RDP 2006-08: A Survey of Housing Equity Withdrawal and Injection in Australia 7. Aggregate Implications of the Survey

Thus far, we have concentrated on the microeconomic results for 2004 arising from the survey. This section aims to draw some aggregate implications from these results. First, we consider what the survey results imply for aggregate flows of housing equity over 2004. Second, we consider factors contributing to movements in aggregate housing equity withdrawal over time. Finally, the implications of housing equity withdrawal for key uses such as consumption over time are considered in light of the survey. As the survey was only for 2004, inference on earlier periods assumes that the findings are broadly representative of how equity was withdrawn and used in other years.

7.1 Aggregate Flows of Housing Equity in 2004

The sample results were aggregated to economy-wide flows by multiplying each household's net injection or withdrawal by the frequency weight attached to that household (that is, the number of Australian households the respondent household represents). Housing equity flows over 2004 based on aggregated survey responses suggest that:

  • households that were net withdrawers of equity by increasing debt on already-owned property withdrew around $20 billion;
  • households that were net injectors of equity by reducing debt injected around $28 billion;
  • households that were net injectors of equity primarily through renovations injected around $16 billion;[16] and
  • households engaging in property transactions were responsible for the largest flows of equity. Of these, net equity withdrawers extracted around $53 billion, while net equity injectors added around $43 billion.

Combining these, the survey findings suggest a net equity injection of around $13 billion in 2004. This contrasts with the favoured aggregate measure, which shows a net withdrawal of $17 billion.[17] Given the vagaries of measuring the flows involved, both at a household and aggregate level, it is not unexpected that the measures do not line up, though it is a caveat to bear in mind.

7.2 Housing Equity Flows Over Time

Section 4 shows that, over 2004, the largest aggregate flows of housing equity came from households transacting in the housing market. The typical housing transaction gave rise to net equity withdrawal, with vendors tending to have less debt remaining than was taken on by buyers, a pattern likely to be exacerbated by a period of rising house prices.

These findings suggest that movements in turnover and house prices are important for movements in housing-secured credit and aggregate housing equity withdrawal, a point borne out by the data. Figure 4 shows that the turnover rate of the national housing stock rose consistently over the mid to late 1990s, reaching a high level in 2002 and 2003 – a period in which housing equity withdrawal was also strong. Turnover then fell sharply through 2004, at the same time as housing equity withdrawal declined. Similarly, nationwide house prices rose rapidly up to late 2003, but have subsequently increased only modestly.

Figure 4: Drivers of Housing Equity Withdrawal
Figure 4: Drivers of Housing Equity Withdrawal

Note: (a) Per cent of household disposable income
(b) Per cent of dwelling stock

Sources: ABS; APM; Australian Treasury; RBA

Another relevant consideration for housing equity flows is the activity of property investors. The share of housing loan approvals made to investors rose from around ⅓ in 2000 to a peak of around 45 per cent in 2003, followed by a subsequent decline. This may have contributed to rising housing equity withdrawal up to 2003 because, according to the survey results, investors tend to purchase with relatively higher LVRs.

The survey results suggest that flows of housing equity due to non-transactors are of less importance. Nonetheless, partial data on these flows, where available, are also consistent with developments in aggregate housing equity withdrawal in recent years. The survey identifies mortgage refinancing as one of the main methods of withdrawing equity by non-transacting households. Australian Bureau of Statistics (ABS) data on refinancing of owner-occupier mortgages show rapid growth in loan refinancing during 2002 and 2003. In addition, borrowing through home-equity line-of-credit products increased by more than 30 per cent over 2003, before slowing. Movements over time in equity injection by non-transactors, however, are difficult to gauge, with various influences likely to have shaped any overall trend in principal repayments over recent years. These include ongoing growth in wealth and income, the increased share of interest-only loans, and flexibility of many mortgage products.

7.3 Housing Equity Flows and Economic Activity

The survey results suggest that movements in housing equity withdrawal need not be associated with large swings in consumption. To the extent that property transactions are a key driver of movements in net housing equity flows, and the bulk of equity extracted from transactions appears to be used to acquire non-housing assets, changes in housing equity flows are likely to be only partly reflected in changes in consumption. Nevertheless, it remains likely that the trend rise in equity withdrawal evident for much of the past 10 to 15 years has been one of the factors supporting strong growth in consumption over that period.

For 2004, the results suggest that around 18 per cent of the aggregate equity withdrawn by net withdrawers was used for consumption, which represents around 2½ per cent of the level of aggregate household consumption. However, this estimate may understate the amount of gross withdrawals used for consumption (see Footnote 15).

The static nature of the survey means that it is not possible to assess contributions to growth from the survey data alone. Nonetheless, it seems likely that the strong growth in housing equity withdrawal over 2001 to 2003 contributed to strong growth in consumption relative to income (and a corresponding decline in the saving rate) over that period (Figure 5). Trends in aggregate financial variables over that period are also consistent with the survey findings on uses of withdrawn equity. Flows into financial assets were above average, and personal credit growth was well below that of housing credit, consistent with households withdrawing housing equity as a substitute for other debts. These trends have subsequently abated.

Figure 5: Selected Uses of Household Funds
Per cent of household disposable income
Figure 5: Selected Uses of Household Funds

Source: ABS

Another channel through which swings in household borrowing affect economic activity is spending on renovations. Borrowing to finance this form of spending does not necessarily lead to a withdrawal of equity, if the borrowed funds are used solely to increase the value of the household sector's housing assets. Nevertheless, the effect on overall activity can be significant. Over recent years, annual spending on renovations has averaged around 4½ per cent of household disposable income, up from an average of around 3½ per cent between 1990 and 1998. The survey data suggest that, in many cases, renovations have been partly funded by drawing down on the equity built up as a result of the large house price increases since the mid 1990s. Around 11 per cent of surveyed households spent money on renovations in 2004, with the median amount spent on the main home equal to $14,000. Around 40 per cent of these households used housing debt to at least partly finance their renovation expenditure, with debt finance being used more often for larger renovations.


These households accounted for around half of overall renovation spending identified in the survey. Remaining renovation spending was dominated by other equity actions, and is captured in other categories. [16]

See Appendix A for discussion of the aggregate measure. The discrepancy between the survey and aggregate data is likely to partly reflect survey respondents reporting less debt than is suggested by aggregate figures, a feature also observed in household surveys in other countries. See Redwood and Tudela (2004). [17]