RDP 2013-04: Home Prices and Household Spending 5. Conclusion

We use a household-level dataset, the HILDA Survey, to explore the relationship between home prices and household spending in Australia. Three main arguments have been put forward in the literature to explain the apparent co-movement between home prices and spending: (1) a ‘traditional wealth effect’, whereby spending rises with home prices due to an increase in households' lifetime resources; (2) the removal of credit constraints, whereby spending rises with home prices due to households' ability to borrow more, given more valuable collateral, and the related buffer-stock savings argument, whereby higher home prices act as a form of precautionary savings for low-saving households, allowing them to increase spending; and (3) that spending and home prices move together due to a common third factor, such as changing perceptions of lifetime income.

Our analysis most strongly supports the second explanation – that credit constraints and/or buffer-stock saving are the vehicle through which home prices affect spending. At both the cohort and household level we find that the spending by younger (and so more credit constrained) households is more responsive to changes in home prices than that of older households. This argues against the traditional wealth effect hypothesis; this wealth effect should be stronger for older households who typically own more housing than they will need over their remaining lifetimes. We also find that young and middle-aged homeowners respond more than young renters to rising home prices. This argues against the explanation of a common third factor, since renters and homeowners should both be affected by non-home-price shocks, for example increased income expectations.

By analysing the same dataset at two different levels of aggregation, we are able to assess the effect that aggregating data has on model results. We find that household-level and cohort regressions imply very similar spending reactions in response to a change in home prices. This suggests that ‘pseudo-panels’ are a reasonably good substitute for actual panels, although the necessary use of aggregate home prices in pseudo-panels seems to inflate estimated wealth effects.