RDP 2001-08: City Sizes, Housing Costs, and Wealth 6. Conclusion

Economic researchers have long recognised the potential for demographic factors to drive medium-term outcomes in the labour market and financial asset returns. This paper argues that spatial aspects of demography are important for the level of non-financial wealth and housing costs. We argue that Australia's flat Zipf curve is a result of its federal political system and sparse population, which interact to produce multiple primate cities. The relationship between these primate cities arises because some serve as national centres to a greater extent than others, and therefore attract relatively more population from the small towns than do the smaller primate cities (state capitals).

The link from urban structure to average dwelling prices is more subtle. Given that large cities have higher housing costs, the argument that national average dwelling prices will be higher in countries with a larger share of their population concentrated in large cities would seem to be a matter of arithmetic. However, this cannot be the only explanation of the pattern of Australia's household wealth – otherwise, we would see Canada's wealth having a similar composition and Germany's less similar. Sweden would have had a much higher share of dwelling wealth had it not been for its large public and co-operative sectors holding market dwelling prices down.

If our arguments about the link between urban structure and national average dwelling prices are right, we would expect that dwelling wealth will be higher relative to income in Australia than in other countries, in the long run. This implies that housing debt-income ratios could be higher in equilibrium in Australia than elsewhere, without this being a cause for concern. If so, the rapid build-up of housing indebtedness over the 1990s may still have some way to go. Nonetheless, growth in housing prices and debt will still have to level out at some stage, to match nominal income growth in the long run. As has already occurred in New Zealand, the windfall gains that accrued to home owners over the 1980s and 1990s will ultimately end.

However, our estimate of the effect of urban structure on dwelling wealth accounts for only one-third – or at best, one-half – of the gap between Australian and US dwelling wealth-income ratios. This may suggest that dwelling prices are too high in Australia and must ultimately fall relative to household income. Fortunately, the prospect of a sudden crash in dwelling prices similar to that seen in the UK in the early 1990s seems remote. Australia's unusually high dwelling wealth-income ratio has built up over fifteen years, not in a brief period of speculation, and the specific circumstances that contributed to that boom-bust cycle do not apply here (Muellbauer 1992; Henley 1999; Bean 2000). It would seem more likely that such an adjustment would occur through an extended period of slow or zero growth in dwelling prices, and perhaps partly through a shift in the composition of the dwelling stock towards higher-density homes rather than prices of particular dwellings falling. Whether that adjustment caused financial distress in some segments of the household sector would depend on whether households had over-extended themselves, in the erroneous belief that dwelling prices would continue to grow strongly.

Given the urban structure of Australia and the concentration of high-income employment opportunities in just two cities, it seems likely that Australia will continue to have a pattern of household wealth being concentrated in dwellings. This has important implications for macroeconomic factors such as savings-investment balances, provision for retirement and the growth in and distribution of wealth.

Finally, the level shift in residential land values that occurred over the past 15 years must have had implications for intergenerational wealth distribution. Households that owned homes before financial deregulation experienced windfall gains in their dwelling wealth that will not be enjoyed by subsequent generations. While inheritance will even out some of this redistribution in the future, the difference between average age of home purchase and average age of inheritance implies that these changes will still affect wealth holdings and saving behaviour over households' life cycles, with home purchase possibly occurring later in life than is common today.