RBA Annual Conference – 2000 Discussion

1. John Freebairn

Edey and Gower provide a comprehensive review of the numbers on Australian saving, the associated policy debates, and policy changes over the 1980s and 1990s. To me a very interesting and challenging policy question is whether Australia is saving too little, and if so why, and then what are the desired policy interventions?

Saving is not an end in itself, but rather a means to a higher goal of maximising consumption, and the utility it provides, over time. An important role for saving is intertemporal consumption smoothing for unanticipated adverse effects, over the business cycle (clearly documented for both private and public saving by Edey and Gower) and over the life cycle. It also has been argued that even for a relatively open capital market economy, including Australia, the observed close relationship between domestic investment and domestic saving (as illustrated in Figure 13 of Edey and Gower) due to high implicit transaction costs associated with different legal systems, distance and so forth, requires Australia to save to fund current investment. However, given the quite wide swings in Australia's dependence on overseas savings, from 2 to 6 per cent of GDP, it seems reasonable to ignore the investment argument for savings and concentrate the discussion on the use of savings for intertemporal consumption smoothing.

Edey and Gower argue that Australia saves too little by drawing primarily on comparisons with other OECD countries, the trend decline in the national saving rate, the larger and growing current account deficit, and the projected ageing of the population. They also point to the myopia and short-termism of individuals. In my view there are also a number of important policy distortions which reduce the incentive to save and reward from private saving.

Australia's system of means-tested aged pensions, cards and other retirement income support result in high effective marginal tax rates on saving for retirement years. The withdrawal rate associated with means testing of the aged pension is 40 per cent (reduced from 50 per cent in the 1999 tax reform package). Further, as shown in Table 5, even individuals on double Average Weekly Ordinary Times Earnings (AWOTE) can expect to receive a part-aged pension. Presumably the system of compulsory superannuation can be justified in part as a policy response to force individuals to increase their saving for retirement. Even here, in addition to the questions of adequacy of the 9 per cent rate discussed by Edey and Gower, policy could move much quicker to remove leakage via early access to superannuation and the current preferential tax treatment of lump sums versus annuities. Also, even if the present scheme might be satisfactory for full-time workers, it is much less likely to be satisfactory for the growing numbers of part-time workers and those with broken careers, including many women.

Important motives to save include making provision for private consumption of health and education services and for consumption in retirement. In Australia, however, governments provide from taxation revenue at least a basic level of health, education and retirement incomes. Given this backstop, the incentive for private saving is greatly diminished, and especially for those on low and middle incomes. Current governments fund these services from current revenues on a pay-as-you-go basis. With governments taking primary responsibility for social expenditures, and projected increases in these outlays in the coming decades, it is arguable that governments should be net savers over the business cycle rather than the present federal government stance of fiscal neutrality over the cycle. Just how big the surplus should be requires more assessment.

The Australian tax system likely distorts intertemporal consumption and saving decisions towards too high a level of current consumption. While the tax system is titled an income tax system, in fact some savings receive a consumption base treatment, including savings invested in owner-occupied housing and business investment in human capital. Others receive close to a consumption base treatment, including superannuation; and there are concessions for the returns on other forms of saving, including the benefits of deferral and lower tax rates on capital gains. Granted these facts, together with the relative importance of these forms of savings, it remains the case that many savings options receive an income tax treatment, including saving via financial instruments and business equity. To the extent these income-taxed forms of saving represent marginal saving, the present tax system distorts decisions against private sector saving.

Assessment of the net effect of compulsory superannuation on private sector saving is a difficult task for all the reasons discussed by Edey and Gower. My guess is that the problems of quality of data and of accounting for the effects of many other determinants of saving make it unlikely that econometric estimates based on macroeconomic data will prove a fruitful avenue for research. Extending the simulation modelling work of Gallagher (1997), and in particular evaluating the robustness of results to likely ranges of key parameters, seems a useful research strategy.

A better understanding of the path of private saving over the 1990s, and peering into the future, seems worthwhile by splitting private saving into the household and enterprise components together with the material on balance sheets provided in Figures 3 and 4 of the Gizycki and Lowe paper (this volume). Using this disaggregation is not to deny the valid points made by Edey and Gower that in general it is better to consider private saving as an aggregate. In part, the marked growth of private saving by enterprises in the 1990s is associated with what Gizycki and Lowe term restoration of their balance sheets after the excesses of the 1980s. To the extent this balance-sheet restoration has been completed, as is indicated in their Figure 4, it seems unlikely that enterprise saving will contribute in the coming years to private saving in the way it did in the 1990s.

The decline in measured household saving in the 1990s may be explained partly by the dramatic increases in both household debt and household financial assets (Figure 3 of Gizycki and Lowe). In part there is a measurement issue with capital gains excluded from the measures of household income and saving. A potentially fruitful area for further analysis of household consumption and saving decisions is the inclusion of balance sheet assets and liabilities in the explanatory model. Poterba (2000) provides an excellent review of these issues for the US. In particular, it would be useful to evaluate the likely order of response of household consumption to a sharp fall in equity (and home dwelling) prices should the asset bubble break.

To conclude, combining the near completion of the compulsory superannuation levy at a rate of 9 per cent, no further significant build-up of enterprise financial assets, and little reason to anticipate a marked increase in the household saving rate implies that the private sector saving rate is more likely to fall rather than rise over the next decade.

References

Gallagher P (1997), ‘Assessing the National Saving Effects of the Government's Superannuation Policies: Some Examples of the New RIMGROUP National Saving Methodology’, The Fifth Colloquium of Superannuation Researchers, University of Melbourne, Conference Paper 97/3.

Poterba JM (2000), ‘Stock Market Wealth and Consumption’, Journal of Economic Perspectives, 14(2), pp 99–118.

2. General Discussion

The discussion of the paper by Edey and Gower focused primarily on the causes of the recent decline in national saving in Australia. Participants debated the effect of compulsory superannuation on private sector saving. There was also considerable discussion of the implications of demographic change for national saving.

In discussing the possible causes of the decline in private saving since the mid 1970s, the various policy distortions identified by Freebairn in his discussion of the paper were considered. One participant made the point that Australia's system of state-provided health care and subsidised education (through the Higher Education Contribution Scheme, for example) may be one reason why households in Australia save less than most OECD countries. Another pointed out that prior to financial deregulation, households needed to establish a saving history in order to get a loan for house purchase. The relative ease with which home loans can be obtained post-deregulation might be an important factor. It was noted that the high effective marginal tax rates on pensions implied by the system of means-tested aged pensions may discourage private saving. A few participants also wondered if the introduction of mandatory superannuation in 1992 had reinforced the view that one need not save for retirement.

Some participants argued that household consumption and saving decisions have recently tended to be significantly influenced by asset market developments. In particular, they pointed out that households are increasingly spending out of capital gains from financial assets, and suggested that a closer examination of household balance sheets might help explain why private saving in Australia is relatively low.

On the issue of superannuation, many felt that the current system – a target mandatory contribution of 9 per cent of gross salary by 2002/2003 – would not provide a sufficient boost to national saving. One participant expressed concern that a substantial portion of the contribution is spent on administrative costs. Citing the Chilean system as being a good model, this participant argued that employers should be required to contribute savings net of administrative costs. This would not only contribute to enhancing national saving, but would also give employers strong incentives to minimise administrative costs. Some participants also remarked on the limited coverage of the current system, pointing out that it does not cover employees earning less than the exemption threshold of $450 per week. One participant made the point that the primary policy objective of compulsory superannuation was to provide retirement income, not necessarily to boost national saving, and that the system had in fact been very effective in augmenting retirement income.

Finally, the implications of the ageing of the population for national saving, and particularly for public-sector saving were considered. Some participants remarked that the combined effect of increased pension liabilities and health-care costs could lead to a substantial decline in public-sector saving. At the same time, it was noted that the problem of an ageing population is perhaps not as pronounced in Australia as it is for many other OECD countries.