June 2010 http://www.bankofcanada.ca/en/fsr/2010/index_0610.html

5. Bank of Canada Financial System Review, Jun 2010, http://www.bankofcanada.ca/en/fsr/index.html

Documents

TD Economics: Canadian Household Debt A Cause For Concern