RDP 2003-08: A Tale of Two Surveys: Household Debt and Financial Constraints in Australia 7. Conclusion

Using the 1998/99 HES data, a logit model was constructed to examine the relationship between the probability of being financially constrained and the economic and demographic characteristics of households in Australia. The proportion of Australian households that were cash-constrained in 1998/99 is around 22 per cent, which is broadly consistent with the findings of previous macroeconomic studies in Australia. Whether a household is constrained is significantly determined by demographic variables such as age, unemployment, disability and marital status. Important economic variables include home ownership, weekly household income, the proportion of income earned from interest and the share of income going to repayments on mortgage debt.

We also apply the model to the 1993/94 HES data and the 2001 HILDA data to determine whether the proportion of financially constrained consumers is likely to have changed over time and, if it has, to identify which economic and demographic variables may have been driving the change. Despite both economic and demographic variables being important determinants of cash flow constraints, only the economic variables have changed significantly between the sample periods. Relatively strong economic growth is predicted to have relaxed financial constraints through growth in earnings, growth in housing wealth and falling unemployment. Partly offsetting this, rising mortgage repayments (as a share of income) and increasing credit card debt are predicted to have increased constraints.

Putting these effects together at the aggregate level, we estimate that fewer households are now constrained. In 1994, 24.1 per cent of Australian households are estimated to have been constrained; this falls to 22.5 per cent in 1999, before falling marginally to a projected 21.6 per cent in 2001 (based on our preferred model). On balance, it appears that the overall level of cash constraints in the economy has fallen. However, the magnitude of the fall is small due to numerous offsetting effects.

To further examine changes in household fragility, we separate the sample into cash-constrained and unconstrained households and examine whether important changes have occurred within the two groups. The mortgage debt-service ratio has hardly changed for constrained households, on average, suggesting that they are no more fragile than in the past. And while there appears to have been a rise in the proportion of constrained households who have incurred at least one form of debt, the relative size of this sector of the community has fallen, as there are fewer constrained households in 2001 than in the mid 1990s.

Accordingly, despite the increase in the aggregate household debt to income ratio to historically high levels, we find little evidence that Australian households are now significantly more financially fragile than in the past. Much of the rise in debt appears to have been due to unconstrained households. There are now more unconstrained households in the population and it is these households that are primarily responsible for the increase in household debt. Indeed, the rise in the aggregate debt to income ratio seems to reflect households reacting to increased household income and low unemployment rather than being an indicator of increased household financial distress or fragility.

Hence, in summary, we find no evidence of an increase in financial fragility from the rise in debt associated with owner-occupier mortgages. This does not preclude an increase in fragility for those households that have significantly increased their exposure to investment housing. Information on this will become available with the next wave of HILDA data.