Section 6 - Review NZS Today


6.         Framing a debate on every aspect of the NZS benefit design[30]

New Zealand should start a full review of NZS and the sooner that happens, the better.  This is a topic that has been canvassed on many occasions in the last 25 years:

  • The Task Force on Private Provision for Retirement (1991-92);
  • Periodic Report Group (1997);
  • Periodic Report Group (2003, report accessible here);
  • Retirement Commissioner’s Review of Retirement Income Policy (2007);
  • The 2010 Review of Retirement Income Policies (report accessible here);
  • The Review of Retirement Income Policies (2013, report accessible here) and, most recently,
  • The 2016 Review of Retirement Income Policies (report accessible here).

None of these seven reports has come close to what is now needed.

This section looks at each of 13 design decisions that go to make up NZS as we know that today.  Of those, the government chose to change two in its 6 March 2017 announcements – the state pension age from 2037 and the minimum residence requirement from, possibly, 2018.  Neither of those changes has been the subject of a research-led analysis[31].  The government also apparently thought about two other aspects of NZS’s design:

“Other settings such as indexing NZ Super to the average wage and universal entitlement without means testing will remain unchanged…” Stephen Joyce, Minister of Finance 6 March 2017[32]

Again, neither of those decisions was the subject of a research-led analysis or consultation.

This kind of ad hoc decision-making has not served New Zealand well over the more than 40 years since 1975 when the then-Labour government introduced the compulsory private savings scheme (the ‘New Zealand Superannuation Scheme’).  These latest decisions have a distinctly political flavour to them, given the upcoming election and the manoeuvring that is an inevitable feature of our MMP environment.

We think that, despite the March 2017 decisions, it is time to have a proper look at the benefit design of NZS.

Here is a summary of the major design decisions that should emerge from the proposed, research-led review[33]:

6.1  Universal or means-tested? Until 1977, the then ‘Old Age benefit’ was income-tested but ‘Universal Superannuation’, payable from age 65 was not.  In practice, by 1975, this meant there was an income-test between age 60 (when the Old Age benefit started) and age 65 (when Universal Superannuation started)[34].  ‘National Superannuation’ changed that in 1977. The state pension age was reduced to 60 and the income-test was eliminated. 

[30] This section is based on the RPRC’s 2015 PensionCommentary 2015-1 – Re-designing New Zealand Superannuation, Michael Littlewood (accessible here).
[31] And if the government uses the Retirement Commissioner’s 2016 review as supporting evidence, our statement stands.  That review did not use research-based findings either.
[32] Press release 6 March 2017 Lifting NZ Super age the right thing to do, accessible here.
[33] NZS is payable under the New Zealand Superannuation and Retirement Income Act 2001 accessible here.  In this section, we refer to that as the ‘Act’.
[34] Retirement income in New Zealand the historical context, David Preston (2001), Office of the Retirement Commissioner (now Commission for Financial Capability, Wellington) provides a good summary of the history of NZS since it started in 1898 (accessible here).


Page 2

However, the Labour government re-introduced income-testing without debate from 1985 (the ‘surcharge’).  It was watered down later and finally eliminated in 1998 by the next National government, again without discussion.

NZS itself has never been subject to an asset-test[35].  Finding out how the asset-test for the Age Pension works in Australia should be an important part of the New Zealand review.

Should NZS be paid to people who don’t, on any reasonable basis, need it?  If we decide to apply a means-test (on income and/or assets), NZS will no longer be a universal pension.  In that case, where should the reductions begin and at what rate should the state pension be withdrawn?  What might be the consequences for the economy, particularly to labour force participation rates, and how might New Zealanders react to such tests?  How might savers react during both their working and retirement periods?  What might the administrative costs be?  We have some experience from the days of the surcharge (1985-1998) and there are useful potential lessons to be learned from Australia that has extensive income- and asset-tests.

Governments cannot dictate how much of the economy’s total output goes to the old because of private, unmanageable responses to the retirement income framework.  We need a better understanding of the present and expected total claims on the economy by the old[36].  With the best of intentions, it is almost impossible to regulate private behaviour so as to achieve the desired overall objectives and any discussion of income- and/or asset-tests must recognise that[37].

6.2  State pension age:  The state pension age of 65 was first set in the 1898 Old-age Pensions Act (accessible here).  We flirted briefly with age 60 between 1977 and 1992 but, by 1 April 2001, it was back to age 65.  The government’s recent announcements would see that increasing from 65 to 67 between 2037 and 2040.

Why the current age 65?  Why the proposed age 67?  There is no particular reason (physiological, physical or gerontological) to pick any age because the appropriate age for an individual will be driven by many issues including health, availability of work, family circumstances, income, personal preferences and wealth.

And why does the under-age-65 partner of a person who is over age 65 collect the pension on a means-tested basis (the ‘non-qualifying partner option’)?  The Retirement Commissioner’s 2016 Review recommended abolishing this without offering any evidence to support that suggestion, other than that the option exists and how much it might save to remove it[38].  The government rejected that recommendation, in an evidence-free response[39].

[35] The Old Age benefit (pre-1977) was the last time New Zealand had an asset-test on an age-related income benefit.
[36] In Turning silver to gold: Policies for an ageing population (2014), Retirement Policy and Research Centre (accessible here), Claire Dale collates what we know about public policy-driven state costs of services and support for the age 65+ population through to 2030.  For the proposed national discussion on NZS, that analysis needs to be deepened and extended.
[37] In fact, it’s even possible that the total (public + private) claims of the old on the economy will be greater in the presence of a means-tested state pension than might be the case with a universal pension such as NZS.  Savers might under-estimate the net amount of state provision they will receive and therefore over-estimate the need for private provision in the presence of means-tests.  Means-tests are by their nature, complicated and expensive to administer.  That could all increase the total economic claims by older people (both public and private) on a demographically ageing economy.  Means-tests also have a tendency to provoke avoidance activities that aim to limit the impact of the asset- or income-test on net incomes in retirement.  Australia shows how the financial planning industry was founded, in part, on means-test avoidance.
[38] Retirement Commissioner’s 2016 Review of Retirement Income Policies (accessible here) at page 24.
[39] In the Minister of Commerce and Consumer Affairs’ letter of 7 June 2017 (accessible here), she said “The Government considers that retaining the non-qualifying partner option is necessary to provide low-income couples with the means to smoothly transition to retirement.  This is because the majority of non-qualifying partners are people approaching the age of eligibility for NZ Super.”  This, we suggest, easily qualifies as a statement of the obvious but is no justification for the government’s decision.


Page 3

Retirement, as a universal ‘entitlement’, is a relatively recent phenomenon.  In 1910, two out of three US men age 65 and over were actively employed.  Even at age 72, male participation in the labour market was over 50%[40].  The percentage of US men age 65 and over who worked fell to about 50% in 1950 and then below 20% in 1980.  By 1990, it had fallen to 16%[41] and has now (2014) climbed back to 17.4%[42], probably because the state pension age for US Social Security is now 66 and is increasing to 67 by 2027.  Of all US citizens over the age of 65, 19% now work[43].

New Zealand’s participation rate for all those aged 65+ fell to as low as 6.4% in 1986.  It’s now (2013) 22.5% and rising[44].

When the government chooses a state pension age, it must balance social issues, labour market efficiencies, voter satisfaction and fiscal considerations[45].  Some suggest that, with improving mortality, we should be seeing a natural increase in the state pension age, certainly by comparison with the position that prevailed in 1898[46].  The state pension age is now, perhaps, one of the most significant single elements of public welfare policy, one that has, in essence, persisted for 119 years.  But what is the impact on other programmes such as ACC weekly compensation?

The amount and quality of the information we have on issues associated with the fixing of a state pension age are inadequate.  We need to discuss the distortions created by the present state pension age on the work/retirement decision.  We do not know fundamental facts such as when New Zealanders stop working (not when they ‘retire’), when they can afford to stop working or when they would prefer to stop and finally what the progress is from fulltime work to ‘fulltime’ retirement.  There is more on this in section 16 (When do New Zealanders retire?).

[40] Retirement Trends and Policies to Encourage Work Among Older Americans, Gary Burtless and Joseph Quinn (2000), The Brookings Institution (accessible here).
[41] Monthly Current Population Survey data in the US cited in Passing the Torch: The Influence of Economic Incentives on Work and Retirement, Joseph Quinn, Richard Burkhauser and Daniel Myers (1990), W E Upjohn Institute for Employment Research (accessible here).
[42] ILO Key Indicators of the Labour Market (KILM, 7th Edition, 2011) cited in Comparison of New Zealand and Australian Retirement Income Systems, Ross Guest (2013) accessible here.
[43] US Labor Department data, Press report, 5 May 2017 here.  Any US data about labour force participation rates after age 65 now need qualification because of the increasing state pension age.
[44] See Retirement Policy and Research Centre (2014) Updating data on older workers, PensionBriefing 2014-4 (accessible here).
[45] Other countries’ decisions about their state pension ages are not relevant to New Zealand’s own decision.  The government claimed support for lifting the age to 67 (by 2040) from the experience of other countries (see the Annex to the undated Cabinet paper on New Zealand Superannuation from Minister of Finance of March 2017 at p. 14, (accessible here).  However, each country’s social and workforce positions are different to others and each must make its own decision based on a full analysis of relevant data.  That hasn’t happened so far in New Zealand.
[46] In 1898, New Zealand males had an average life expectancy of 12.8 years at age 65; females had an average of 16.9 years – Life Expectancy (ex) by age, sex, and birth cohorts: 1876-1941, Statistics New Zealand, accessible here.  Currently males have an 18.9-year life expectancy at age 65; females will survive, on average for 21.3 years (New Zealand Period Life Tables 2012-14, Statistics New Zealand accessible here).  Over the 119 years, males at age 65 have added an average 6.1 years; females 4.4 years.


Page 4

6.3  Residency test:  We pay NZS from age 65 to anyone who has lived in New Zealand for ten years after age 20 with five of those being after age 50.  The government’s recent announcements will see that increase to 20 years, with full effect from 2037 (after some as-yet-unclear transitional provisions for current residents and returning New Zealanders).  Why 20 years?  Why 10 years and why ages 20 and 50?[47].

6.4  How much?  The size of the pension has had a more varied history.  The age pension was modest (and both income- and asset-tested[48]) when it started in 1898.  By 1940, the single person’s pension was about 29% of the then national average wage (on a ‘gross to gross’ basis).  Over the following 35 years to 1975, it fluctuated between, roughly, 27%-35%.  The introduction of ‘National Superannuation’ in 1977 saw a major lift but, in spite of the highly politicised nature of the issue since then, has fluctuated over the last 30 years between about 40-47% of the national average wage (which was $58,745 before tax as of December 2016[49]).  Currently, NZS is 40% on a pre-tax basis[50] for a single person living alone (43% on a net-to-net basis).

In 1989, the then Labour government decided that the after-tax married couple’s rate should lie between 65% and 72.5% of the after-tax national average wage[51].  Currently, it is a net 66%.  There was no public debate at the time about this and no ‘science’ to it other than that it was less than the then-current rate and was expected to save a significant amount.

Is 65% enough or too much?  One measure of adequacy might be to eliminate poverty in old age; another might be to ensure ‘participation and belonging’.  New Zealand needs to decide what the welfare objective of NZS should be, how to test that and how to measure changes over time to ensure the objective is reached[52].

6.5  How re-valued?  Until ‘National Superannuation’ of 1977, there had been no formal link between the pension and any measure of real value.  National Superannuation made that link with the national average wage in 1977.  As a country, we have never discussed whether the pension should be linked to anything in particular though some reviews have recommended that it be changed[53].  Currently the measure is a combination of the after-tax, national average, ordinary-time wage and the Consumer Price Index.  Is that the best?  Some say it should instead be linked to economic output; others to prices alone; yet others to a mix of prices and incomes.  Some say that the CPI does not fairly reflect the prices faced by pensioners and that NZS should be measured against a ‘superannuitants’ price index’[54].  We have never had a research-led discussion about the alternatives and their implications.

[47] Before 1937, the qualifying period was 25 years - the Pensions Amendment Act 1937 (accessible here) reduced that to 10 years.
[48] The means-tests were administered by the Magistrate’s Court and the weekly amount decided was published in the local newspaper: see Part of our pensions past: the 1898 Old Age Pension (an RPRC PensionBriefing of 2012, accessible here).
[49] The December 2016 national average wage was used to set the 1 April 2017 NZS rates.
[50] The pre-tax, annual NZS for a single person, living alone is $23,405 a year; $450.10 a week (2017) – see here.
[51] This is now in section 16(1)(a) of the New Zealand Superannuation and Retirement Income Act 2001 (the Act), accessible here.
[52] Given that KiwiSaver is expected to cost taxpayers $738 million for the year ended 30 June 2017 (The Treasury, Long Term Fiscal Model 2016), the expected benefits members will receive at the pension age could form part of the discussion about the size of NZS itself.  As KiwiSaver is voluntary, that discussion will become complicated.
[53] The Retirement Commissioner’s 2013 Review of Retirement Income Policies (accessible here) recommended that the Treasury develop a model “…that will show the likely impacts on living standards among older New Zealanders of a new method of indexation of NZS, based on the average of percentage change in consumer prices and earnings but no less than price inflation in any year.” (at page 48).  Adopting that would mean a long-term reduction in the real value of NZS.  There was no sign of that suggestion or of any results in the 2016 report.
[54] Statistics New Zealand has started publishing price movements as they affect 13 different household groups, including superannuitants – see Household living-costs price indexes: Background, 2016 accessible here.  This analyses the impact that price rises have on the different groups, based on their spending patterns.  Applying these retrospectively, they show that superannuitants as a group had inflation of about 0.7% p.a. more than the ‘all households’ group over the period June 2008 to September 2015.

Page 5

6.6  How paid for?  Until 2001, New Zealand paid for NZS on a pay-as-you-go (PAYGO) basis.  There was a ‘Social Security Fund’ between 1938 and 1964 but that was little more than a bookkeeping arrangement.  In 2001, the government decided, without debate, that New Zealand needed to partially pre-fund the expected cost of NZS through contributing to the New Zealand Superannuation Fund (NZSF) that would invest in capital markets. This means that NZS will still be largely PAYGO, but a little bit pre-funded[55].

We did not have a research-led debate when the NZSF started in 2001.  Some think the role of the NZSF should be significantly extended[56]; others that the NZSF be dismantled and the proceeds used to reduce government debt.  Regardless, New Zealand needs to understand the economic and political considerations of the alternatives.  The next section 7 (The place of the New Zealand Superannuation Fund) looks at this.

6.7  Payments to single people: Why is a single person, living alone, entitled to at least 65% of the married couple’s combined rate (section 16(1)(b) of the Act accessible here)?  Why is a single person living with others entitled to 60% of the married couple’s combined rate (section 16(1)(c) of the Act)?  Why do we pay a married couple less in total than two single people who live together?  Are these amounts adequate (or too much)?  When was the empirical work done to see whether the proportions might be other than they are?

6.8  Overseas pensions: The present regime for deducting equivalent overseas pensions from a resident’s entitlement to NZS under section 70 of the Social Security Act 1964 is a confused, inconsistent mess.  Many commentators have suggested that the treatment of overseas pensions needs an urgent review and that should happen anyway.  However, it should preferably be part of the proposed research-led debate on NZS as a whole.  Section 8 (Overseas pensions and section 70 deductions) below looks at this issue in more detail.

6.9  ACC entitlements: given that NZS is not income-tested, why can’t an ACC recipient receive both NZS and an ACC pension (section 7(2) of the Act, accessible here)?  The present law may be correct but the question needs analysis and an answer.

6.10  Periods of absence – VSA or missionary service: Sections 9 and 10 of the Act include periods of absence on Volunteer Service Abroad (here) and missionary service (here) for the residency qualification in section 8 (see paragraph 6.3 above).  Why not include other charitable bodies such as the Red Cross and other “recognised aid agencies” as in section 24 (here)?  Why include any of these periods of absence?

6.11  No reduction to benefits: Section 15(4) of the Act (here) says that no future CPI adjustments to the amounts payable shall result in the benefits being reduced.  Why?  If we go through a sustained period of falling incomes, why should the incomes of retired New Zealanders increase in relative terms (by standing still nominally) while all other New Zealanders are forced to adjust to new, lower standards of living?  If 65% of the average wage is the right answer on the married couple’s pension, why does it become wrong if average incomes were to fall?

[55] The Treasury estimates that, by 2060, capital withdrawals from the NZSF will be only 6.5% of the net annual outlay on NZS in 2060 (The Treasury’s 2016 NZSF Contribution Rate Model, accessible here).
[56] See, for example, To Save or Save Not: Intergenerational Neutrality and the Expansion of New Zealand Superannuation, Andrew Coleman (2014) accessible here.


Page 6

6.12  Hospital rates: Section 19 of the Act (here) says that the amount of NZS should reduce after 13 weeks in a public hospital to, currently, a net $48.49 a week.  Why 13 weeks?  Why not four weeks?  What is the logic of $48.49?

6.13 Payments overseas: A number of aspects of overseas payments of NZS deserve debate:

(a) Why should NZS be payable to anyone who is overseas for up to 26 weeks (section 22 of the Act here)?  Why not 13 weeks (as with hospital rates)?  Why not four weeks?

(b) Why should the 26 weeks in section 22 (referred to in the last paragraph) become 156 weeks (three years) if the recipient is working (albeit on an unpaid basis) with a “recognised aid agency” (section 24 here)?[57]

(c) Why should someone who lives in a country with no “reciprocity of social security monetary benefits” receive a proportion of NZS (sections 26 here and 26A here of the Act)?  Why do they get anything?  What welfare obligation do New Zealand taxpayers have towards people who are no longer resident?  If we do that, should that proportion be based on years of residence between ages 20 and 65 (section 26A(1)) and why might years after age 65 be excluded?  If this is appropriate for emigrants, might that test also be appropriate for immigrants with overseas pensions (see paragraph 6.8 above)?  Why must the person be ordinarily resident on the application date (section 26B(b)(i)) here but not before or after?  Finally, why do we pay that pension gross?  Why not deduct tax?

(d) Why should NZS be payable to people who leave New Zealand and live in a “specified Pacific country” (section 31 of the Act here)?  Why is each of the countries listed in Schedule 2 of the Act here included[58]?  Why does each person so affected need to have lived in New Zealand for 20 years (section 32(1) here)?  Or to receive a proportionately reduced benefit if they have lived in New Zealand for at least 10 years?

The final important step in this process is agreeing transition provisions that move NZS from its present basis to the new 21st Century programme.  The benefit design decisions will affect different groups in different ways so the transition will need tailoring to individual needs.  Each of the issues described in this section raises benefit design implications that should at least be debated in the context of good evidence.  The debate must acknowledge that today’s decisions will be made under conditions of great uncertainty so flexibility will be an important component of those decisions.

[57] And the extension from 26 to 156 weeks applies only if the Ministry’s Chief Executive is satisfied that the applicant “has not deprived another person of paid employment to engage the person to do that work on an unpaid basis”.  It will be interesting to know how the Chief Executive might arrive at such a conclusion and, indeed, whether the test in section 26(1)(c)(ii) has ever been applied.  If it hasn’t, this particular test should go.
[56] The countries included in the list of “specified Pacific countries” (Schedule 2 of the Act) are a curious mixture.  Some have long-standing relationships with New Zealand (Cook Islands, Nauru, Niue, Samoa, Tokelau and Tonga).  For these, the payment of NZS might be considered as part of our aid programmes.  However, why does the list extend to American Samoa, French Polynesia, Guam, Marshall Islands, Palau, New Caledonia, Northern Mariana Islands, Pitcairn Island and Wallace and Futuna?  All of these are either colonies of or have political ties to other countries (US, UK and France).  The case is even more curious for the US and French colonies given that New Zealand does not have social security reciprocity agreements with either country.  There is no obvious reason to include the other seven countries (Federated States of Micronesia, Fiji, Kiribati, Papua New Guinea, Solomon Islands, Tuvalu and Vanuatu) other than they are all in the Pacific.


Page 7

New Zealand needs to agree social policy reasons as to why the things described in this section should be so and what might be ‘better’ ways of achieving common goals.  New Zealand has never had such a debate[59].

Each of the benefit design elements should be agreed without, at least initially, regard for the expected cost to taxpayers.  What, in each case, is the ‘best’ answer to each benefit design question posed above?

Costings of a 21st Century NZS should be done only once all aspects of the benefit design have been tentatively settled.  It is quite likely that the agreed scheme and the transition arrangements will cost more than might be acceptable to today’s taxpayers.  The debate on benefit design should then pull back from the ‘ideal’ so as to bring the new NZS within an acceptable budget.  That will probably be an iterative process.

Eventually, we will arrive at a benefit design that achieves the country’s agreed objectives at a cost that is likely to be acceptable now and over coming years.

A word about the Accommodation Supplement: For the reasons described in section 12 below (Housing and home-ownership), we do not know what proportion of New Zealanders over age 65 own the home they are living in.  Probably about one-fifth do not.  Having a paid-off home to live in is an important part of financial preparations for retirement.  The state supports those who do not own, or who do own their home but are having difficulty meeting outgoings, by providing an ‘Accommodation Supplement for those not living in ‘social housing’.  Currently that is up to $130 a week for a superannuitant (depending on where the recipient lives) and is income- and asset-tested.  Higher amounts are possible if there are dependent children (see here for more)[60].

We flag this as an issue that is related to the review of the state’s involvement in financial support for older New Zealanders but do not have the knowledge or experience to analyse its implications[61].  However, as with NZS itself, we need more, better data about those who retire and who do not own a debt-free home.

The Retirement Commissioner’s 2016 Review recommended an increase in the Accommodation Supplement (here at page 21) but in an evidence-free kind of way.  There was no discussion or basis for her recommendation, other than “The maximum amount of accommodation supplements were last reviewed in 2005.”  The 2017 Budget has in fact increased them but we don’t know the basis, nor whether the rates from 2018 will be ‘enough’ but they will be more.

[59] Overall, New Zealand is doing these things generally ‘well’.  That is reflected in ‘material deprivation rates’ that, for over-65s are amongst the lowest of all social groups within New Zealand.  They also compare very favourably in comparisons over over-65s compared with 27 EU and two other European countries – see The material wellbeing of New Zealand households: trends and relativities using non-income measures, with international comparisons, Bryan Perry (Ministry of Social Development) 2016, accessible here, at page 19.  As explained in this section, our current overall performance is no justification for allowing things to continue as they are.
[60] Rates and the geographic basis for their calculation will change from 1 April 2018 – see here (Budget 2017 announcements) for more.
[61] The total cost of the Accommodation Supplement for the year ended 30 June 2017, at all ages, is expected to be $1.13 billion.  The changes announced in the 2017 Budget are expected to increase that to about $1.5 billion from 2018 on – see the Budget Economic and Fiscal Update 2017 (accessible here).  We could not find out how much of that related to ‘retired’ people.


Page 8

Questions New Zealand needs to discuss on the review of NZS’s design:

Based on the analysis in this section, here are the questions New Zealand needs to discuss on the design of NZS:

  1. Should NZS be a universal pension (as now) or means-tested (assets, income or both)?  How administratively do the income and asset tests work in Australia and how do Australians respond to those?
  2. Should the state pension age be 67 (as currently proposed) or later?  What about earlier or, as some suggest, ‘flexible’?
  3. Should the minimum residency period be 10 years, 20 years or more?
  4. How much should NZS be for a couple?
  5. How should NZS be re-valued each year?  Why is there a floor?
  6. Should NZS be pre-funded in full, partially (as now) or paid on a ‘pay-as-you-go’ basis (as was the case until 2001)?  The discussion in the next section 7 (The place of the New Zealand Superannuation Fund) is relevant here.
  7. Should single pensioners get more than each of a couple?  How should the single person’s rate be set?
  8. Which overseas pensions should be deducted under section 70?  There is more on this in section 8 below (Overseas’ pensions and section 70 deductions).
  9. Should ACC recipients lose their NZS?
  10. Should there be ‘approved’ absences overseas for the residence test?
  11. How much should the ‘hospital rate’ be?
  12. Should overseas residents be entitled to any NZS?  Should that be tax-free?
  13. How should all this be reviewed?  We think the present three-yearly review isn’t working – there is more on this in section 21 below (The review process).
  14. Then we need to agree the transition between current and future benefits (if changed).

Discussion of any reform should begin with agreement on the principles before the detail of any reform is examined.