RDP 9603: Australia's Retirement Income System: Implications for Saving and Capital Markets 4. Role of Superannuation in the Financial Sector

Assets of superannuation funds and life insurance offices have fluctuated mainly in a range of around 20 to 25 per cent of the Australian financial system in recent decades.^[12] They are currently around 26 per cent, having risen strongly in recent years, and this share could be expected to increase further in future decades as compulsory contributions accumulate. The historical importance of these institutions reflected the significant use of superannuation as a voluntary savings vehicle, as has been discussed above, and was in part a result of their tax-favoured status. There are currently over 100,000 superannuation funds in Australia, which range from the very large (the ten largest fund managers control around 60 per cent of the assets) to the so-called do-it-yourself (DIY) funds with only a few members.^[13]

Trends in the superannuation sector's overall size and its sources of funds are summarised in Figures 1 and 2. Broadly, the historical growth of the superannuation sector can be divided into three phases. The first phase, which ended in the early 1970s, was one of moderate and fairly steady growth. In the second phase, which comprised most of the 1970s, superannuation assets shrank relative to nominal GDP, largely reflecting poor earnings performance and high inflation. The third phase, from the early 1980s onward, has been one of rapid expansion in which total assets more than doubled as a ratio to GDP, although this may have slowed down in the latest few years. The data presented in Figure 2 divide the sources of superannuation asset growth between net new contributions and a residual representing earnings on existing assets and capital gains. Although net contributions have fluctuated significantly in some periods, it is apparent that most of the variation in overall growth performance is attributable to variation in the earnings and capital gain component, rather than in contributions.^[14] The three growth phases outlined above correspond broadly to periods of moderate, negative, and high real rates of return on financial assets, as summarised in Table 4.

On the basis of currently available data, aggregate net contributions to superannuation funds do not yet show the upward trend expected to result from the compulsory plan.^[15] A number of possible reasons can be given for this. First, there is likely to be a strong cyclical influence on net contributions. They fell substantially in the recession of the early 1980s, when withdrawals related to early retirements were likely to have been particularly important. This may again have been a factor in the early 1990s. In addition, many voluntary schemes contain a tranche of employee-contributed funds which do not have to be preserved to retirement but can be withdrawn on leaving a job.^[16] There is also provision to allow early withdrawal of funds in cases of hardship. For all these reasons, recessions can be expected to result in significantly increased withdrawals from superannuation funds as jobs are lost. Second, many employers were already satisfying, at least partly, the requirements of the compulsory plan under pre-existing voluntary arrangements. This has allowed some scope for absorption of the compulsory scheme into existing arrangements, and has meant that the aggregate effect of the new compulsory schedule has so far been relatively small; but it can be expected to increase as the mandatory contributions rate increases significantly above levels currently prevailing. Third, an important factor in the second half of the 1980s was the phenomenon of overfunding of existing defined-benefit schemes. High rates of return meant that surpluses were accumulated in many of these schemes, enabling the employers who sponsored them either to withdraw funds, or to finance their superannuation liabilities with reduced contributions. Finally, it is possible that increased tax rates on superannuation savings after 1983 have discouraged voluntary contributions.^[17]

These factors provide a useful qualitative explanation for the behaviour of aggregate contribution rates. However there is no direct way of measuring their quantitative impact and thus arriving at some measure of an ‘underlying’ trend in contributions. This is an important issue for further investigation since, as discussed below in Section 5, the capacity of the scheme to meet its objectives hinges critically on its compulsory nature and on the ability to discourage unintended leakages.

One important dimension of this issue is the growth of ‘rollover’ funds, created in 1983 as a vehicle for deferring tax liabilities by preserving withdrawn benefits within the tax-favoured system.^[18] Funds withdrawn as a result of leaving a job can be deposited in a rollover fund until required to be drawn upon, and continue to be treated for tax purposes like other superannuation funds. They can also be moved from one such fund to another at the discretion of the member. Rollover funds are a relatively small component of the superannuation system by assets (around 5 per cent in 1995) but, because they are mobile at the member's discretion, they are responsible for a large part of the gross flows illustrated in Figure 3. Part of the impetus for this increased turnover in the early 1990s probably came from increased redundancies and early retirements.

Assets of superannuation funds are invested across a wide spectrum of traditional investments, with no important portfolio restrictions other than a limit of 10 per cent on the proportion of funds that can be invested with the sponsoring employer. Investments in the broad categories of equities, bonds and property are shown in Figure 4. The predominant trends have been a substantial reduction in the portfolio share of bonds and a rise in that of equities over the past three decades. Property investments had also been on an upward trend over much of the period but fell sharply at the end of the 1980s and in the early 1990s, largely reflecting valuation effects following the collapse of the property market. The long-term reduction in bond portfolios is likely to have been a consequence of removal of earlier portfolio restrictions setting minimum holdings of government bonds,^[19] along with a trend decline in public sector debt ratios which reduced the available supply. Holdings of foreign assets are not separately shown on the figure as consistent data are unavailable for much of the period. However, their portfolio share has grown rapidly in recent years and is currently around 13 per cent. A more detailed snapshot of the asset allocation as at end 1995 is presented in Table 5.

The superannuation sector is projected to expand considerably in future decades as the compulsory increases in contributions take effect. One estimate suggests an approximate doubling of the sector in relation to GDP, from 40 to 76 per cent of GDP by the year 2020.^[20] This policy-induced expansion raises a number of issues concerning the competitive position of superannuation within the financial system and the size of superannuation funds in the markets in which they operate. Some observers have argued that growth of the superannuation sector will in some degree occur at the expense of banks, or will occur in a way that increases competitive pressure on banks.^[21] Another issue is the possibility that the superannuation funds will ‘run out’ of domestic assets to purchase as they expand, or that their holdings of such assets will grow to a point where they significantly change the characteristics of domestic asset markets. These issues are closely related to the question of how effective compulsory superannuation will be in generating additional saving rather than displacing existing forms of saving. To the extent that new saving is generated, it could be expected to lead to a general expansion of the financial system and of the supply of domestic assets, along with an accumulation of foreign assets, rather than drawing funds from other domestic financial institutions.

A good general case can be made that there has in the past been relatively little competitive overlap between banks and the superannuation sector, although in some respects this competitive separation seems to be breaking down, particularly on the liabilities side. On the asset side of these institutions' balance sheets, the competitive separation has been strong. Superannuation funds invest primarily in securities while the traditional core business of banks is in non-securitised lending.^[22] Banks' traditional lending activities now represent a declining proportion of their balance sheets and profits, but this is part of a worldwide phenomenon related to improvements in financial technology associated with securitisation,^[23] and does not particularly seem to reflect competition arising from the growth of superannuation funds. While the trend of increasing securitisation seems likely to continue, the potential erosion of banks' competitive position with respect to traditional lending can easily be overstated. As noted by Tease and Wilkinson (1993), banks continue to have a natural specialisation in borrower risk assessment, and this is likely to remain important even when loans are increasingly in securitised form.

There is also a clear difference between the liability structures of these two classes of financial institutions. Superannuation fund liabilities are the long-term savings of their members, whereas bank liabilities are a combination of transaction balances, short-term savings and marketable debt instruments. As is documented by Edey, Foster and Macfarlane (1991), the banking system in Australia has not traditionally been an important vehicle for longer term saving, and the shorter-term balances held by households with banks bear a fairly stable relationship with household income. These balances do not seem likely to be closely substitutable by compulsory superannuation balances. Nonetheless, the competitive separation between banks and superannuation funds on the liabilities side seems to be breaking down at the margin. One important aspect of this is the growth of rollover funds, which are tax-favoured superannuation vehicles but which do have some of the characteristics of shorter-term savings, since their funds are highly mobile and not necessarily locked in for long periods. Also important is that the superannuation sector is itself an important provider of funds to other parts of the financial system. From Table 5, around $40 billion, or 15 per cent of superannuation assets are currently held as bank securities or deposits with financial institutions, a significant proportion of these institutions' liability base. Growth of these ‘wholesale’ sources of funds to the banks represents a potential source of upward pressure on their average cost of funds. However, the role of superannuation in this process should not be overplayed because it is part of a trend that would be likely to occur anyway, through the growth of money market mutual funds and the increasing sophistication of retail depositors.

These competitive issues have led some banks to move into the superannuation area by establishing life-office subsidiaries or forming partnerships with existing major life offices. More recently it has been announced that banks will be allowed to participate directly in some superannuation business by offering retirement savings accounts. Further issues concerning institutional distinctions between different parts of the financial sector, and their regulatory-policy implications, are the subject of a current government inquiry.

Footnotes

For statistical purposes it is useful to treat life insurance and superannuation funds as a single aggregate because their activities are similar and much of the historical data does not distinguish between the two. [12]

The situation is complicated by the fact that the major fund-management groups can run large numbers of separately constituted superannuation funds. [13]

Capital gains are likely, however, to be understated in the 1960s and 1970s, and overstated in the early 1980s, as a consequence of the widespread use of historical-cost valuations prior to the 1980s. [14]

These data should be interpreted cautiously, however, as they have in the past been subject to substantial revision. [15]

Recent regulatory changes restrict this right of withdrawal, subject to grandfathering of existing withdrawable amounts. [16]

There is also a serious longer-term policy concern: the potential for funds to leak from the compulsory scheme due to incentives favouring early retirement and dissipation of accumulated savings. See FitzGerald (1996). [17]

Following rule changes in 1992, rollover-fund operations as described here can now be carried out within ordinary superannuation funds. [18]

Until 1981, funds were required to hold at least 30 per cent of their assets as government bonds. [19]

Knox (1995). The estimates are for the superannuation sector excluding life-office business. [20]

See, for example, Thom (1992). [21]

This distinction is discussed in the Australian context by Tease and Wilkinson (1993). [22]

For a recent analysis of this global trend, see Bisignano (1995). [23]