Household Debt: RBA says She'll Be Right

The RBA does not appear to be too worried about rising household indebtedness.  A new paper by Tom Bilston, Robert Johnson and Matthew Read, "Stress Testing the Australian Household Sector Using the HILDA Survey" contends that:

Higher levels of household indebtedness are an endogenous – and expected – response to permanently lower nominal interest rates. ... Additionally, higher household indebtedness can facilitate consumption smoothing, consistent with lifecycle ... or permanent-income ... hypotheses. Furthermore, to the extent that higher household indebtedness increases entrepreneurship and access to further education, it may raise living standards and long-run economic growth. Nevertheless, higher household indebtedness can also amplify the effects of economic and financial shocks on households. ... 

Despite the increase in aggregate household indebtedness over the 2000s, the distribution of household debt was little changed. The share of households with some debt rose slightly over the 2000s, to be around 70 per cent in 2010. Higher-income households (those in the top 40 per cent of the income distribution) owed around three-quarters of household debt (Table 1); these households generally have the lowest debt-to-income and debt-servicing ratios. Similarly, the most asset-rich 40 per cent of households owed around three-quarters of household debt. 

Households where the head was prime working age (35 to 54 years) owed about 60 per cent of household debt. However, the share of household debt owed by older households rose slightly over the decade. This probably reflected a decrease in the rate of property downsizing, increased life expectancies and a trend toward geared property investment. Even so, because older households tended to be among the wealthiest households, their increased indebtedness did not necessarily reflect a rise in the household sector’s overall financial vulnerability.

Household debt generally appears to have been well collateralised during the 2000s. The share of household debt secured by property rose slightly over the decade, to be nearly 90 per cent in 2010 (Table 2). About half of household debt was for the purchase of owner-occupier property (‘primary mortgages’). However, the rise in the share of household debt secured by housing was due to an increase in ‘other’ housing loans, such as second mortgages secured against owner-occupier property (e.g. home equity loans) and loans for the purchase of investment property. The value of credit card and other personal loans as a share of household debt both fell slightly over the decade.

The paper conducts a stress test of the household sector and finds that: The results suggest that the share of households whose incomes are estimated to be less than minimum expenses (i.e. with negative financial margins) fell from around 12 per cent in 2002 to 8 per cent in 2010. These households tend to have lower incomes, be younger and live in rental accommodation; however, these groups tend to hold a relatively low proportion of 30 total household debt. Households that were more indebted did not necessarily appear to be more likely to have negative financial margins than households that were less indebted. This could be interpreted as evidence that the screening lenders carry out in assessing loan applications is effective. 

Lenders’ exposure to households with negative financial margins appears to have remained limited, with expected loan losses (based on the assumptions underlying our model and in the absence of any adverse shocks) increasing over the 2000s, but remaining fairly low. This increase occurred despite the share of households with negative financial margins falling over this period, implying that these households owed an increasing share of debt and/or held less valuable collateral relative to this debt. The limited increase in expected loan losses is despite a substantial increase in aggregate household indebtedness, as well as the impact of the global financial crisis on the labour market and asset prices. This suggests that aggregate measures of household indebtedness may be a misleading indicator of the household sector’s financial fragility.  

Although the stress-testing model used in this paper is relatively simple and relies on a number of assumptions, it generates plausible results in response to shocks to interest rates, the unemployment rate and asset prices. The results from the two stress scenarios considered – both of which incorporate a substantial increase in the unemployment rate and a substantial decline in asset prices – imply a high level of household financial resilience and limited expected loan losses for lenders. That said, the effect on expected household loan losses of a relatively severe stress scenario, under which unemployment rises, asset prices fall and interest rates are unchanged, increased over the 2000s, suggesting that the household sector’s vulnerability to macroeconomic shocks may have increased a little. However, expected loan losses are actually lower under the less severe of the two scenarios, which has rising unemployment and falling asset prices comparable to Australia’s experience during the financial crisis. This is due to the offsetting effect of lower interest rates, highlighting the potential for expansionary monetary policy to offset the effect of negative macroeconomic shocks on household loan losses.