Statement on Monetary Policy – May 2006
Box D: Capital Gains and Measures of Household Saving
The most frequently-cited measure of household saving is the net saving rate, which is measured as a residual by subtracting consumption and depreciation from household disposable income. On this measure, the saving rate of Australian households has been declining, and in recent years has been negative (Graph D1). This is lower than in many other developed countries, though not dissimilar to developments in the US. Adding back the amount deducted for depreciation (most of which is on housing) lifts the saving rate, though the position of Australia and the US remains below average. These conventional measures of saving are calculated from the household income account in the national accounts and do not include in income (and therefore saving) that part of the return on investments that comes in the form of capital gains. In the case of equity investments, for example, only the return that is in the form of dividends is included as income even though historically this has accounted for only about one-third of the overall return on equities.1
In countries such as Australia where the composition of household investments has shifted from bank deposits and fixed income to equity holdings, this measure of saving has increasingly understated the true extent of household saving. In such countries a more meaningful measure of household saving is the change in household net financial wealth, defined as household financial assets (bank deposits, bonds, equities and unit trusts) less non-housing debt. This measure, which can be directly measured using financial data, focuses only on financial wealth and abstracts from wealth in the form of dwellings and the household debt used to fund that investment in dwellings. To the extent that households borrow against dwellings to buy financial assets (i.e. engage in housing equity withdrawal) the measure of financial wealth used here would be overstated, but on average over the 1990s the impact of this has not been very significant.
Graph D2 shows net financial wealth of Australian households, per head of working-age population. It has risen from $32,000 in 1991 to $90,000 in 2005. Also shown on the graph is the increase in wealth that would be implied by the cumulation of (net) saving as measured in the national accounts. This latter series shows a much smaller increase, with the bulk of the gap between the two lines due to the compounding of capital gains on equities which, as noted, are not captured in the conventional saving rate.
Graph D3 shows a broader measure of the saving rate calculated from household net financial wealth – i.e. as the annual change in net financial wealth expressed as a proportion of household disposa ble income. This broader measure is more volatile than conventional measures because of year-to-year changes in equity valuations, but it is also significantly higher on average and does not exhibit the same downward trend. This measure, which is more rele vant to a country where households maintain a large proportion of their wealth in equities, pre sents a more realistic picture of Australian household saving behaviour than conventional measures.
It is also interesting to note that on this broader measure of the household saving rate, Australia is not out of line with international experience. Countries which have relatively high household saving rates as conventionally measured (e.g. countries in Europe) tend also to be those where household investments are mainly in bank deposits and fixed income, whereas in countries with low household saving on conventional measures (of which the US and Australia are examples) households hold a greater proportion of their investments in equities. Once the capital gains in these are included in saving measures, the international dispersion in saving rates is reduced (Table D1).
- See ABS Australian System of National Accounts, 2002â€“03, Cat No 5204.0, ’New analytical measures of income, saving and wealth’, for more detail of the shortcomings of this saving measure.