9.5  The impacts of KiwiSaver on saving (7.3)

The Australian experience

One way of seeing what a compulsory KiwiSaver scheme might do for saving is to look at what has happened in Australia. Australia has had compulsory superannuation as a tier 2 scheme for around 20 years now. Australia also has a tier 1 pension, but unlike New Zealand's, this is means tested. Also, all of the regular contributions to the superannuation scheme are paid by employers, unlike the situation in New Zealand for KiwiSaver where contributions are paid by both employees and employers, with a tax credit contribution from government. Employers' contributions in Australia are currently set at 9% of an employee's wages.

The definitive study on how the tier 2 scheme has affected household saving still seems to be Connolly and Kohler (2004). They found that that of every dollar contributed to the scheme, around 50 to 60 cents of this was additional household saving. The remainder of the contribution was due to “reshuffling” with households transferring saving that was formerly done in other areas into superannuation. The authors found that in 2001/02 household saving would have been around 1.5 to 2.0% of GDP lower if the compulsory superannuation scheme had not been in place.

We have used the Connolly and Kohler methodology to get an estimate for 2008/09.[48] The additional saving is estimated at around 2.3% of GDP. However, this is the increase in household saving, not national saving. The impact on national saving is likely to be lower, since government provides taxation concessions for superannuation contributions, taxing them at only 15% compared to normal marginal rates, which would probably average somewhere around 30%. These tax concessions were equal to around 1.5% of GDP in 2008/09. So if the tax concessions resulted in government saving being lower by 1.5% of GDP, then the final impact on national saving would be 0.8% (2.3% minus 1.5%).[49]

Even if the impact on national saving was only 0.8% of GDP, then the impact would still be substantial. This would mean that the annual current account balance was 0.8% of GDP higher than it would have otherwise been, and this would have a similar impact on net foreign assets. Over a 10-year period, net foreign assets, as a share of GDP, would have been 8% of GDP higher than they would have been without the compulsory superannuation scheme.[50]

Australia's net foreign assets are currently around minus 58% of GDP, compared with New Zealand's minus 85%. However, there are likely to be other factors besides superannuation which account for the difference between the two countries.

It might be argued that the biggest impact from the Australian superannuation on national saving is yet to come. As noted earlier, the Australian tier 1 pension is means tested. At present only 55% of people over 65 get a full tier 1 pension, with the remainder getting part pensions or no pensions. However, as the superannuation scheme matures and a larger proportion of the population receive substantial retirement incomes from this source, the proportion of the population getting a tier 1 pension will decline. By 2050, it is estimated that only 28% of the population over 65 will be receiving a full tier 1 pension. This means that government spending will be substantially lower than it would have been without the compulsory superannuation scheme.[51]

Another positive aspect of Australia's compulsory scheme is that a high proportion of savings is directed via financial institutions. This is likely to result in better performing investments than those made by the household sector directly. For example, the scheme provides a channel by which the household sector can tap into higher earnings in overseas economies. Overall, the Australian scheme probably results in higher returns for households relative to those achieved by New Zealand households.

The estimated impact of KiwiSaver on saving

Results from a survey of KiwiSaver members undertaken in March 2010 indicate that just under 40% of the contributions being made by members are truly additional saving with the rest being due to reshuffling of other types of saving.[52] We used this number to estimate KiwiSaver's impact on household saving.

We did this for a number of scenarios.[53] In the base scenario – scenario A – it was assumed that contributions to KiwiSaver remained as they are now – 2% from the employee, 2% from the employer, and a contribution from government that matches the employee's contribution, up to a maximum of $1043 per year. It was assumed that the KiwiSaver membership rate would rise to 60% of all employed persons. It is currently around 40%. The 60% level was thought to be a reasonable assumption regarding the maximum rate for a voluntary saving scheme.

The results showed that KiwiSaver contributions to household saving were equal to 1.7% of GDP. However, this included employer and government contributions and tax credits as well as household contributions.[54] In terms of “truly additional” household saving – estimated as 40% of the household sector's contributions - this was equal to only 0.2% of GDP (Table 6). This is the impact of the household sector's contributions on national saving.

It will also be the total impact on national saving if we assume that the contributions to KiwiSaver from both government and employers do not affect national saving. This assumption assumes that in the government's case, for example, the government's KiwiSaver contributions result in a rise in household saving but that this is offset by corresponding decline in government saving. Hence there is no impact on national saving. The situation is similar regarding employers contributions.

However, we tried using an alternative assumption, on the basis that both the government's and employers' contributions to KiwiSaver will result in changes to government and business consumption as well as saving.[55] This resulted in a significantly higher figure for the impact on national saving -0.8% of GDP. However, in view of the uncertainty regarding the assumptions behind this result it was decided to use results based on the initial approach in comparing scenarios.

Given that government's contributions to KiwiSaver in scenario A are equal to around 0.5% of GDP there is a question as to whether government would be better off saving this money directly – for example by paying down government debt – rather than putting it into KiwiSaver. However if government spent this money rather than saved it, this would end up actually reducing national saving.

Source: SWG Secretariat

We also looked at a compulsory scenario – scenario B – with the KiwiSaver membership rate set at 100% of the employed. This lifted the impact on national saving from 0.2% to 0.4% of GDP. This is still not large, but it would increase net foreign assets by around 4% of GDP over a 10-year period and 8% of GDP over a 20-year period.

With compulsion, government contributions can be dropped, since they are no longer needed in order to make joining KiwiSaver attractive. Dropping these contributions does not affect our estimates of the impact on national saving, since the assumption is that the resulting fall in household saving will be offset by a rise in government saving. As noted above though, this may not be the case in reality.

In another compulsory scenario – scenario C – we increased the employees' contribution rate to 4% of income. This lifted the impact of contributions on national saving to 0.8% of GDP. This would increase net foreign assets by around 8% of GDP over a 10-year period and 16% of GDP over a 20-year period.

Note that we have looked here at only the impact of household contributions on national saving and net foreign assets. There is another effect on national saving over and above this. This is the effect of investment earnings from the household sector's “additional” contributions. These investment earnings are in fact a part of total household saving (the flow measure). Also, there is a compounding effect with respect to these earnings since they are reinvested by superannuation funds.

The extent to which these compounded earnings affect household saving and national saving will depend crucially on the returns that superannuation funds produce and the fees that they charge. Relatively small changes in each of these can have large impacts on the size of the funds over the long term. This is behind the SWG's view that it is important to focus on developing a low-cost default scheme where KiwiSaver members are matched to the most appropriate type of fund for their age.

More work is needed in order to assess accurately the possible impact of KiwiSaver earnings on household saving, and on national saving and net foreign assets.

It seems that in its present form, KiwiSaver is basically an adjunct to NZ Superannuation, not a replacement for it. The Treasury has estimated the value of the annuity that a KiwiSaver member on the average male wage could buy with his KiwiSaver fund at the end of a 40-year working life. With employee and employer contributions at 2%, the annuity would be equal to around only 17% of the person's wages in the final year of work (Savings Working Group Secretariat 2011c). Clearly the level of aggregate contributions would have to rise substantially, if KiwiSaver was going to become the major vehicle for providing retirement pensions (and potentially creating the opportunity to scale back the universal pension).