Putting the second leg back on the super stool

I would like to thank you for the invitation to address your annual general meeting, and to thank you also for the advance warning via the media a few weeks ago that some of your members would be seeking to give me a roasting today on the question of progress on policy related to incentives for retirement savings.

To misquote Dr Johnson, when a man knows he is to be roasted in a fortnight, it concentrates the mind wonderfully.

I am happy to meet the challenge head on, and would like to speak frankly about where I stand on the key policy questions and about where I see potential ways forward.

It is no secret that I have not seen eye to eye with my officials’ line of advice on this topic, much as I value the free and frank exchange of ideas. I do believe we have a problem with the level of private saving in New Zealand. Whereas the Treasury’s advice over a number of years has been more equivocal on this point.

Similarly, I believe that tax incentives can, if properly designed, have some effect on the level of private savings, where officials tend to be convinced sceptics on this point, arguing that incentives only alter the form that saving takes, rather than the total quantum of saving.

There are two important points, though, where there is little dispute:

• First, it is possible for tax incentives to be inefficient. Poorly designed incentives can impose a high cost for low levels of net increase in savings. And even well designed incentives can fall victim over time to changes in interpretations of tax law, evolving remuneration practices or other factors.
• Second, it is also possible for tax incentives to be inequitable. Those who have such low incomes that they cannot save do not have the potential to access the incentives, and a result the net effect can be to make the tax system more regressive.

From a policy point of view, any incentive regime has to pass muster on five basic criteria:

1. It must encourage genuine savings, not simply the temporary diversion of spending through vehicles that qualify for the tax break.
2. It must limit the switching of past savings into a tax-advantaged form.
3. It should increase net new savings into the future, and not simply divert what might be saved in the future down new avenues.
4. It must be broadly equitable (although this is difficult to achieve given that a precondition for an equitable incentive regime is that lower income groups have the capacity to save).
5. It should have a kind of "multiplier effect" in that it changes attitudes and habits, and contributes to a shift in the savings culture. Incentives that rely purely on a "bribe" effect tend to limit the net longer-term benefits that they produce, rather like the ploys parents use to induce their children to eat their greens.

It is also important to recognise that incentives only make a contribution to improved savings: they are not the total solution, and it would be dangerous to base a policy of encouraging a savings culture purely on an incentive scheme.

For example, the importance of direct education about savings opportunities (as opposed to generic advertising through mass media) is often overlooked as an ingredient in the savings formula. While New Zealanders are constantly receiving detailed, personally-targeted advice on our consumption options, the same cannot be said of our savings options.

I understand that in the USA, studies have shown that when workplace-based, small group programmes have been conducted focusing on how to develop a balanced portfolio of savings, the result has been that savings levels have increased, even when there has been no incentive like an increased employer contribution to retirement savings, or a tax break.

Another focus on recent research has been the impact that early personal experience with savings has on subsequent savings behaviour. This lies behind some of the "asset based welfare" initiatives that are being trialed in the UK. The programmes draw on empirical studies that suggest that if people have cash assets to look after, they get into a savings habit and start to save more themselves. This then raises the question of how early in life to try and kick-start a small savings portfolio, who pays for it, and how big it has to be to have a significant influence upon subsequent behaviour.

As far as I can tell, there have been no studies into whether or not there are reverse generation effects with asset based welfare: that is, whether kick-starting the savings of children gives parents an exposure to savings practices and influences their savings behaviour.

I am not going to dwell on these education and experience factors in improving savings habits, but simply note that they need to be a part of the equation going forward. A shift in attitudes towards saving is the cake; not the icing.

If I can draw an analogy with modern mental health treatment regimes, tax breaks function rather like drug therapies. They have a powerful effect, and can, as it were, manufacture the desired outcome quickly. But they are expensive, and tend to crowd out programmes aimed directly at achieving behavioural change through counselling, modelling positive behaviours and teaching families to recognise and respond to risk factors. Therapeutic drugs are an important part of a treatment regime; but they do not stand alone.

I would like now to consider the alternative designs for tax incentives against the criteria I mentioned earlier. There are several possible ways to design incentives, and each has its problems and challenges.

The up-front exemption of a set dollar figure per year lodged with a superannuation scheme is probably the worst incentive design. It is very hard to stop past savings or future savings migrating into the tax vehicle. Because almost no scheme can limit withdrawals over age 65, the effect will be, at a minimum, a diversion of the exempt amount through a tax vehicle by those in or near retirement, with a view to lowering their tax burdens, not increasing their savings. By definition, an exemption gives the biggest gain to those on the highest marginal tax rates, so it is particularly inequitable. Finally, because it encourages "tax farming" it is least likely to have a material impact on changing attitudes and savings habits.

A slightly better option is a rebate equal to a percentage of the dollar amount per year placed with a superannuation scheme – up to some maximum amount. It is better than a simple exemption in that it is more equitable. However, it does suffer from the disadvantage of being more expensive if it is going to be sufficiently attractive to leverage extra savings.

These considerations lead me to favour an incentive package that concentrates on trying to reinvigorate employment based retirement savings schemes.

In 1990, 311,000 employees – or 22.6% of the workforce – contributed to employee superannuation funds. Now, the number is 248,000, or 15% of the workforce. That is a 20% decline by number and a 33% decline in the relative proportion of the workforce participating in schemes.

I suspect that there are a number of factors at work. Partly it is the effect of a less attractive tax regime working its way through. Typically the new tax regime discourages new entrants to schemes, but existing members remain and only exit with retirement or change of employer. It has undoubtedly been affected by the decision to close the Government Superannuation Scheme to new members. The shift towards total remuneration packages – itself both the intention and the effect of the TTE regime – also contributed. Increased labour turnover and the decline of career service with a single employer may also have played a part.

If we want to reinvigorate employment-based savings, we need to respond to all of these factors – not just the tax change. It also has to be cost effective from a fiscal point of view. I stress that no decisions have been made; but I will outline the factors that I have asked to be taken into account with the policy development work that is going on.

The first goal is to try and make it more attractive for employers to offer a retirement savings plan, and to divert money that would otherwise have gone into salary to that form of remuneration. The payback for an employer is to reduce staff turnover and thereby raise productivity; but as the declining numbers of participants shows, this incentive has not been enough to maintain employer interest.

The bottom line here is that it will only be attractive to employers to apply part of a remuneration package as a contribution to retirement savings if the marginal rate of tax on that part is lower than the marginal rate of tax on an equivalent amount paid as a wage. At the moment, we tax those contributions at 33%, and so overtax contributions for those with incomes below $38,000 a year. The best estimate I have is that we overtax employer contributions to super by about $29 million a year – a not inconsiderable disincentive by any standard.

If we want to go further, it will cost money. There is no exact science in the numbers I am going to give you, because a lot depends on take-up rates if a new incentive is applied. As a rough guide, if we were to "cascade" the six cents marginal tax rate enjoyed on contributions for those on the 39 cent tax rate down through each of the marginal tax rates, it would cost around $80 million a year in forgone revenue.

I would be interested to hear the industry’s appraisal of what sort of savings response they might realistically expect from that sort of carrot.

The second thing that employers need to encourage them to offer schemes is a reduction in compliance costs. My impression is that progress to simplify disclosure rules was jolted by the Enron fiasco, and while the exercise has not halted, there is a pause for reflection.

Employers themselves have some responsibilities to design schemes to make them attractive to employees. In particular, vesting rules and an ability to remain in a scheme – or transfer to a new one without punitive transfer valuations – are crucial in an age where the expectation is that employees will change jobs from time to time.

Making schemes attractive to workers is the potentially expensive bit. Up front concessions are possible. Limiting concessions – either through exemption or rebate – to source deductions limits the amount of diversion of past savings into new tax advantaged vehicles. It is less clear that it will limit the amount of diversion that takes place out of future savings. Indeed, the leakage from a switch in the form of saving could be considerable.

On pure cost effectiveness grounds, I do not favour an upfront E for part or all of the employee contribution. I favour here a middle E – exempting from tax the earnings of the funds of schemes that register for the tax-advantaged regime. This sort of regime would need to be accompanied by the taxation of withdrawals.

I have looked at the option of applying rules on the amount that can be withdrawn as a lump sum, or at requiring benefits to be taken out in the form of an annuity; but the rules would be complex and could make the schemes inflexible and unattractive. It would be far simpler to tax withdrawals at the individual’s marginal tax rate in the year of withdrawal. This tends to both delay and spread the draw-down of balances, and puts the pressure back on the industry to design attractive products.

The argument is that although there is a potential cost through a delay in when the tax on the earnings of the fund is collected, to the extent that these are new savings, there would not have been any tax collected anyway. The tax deferral is the only fiscal cost involved.

A further point I would make about tax incentives is that there is scope for applying several different models to differing types of scheme. Schemes which have features that more stringently meet the criteria I outlined above could attract the most favourable tax treatment. Other schemes, which provide more flexibility, but have more risk of displacing other savings, could attract tax treatment which is not quite as favourable.

Finally, as I mentioned earlier, we need to consider how best to ensure that both employers and unions reinforce any new scheme offer with education programmes, and what information gathering systems we might put in place to monitor the impact on savings behaviour.

I also do not discount the importance of leadership and example. This is an area where there has been a tendency for individual employers to stand back and wait for somebody else to take the initiative. I have some discomfort in the fact that by closing the GSF, the government did in fact set a bad example.

There isn’t a simple solution to restoring employment based retirement schemes in the state sector. Many state employees are still contributing to the GSF. Replacement schemes have been established in some departments. Other state workers have had the equivalent of the employer contribution incorporated into their remuneration packages.

The government is, though, interested in working with state unions and departmental chief executives in charting a way back. A small, but significant start was made during the primary teachers pay negotiations when a part of the money that could have been incorporated into base pay was diverted into a retirement savings scheme.

Since then my colleague the Minister of State Services and I have set up a small working party to scope the potential to create a new scheme that would roll out through the collective bargaining process. It would involve an element of diversion of potential pay increases into retirement savings, and there would be matching contributions of new money from the government to state employers to leverage the employer contribution. We have had reports that it is difficult to design an attractive scheme in an unattractive tax environment, so we are grappling with sequencing: do we try and implement a new scheme and take advantage of better tax treatments when they are implemented, or hold off finalising design details until we know how we will link the scheme in with tax advantages?

We have not gone far enough to report back to a wider stakeholder reference group of state unions and employer representatives; but I can say that we are walking our talk, and stand ready to put money in to make sure that a scheme gets off the ground.

The point of my explaining this to you is to stress the importance of letting others know when employers take steps to reverse the decline in the coverage of work based superannuation. If we are to make progress, it has to become the normal and acceptable behaviour to be using the S word again.

Thank you.