KiwiSaver evaluation: Final summary report A joint agency evaluation 2007–2014 Prepared for: KiwiSaver Evaluation Steering Group Prepared by: National Research and Evaluation Unit February 2015 KiwiSaver evaluation: Final summary report KiwiSaver evaluation: Final summary report Disclaimer Acknowledgements The views, opinions, findings, and conclusions expressed in this report do not necessarily reflect the views of Inland Revenue. This huge undertaking wouldn’t have been completed without the immense contributions from the following: 48T Inland Revenue has made every effort to ensure that the information contained in this report is reliable, but does not guarantee its accuracy or completeness and does not accept any liability for any errors.  Lorrae Ward Evaluator/Director, Cyperus Ltd, for her work synthesizing the findings from the reports over the seven-year life of the KiwiSaver evaluation program  the current and previous members of the multiagency KiwiSaver evaluation steering group from: The information and opinions contained in this report are not intended to be used as a basis for commercial decisions and Inland Revenue accepts no liability for any decisions made in reliance on them. The material contained in this report is subject to Crown copyright protection unless otherwise indicated. The Crown copyright protected material may be reproduced free of charge provided it is reproduced accurately and not used in a derogatory manner or in a misleading context. Where the material is being published or issued to others, the source and copyright status should be acknowledged. The permission to reproduce Crown copyright protected material does not extend to any material in this report that is identified as being the copyright of a third party. Authorisation to reproduce such material should be obtained from the copyright holders. © Crown copyright 2012 ISBN: 978-0-478-41828-6 - the Ministry of Business, Innovation, and Employment - The Treasury - the Commission for Financial Capability - the Financial Markets Authority - the Ministry of Social Development - Victoria University of Wellington - Statistics New Zealand who through their expertise provided guidance that helped shape the report into its current version  research companies and contractors who have worked on various research and evaluation projects under the evaluation program  Inland Revenue analysts, researchers, technical experts and evaluation programme managers, past and present, who provided a steer and/or worked on the evaluation. To all, a big thank you. National Research and Evaluation Unit Te Wāhanga ā-motu mo te Rangahau me Aromātai Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report Table of Contents Section 1. Executive summary .................................................................................................................................1 1.1 Introduction ...........................................................................................................................................................1 1.2 Method ..................................................................................................................................................................1 1.3 Report structure ....................................................................................................................................................1 1.4 Key findings ...........................................................................................................................................................2 1.4.1 Implementation and delivery .................................................................................................................................... 2 1.4.2 Scale and patterns of uptake ..................................................................................................................................... 2 1.4.3 The efficacy of the key features of KiwiSaver ............................................................................................................ 2 1.4.4 Savings habits and asset accumulation...................................................................................................................... 3 1.4.5 Superannuation markets and the financial sector ..................................................................................................... 3 1.4.6 Value for money ........................................................................................................................................................ 3 1.5 Conclusions ............................................................................................................................................................4 1.6 Limitations .............................................................................................................................................................5 Section 2. KiwiSaver: A brief introduction ...............................................................................................................6 2.1 The purpose of KiwiSaver ......................................................................................................................................6 2.2 The mechanics of KiwiSaver ...................................................................................................................................6 Section 3. The KiwiSaver evaluation........................................................................................................................7 3.1 The evaluation strategy: Purpose, focus and scope................................................................................................7 3.2 Evaluation objectives .............................................................................................................................................7 3.3 Evaluation design ...................................................................................................................................................7 3.4 Evaluation method ................................................................................................................................................8 3.5 Limitations to the evaluation .................................................................................................................................8 Section 4. The implementation and delivery of KiwiSaver.......................................................................................9 4.1 Introduction ...........................................................................................................................................................9 4.2 Employers ..............................................................................................................................................................9 4.2.1 Support for KiwiSaver ................................................................................................................................................ 9 4.2.2 Impact on workload ................................................................................................................................................... 9 4.2.3 Inland Revenue communications .............................................................................................................................. 9 4.2.4 Awareness of obligations ......................................................................................................................................... 10 4.2.5 Compliance costs ..................................................................................................................................................... 10 4.2.6 Remuneration approaches ...................................................................................................................................... 11 4.2.7 Employer contributions ........................................................................................................................................... 11 4.2.8 Existing superannuation schemes ........................................................................................................................... 11 4.2.9 Informing employee decisions ................................................................................................................................. 11 Prepared by: National Research and Evaluation Unit 4.3 Providers ............................................................................................................................................................. 11 4.3.1 KiwiSaver schemes ................................................................................................................................................... 11 4.3.2 Engagement with Inland Revenue ........................................................................................................................... 12 4.3.3 Communication with their members ....................................................................................................................... 12 4.3.4 Processes for withdrawal of funds........................................................................................................................... 12 4.4 Employees (individuals) ....................................................................................................................................... 13 4.4.1 The efficacy of the design features in promoting membership ............................................................................... 13 4.4.2 Learning about KiwiSaver features .......................................................................................................................... 13 4.4.3 Making decisions about their investments .............................................................................................................. 13 4.4.3.1 Choosing a provider .............................................................................................................................................. 13 4.4.3.2 Choosing an investment fund ............................................................................................................................... 14 4.4.4 The withdrawal experience ..................................................................................................................................... 14 4.5 Summary of key points ........................................................................................................................................ 14 Section 5. KiwiSaver usage: Scale and patterns of uptake ..................................................................................... 16 5.1 Introduction ......................................................................................................................................................... 16 5.2 KiwiSaver membership ........................................................................................................................................ 17 5.3 Enrolment trends ................................................................................................................................................. 17 5.4 Individual member demographics ........................................................................................................................ 18 5.5 Behavioural characteristics of members .............................................................................................................. 19 5.5.1 Risk taking ................................................................................................................................................................ 19 5.5.2 Knowledge of KiwiSaver features ............................................................................................................................ 19 5.5.3 Engagement with KiwiSaver .................................................................................................................................... 20 5.5.4 Attitudes of members towards savings ................................................................................................................... 21 5.6 Summary of key points ........................................................................................................................................ 21 Section 6. The features of Kiwi Saver .................................................................................................................... 23 6.1 Introduction ......................................................................................................................................................... 23 6.2 Making it easy to join........................................................................................................................................... 23 6.2.1 Kiwi-Saver as a work-based scheme ........................................................................................................................ 23 6.2.2 Automatic enrolment .............................................................................................................................................. 24 6.2.3 Choosing a KiwiSaver scheme .................................................................................................................................. 24 6.2.4 Changing schemes (providers) ................................................................................................................................. 25 6.2.5 Default contribution rates ....................................................................................................................................... 26 6.2.6 Contribution amounts ............................................................................................................................................. 27 6.3 Locked-in features ............................................................................................................................................... 28 6.3.1 Optingout of KiwiSaver ............................................................................................................................................ 28 6.3.2 Locked-in savings ..................................................................................................................................................... 29 6.3.3 Withdrawal for retirement ...................................................................................................................................... 30 6.3.4 The funds available to early retirees ....................................................................................................................... 30 6.3.5 The withdrawal behaviour of the early retirees ...................................................................................................... 31 6.3.6 Withdrawal for reasons outside retirement ............................................................................................................ 32 6.4 Making KiwiSaver attractive ................................................................................................................................ 33 6.4.1 Employer contributions ........................................................................................................................................... 33 6.4.2 Crown contributions ................................................................................................................................................ 33 Prepared by: National Research and Evaluation Unit 6.4.3 Contributions holidays ............................................................................................................................................. 33 6.4.4 Purchasing a home .................................................................................................................................................. 34 6.4.4.1 Intentions to use the features .............................................................................................................................. 35 6.4.4.2 Uptake of the features .......................................................................................................................................... 35 6.5 The overall efficacy of KiwiSaver features ............................................................................................................ 36 6.6 Summary of key points ........................................................................................................................................ 36 Section 7. Individuals’ savings habits and asset accumulation .............................................................................. 38 7.1 Introduction ......................................................................................................................................................... 38 7.2 Additional savings and retirement savings substitution ....................................................................................... 38 7.2.1 Additionality in 2010 ................................................................................................................................................ 38 7.2.2 Additionality in 2014 compared with 2010 ............................................................................................................. 39 7.3 Attitudes and behaviours towards retirement ..................................................................................................... 41 7.3.1 Financial planning for retirement ............................................................................................................................ 41 7.3.2 Expected standard of living in retirement ............................................................................................................... 41 7.3.3 Expected length of retirement ................................................................................................................................. 41 7.4 Income in retirement ........................................................................................................................................... 41 7.4.1 Sources of income.................................................................................................................................................... 41 7.4.2 The perceived adequacy of retirement income....................................................................................................... 43 7.5 The accumulation of net wealth .......................................................................................................................... 44 7.6 Using KiwiSaver in retirement.............................................................................................................................. 44 7.7 Summary of key points ........................................................................................................................................ 46 Section 8. Superannuation markets and the financial sector ................................................................................. 47 8.1 Introduction ......................................................................................................................................................... 47 8.2 KiwiSaver scheme fund balances ......................................................................................................................... 47 8.3 Supply side impacts ............................................................................................................................................. 47 8.4 Summary of key points ........................................................................................................................................ 48 Section 9. Value for money of KiwiSaver ............................................................................................................... 49 9.1 Introduction ......................................................................................................................................................... 49 9.2 Considering value for money ............................................................................................................................... 49 9.3 The costs to the Crown ........................................................................................................................................ 49 9.4 Effectiveness of KiwiSaver in meeting the target market ..................................................................................... 49 9.5 Cumulative costs and benefits ............................................................................................................................. 50 9.6 Summary ............................................................................................................................................................. 51 Section 10. Conclusions ...................................................................................................................................... 52 Section 11. The KiwiSaver evaluation programme .............................................................................................. 53 11.1 The effectiveness of the evaluation ................................................................................................................... 53 11.1.1 Evaluating against policy objectives ...................................................................................................................... 53 11.1.2 Survey of Steering Group members....................................................................................................................... 53 11.2 The future of the evaluation programme ........................................................................................................... 53 11.3 Research ideas from the evaluation ................................................................................................................... 54 11.4 Discussion / recommendations .......................................................................................................................... 54 Appendix 1: KiwiSaver features and glossary............................................................................................................. 56 Appendix 2: Research included in this synthesis ........................................................................................................ 58 Prepared by: National Research and Evaluation Unit Appendix 3: Changes to KiwiSaver features over time ............................................................................................... 60 Appendix 4: Membership profiles from linked data set as at September 2010 .......................................................... 61 Appendix 5: Demographic profiles by membership status and enrolment type ......................................................... 63 Appendix 6: Change of scheme demographic characteristics at September 2010 ...................................................... 66 Appendix 7: Mean and median for cumulative member contribution value ($) from October 2007 to September 2010 .......................................................................................................................................................................... 68 Appendix 8: Demographic characteristics of those on a contributions holiday .......................................................... 70 Appendix 9: Sources of income of respondents to 2010 survey (percentages) ........................................................... 72 Appendix 10: Survey of Steering Committee ............................................................................................................. 73 A10.1 Survey respondents ......................................................................................................................................... 73 A10.1.1 The effectiveness of the evaluation .................................................................................................................... 73 A10.2 Future evaluation activities ............................................................................................................................. 76 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report Section 1. Executive summary undertaken for the purposes of this report. Readers are referred to these reports should they require more detail about the information presented here. 1.1 Introduction This is the final report from the evaluation of KiwiSaver, undertaken between July 2007 and June 2014, a period of seven years. Over that time a substantial number of research and monitoring activities were undertaken resulting in a rich picture of KiwiSaver from its initial implementation through to early outcomes. The evaluation was a joint agency evaluation and, as such, it collected evidence from a range of perspectives. This material has been collated under each of the five objectives of the KiwiSaver evaluation: The primary legislative objectives of KiwiSaver are to:    To monitor KiwiSaver usage to understand the scale and pattern of uptake (Section 5)  To assess whether the key features of KiwiSaver are generating expected outcomes (Section 6)  increase individuals’ wellbeing and financial independence, particularly in retirement, and to provide retirement benefits. To determine the impact of KiwiSaver on individuals’ saving habits and asset accumulation (Section 7)  To determine the impact of KiwiSaver on superannuation markets and the financial sector (Section 8). In addition, members of the steering group were interviewed about the efficacy of the evaluation. This survey is reported in Appendix 10. 1.3 Report structure The report begins with a brief description of KiwiSaver and a summary of the design and method of the evaluation. The purpose of this report is two-fold:  To assess the implementation and delivery of KiwiSaver to inform on-going development and services delivery (Section 4) encourage a long-term savings habit and asset accumulation by individuals who are not in a position to enjoy standards of retirement similar to those in pre-retirement This report brings the comprehensive information collected through this evaluation together in one place. Previously, this information has been reported in regular monitoring reports to the Minister of Finance and in stand-alone research reports. These reports are published on Inland Revenue’s website. The report also offers recommendations for any further research and/or evaluation of KiwiSaver.   to provide evidence of the effectiveness of KiwiSaver, when considered against the objectives of the policy Findings related to each of the objectives of the evaluation are then collated in Sections 4 to 8. Each section has a brief summary at the end. to consider the effectiveness of the KiwiSaver evaluation and make recommendations for possible future evaluation and/or research into KiwiSaver. Note these recommendations do not assume further work will be undertaken. The final three sections of the report consider:  the value for money of KiwiSaver, including considerations of additionality, the extent to which the estimated target market has been reached, the impact on the accumulation of net wealth and the cost to the Crown of every dollar saved (Section 9)  the conclusions section brings together the key points from each of the previous sections of the report (Section 10)  the effectiveness of the evaluation and recommendations for future research and 1.2 Method This report was primarily developed through a desktop exercise, which involved reading all research and monitoring reports from across the seven years of the KiwiSaver evaluation. No additional analysis of any data collected during the KiwiSaver evaluation was 1 Prepared by: National Research and Evaluation Unit evaluation activities are discussed in the final section (Section 11). salaries it is likely these are individuals who have not changed jobs or who initially opted-out and then decided to join. The appendices contain a number of large data tables referred to in the report and copied from the relevant research/evaluation report. 1.4 Key findings  It appears that KiwiSaver knowledge is associated with income and the accumulation of assets. The more assets individuals have, the higher their level of knowledge about KiwiSaver.  Of respondents to the 2010 survey 21% were highly engaged with KiwiSaver, ie, they will have made informed decisions about whether to join, their provider and their investment type. There was a very strong correlation between levels of engagement and knowledge. 1.4.1 Implementation and delivery The following is a summary of findings reported in Section 4.0 of the report about the implementation and delivery of KiwiSaver as experienced by employers, providers and employees.   Employers reported the impact on their workload had been minimal. The average compliance cost for small and medium enterprises (SMEs) in 2009 was $770. This decreased to $661 in 2013.  Few employers (18%) reported changing their remuneration practices as a result of KiwiSaver. Those who did applied a “salary sacrifice”. It is important to note that KiwiSaver can have complicated effects on wage rates which may not have been captured in this research.   1.4.3 The efficacy of the key features of KiwiSaver Both employers and providers interviewed/surveyed during the early years of KiwiSaver were generally supportive of KiwiSaver. They believed the implementation had been effectively managed. In particular, the engagement models used were seen as highly successful. KiwiSaver was designed to make:  KiwiSaver attractive to individuals  it easy for people to join  it hard to get savings out. It was aimed at encouraging members to firstly join and secondly to save. The data in this report suggests these incentives have been more successful at getting people to join KiwiSaver than promoting additional savings. Key findings related to the efficacy of the features as reported in Section 6 include the following: The number of providers dropped from 54 in 2009 to 45 in 2013. Of these, four had assets of over $3 billion. While it was reported that many employees found KiwiSaver an easy way to save money there were also difficulties. These were the 4% contribution rate in 2008, a concern that future governments might change KiwiSaver and the apparent inflexibility of the scheme.  KiwiSaver is providing an important platform for employees to save. The deductions of up to 56% of KiwiSaver members are being made at source from salary or wages.  The proportion of members opting in remains high compared with those auto-enrolled, reflecting active membership. Further, there is evidence to suggest increased engagement with KiwiSaver accounts when choosing providers and schemes.  In 2013, 58% of members who had KiwiSaver deductions taken from their salary and wages were contributing the 3% default rate. This reflects the importance of setting the default rate at an appropriate level.  There is a considerable group, 38% of all members in 2013, not making any contributions. This is larger than can be accounted for by contributions holidays; particularly when one considers that many on holidays do make voluntary contributions.  It has become less common for people to opt out. In 2013, for the first time, the number who had opted out had decreased after four years of rises. It should be noted that the percentage change had been decreasing over that time.  For those earning solely wages or salary the median contribution in 2013 was $859; for those earning 1.4.2 Scale and patterns of uptake The following is a high-level summary of key points from Section 5 of the report.  By June 2013, the cumulative KiwiSaver membership had reached 2.15 million, equating to 53% of the eligible population. This was a net increase of over 1.4 million from the 716,637 members in June 2008.  Although membership continues to grow, the yearon-year increase in net membership has been steadily declining, with only a 9% increase in 2013 compared with 54% in 2009.  The proportion of members opting in has remained consistent over time (between 61% and 63%). Given the high proportion of members on wages or 2 Prepared by: National Research and Evaluation Unit only non-salary or wage income the median was $1,040.  Administrative data shows that the number of members on long holidays (five years) continues to grow while the number on shorter holidays has decreased. There also appears to be a decrease in the number of members commencing a holiday.  The cost of the home ownership features is likely to increase as more members begin to access them. The proportion of residential sales drawing on First Home Deposit Subsidy support increased from 0.1% in July 2010 to 5.3% in December 2012 as more members became eligible to access the subsidy.  The main reason members gave for joining in the 2010 survey was that it was an easy way to save. The second reason was the contributions from the Crown and their employer. There is likely to be an interaction effect between these two reasons with the contributions also making it easy to save more.  Analysis of the 2010 survey showed that 78% of respondents had an expected excess retirement income for meeting needs; 50% had an excess for being comfortable. There was no evidence that KiwiSaver membership was associated with any expected shortfall in retirement.  In a 2013 survey of early retirees withdrawing funds from KiwiSaver, 73% believed that their retirement income was either adequate or more than adequate. Only 16% did not have other savings and investments. For 42% of respondents these exceeded $100,000. The more savings they had in KiwiSaver, the more they were likely to also have other savings and investments.  The analysis of the linked KiwiSaver and Statistics New Zealand’s Survey of Family, Income and Employment (SoFIE) data set, for asset 2 accumulation found that KiwiSaver had not been successful in improving the accumulation of net wealth of its members. There was no evidence of a positive effect on net wealth accumulation from KiwiSaver for any subgroup when classified by gender, homeownership or income. The only group for whom there was a large positive effect was the one for those aged between 25 and 34. However, an independent review suggested some caution needed to be taken in considering these findings. 1F 1.4.4 Savings habits and asset accumulation Key findings related to the savings habits and asset accumulation of New Zealanders are summarised here. These are a snapshot rather than a detailed discussion of the long-term impact of KiwiSaver.      In 2010, using a flow measure of savings, the estimated level of “additionality” (ie, KiwiSaver contributions that were additional savings rather than substituting from other forms of saving) was 36%. This has not significantly changed in 2013. 1.4.5 Superannuation markets and the financial sector There has not been a lot of work done in this area through the evaluation. The main study was undertaken in 2010. In 2013, KiwiSaver contributions appear to have increasingly been made at the expense of paying off mortgage or other debt. Higher levels of knowledge regarding KiwiSaver and more planning around retirement income were associated with higher incomes and higher net worth. This relates to findings in other sections, which suggest engagement is related to the value of savings and assets. Of all respondents to the 2010 survey 53% thought their standard of living in retirement would stay the same, 27% expected it to decrease and 18% expected it to increase. KiwiSaver members were significantly more likely to report their standard of living would decrease than non-members.  The 2010 study concluded that KiwiSaver was providing a stimulus to the New Zealand financial sector and that it was a medium to long-term growth opportunity for the sector.  Since then more recent analysis has shown that a significant proportion of KiwiSaver funds are invested overseas and relatively conservatively. The impact on the capital markets remains small.  As of March 2013, scheme providers held $16.6 billion in KiwiSaver schemes; an estimated 19% of the superannuation and managed funds market. 3 2F 1.4.6 Value for money Value for money can be difficult to measure due to difficulties defining the target population for KiwiSaver and the limitations of using data from only the early stages of KiwiSaver. The main retirement income source for both KiwiSaver households (44%) and non-KiwiSaver households (38%) was New Zealand Superannuation. The next most commonly reported main source for KiwiSaver households was KiwiSaver (25%) compared with income from the sale or rent of property for non-KiwiSaver households (22%). 1 Law, D, Meehan, L & Scobie, G (2011). KiwiSaver: An Initial Evaluation of the Impact on Retirement Saving, Wellington, The Treasury, Treasury Working Paper 11/04 (2011 TWP) 2 Law, D & Scobie, G (2014). KiwiSaver and the Accumulation of Net Wealth. Wellington, The Treasury, Treasury Working Paper 14/22 (2014 TWP) 3 Retirement Policy and Research Centre (2014). Observations on Reserve Bank’s Household Financial Assets 2003- 2013. Pension Briefing 2014-1. University of Auckland Business School 3 Prepared by: National Research and Evaluation Unit 1 0F     position to enjoy standards of living in retirement similar to those in pre-retirement In 2013, KiwiSaver cost the Crown $677 million in contribution payments to members. This was a decrease from $1.5 billion in 2012 when the member tax credit was halved. Overall, in the first five years payments to members and employers have cost the Crown $5.3 billion.  Based on evidence collected across the first seven years of KiwiSaver, and in particular the first 3.5 years the success of KiwiSaver in achieving these is marginal at best. In 2010 the KiwiSaver scheme appeared to be reaching about one third of the estimated target population. This implied that the ongoing cost of the scheme for each target member, based on one particular approach to estimating the target group, was around $4,000 per year. The features designed to attract individuals to the scheme have been successful in doing so. However, the extent to which it has attracted the estimated target market, as defined in the 2011 Treasury Working Paper, appears to be limited, with substantial leakage. Further, much of the savings accrued through KiwiSaver are the result of substitution from other forms of savings and debt reduction. There is evidence to suggest that in the short term KiwiSaver has not had a positive effect on the accumulation of the net wealth of its members when compared with non-members. A costs and benefit analysis shows that for the period 2007/08 to 2013/13, the additional savings amongst the estimated target group for each $ of government spending ranged from $0.20 to $0.38 as the level of government contributions dropped with fewer new enrolments and policy changes. The analysis also used a narrower target group which produced a lower value for money. Given the importance of the assumptions used, this analysis could produce different results (better value for money) if a wider definition of the target group was used. The costs to the Crown are reducing over time as membership growth slows. The reduction to the member tax credit in 2012 and the increase in the default contribution rate will have, at least partially, caused this. However, there is the potential for the First Home Deposit Subsidy to reverse this trend should the uptake increase significantly. There is also evidence to suggest that any accumulated wealth has come at a significant cost to the Crown through contributions. The cumulated cost-benefit analysis undertaken does suggest that this could be improving, albeit from a low base. Consideration needs to be given to the extent to which judgements of KiwiSaver based on data collected primarily in the first 3.5 years of its existence can be seen as anything other than short-term impacts. The question that remains is whether these outcomes are likely to be different in the future. What, for example, could the impact of KiwiSaver be for those who are only just entering the workforce and will be contributing for thirty plus years compared with those who withdrew their funds in 2013 and contributed for approximately five years? One could also question the impact of the global financial crisis on these findings. 25% of the Crown subsidies were paid to the highest income quartile. The same group contributed more than 45% of the savings. 1.5 Conclusions This report has drawn together seven years of evaluative activity under the key objectives of the evaluation. It also includes a 2014 value for money study. With the exception of administrative data, much of the data reported was collected in the first 3.5 years of the implementation of KiwiSaver. That notwithstanding, the evaluation has provided a comprehensive picture of the implementation and early outcomes of KiwiSaver. It has been praised internationally as the most comprehensive 4 evaluation of a retirement scheme undertaken . P3F P Essentially KiwiSaver needs to be measured against its policy objectives, which were to:  encourage a long-term savings habit and asset accumulation by individuals who are not in a 4 Collard, S, & Moore, N (2009). Review of International Pension Reform. Research Report No 663. Department for Work and Pensions. Available from https://www.gov.uk/government/uploads/system/uploads/attachmen t_data/file/214434/rrep663.pdf 4 Prepared by: National Research and Evaluation Unit increase individual’s well-being and financial independence, particularly in retirement and to provide retirement benefits. 1.6 Limitations There are a number of limitations to the data reported throughout this evaluation. These relate, amongst other things, to the timing of the evaluation and to the nature and size of the sample populations. More specifically the limitations include:  the self-reported nature of much of the data about retirement income expectations and living standards in retirement  the timeframe of the evaluation and its inability to do more than provide indications of early outcomes with regard to policy objectives. Key data were collected in 2010 or earlier, when KiwiSaver was 3.5 years old  the focus, in 2014, on early retirees who are likely to be an outlier group whose behaviour and experiences are not generalisable to those who will be in KiwiSaver for much longer prior to retirement  the difficulty of determining causality in a complex environment  the possible influence of other factors (eg, the global financial crisis) on the outcomes  the small sample sizes in some of the qualitative work. 5 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report Section 2. KiwiSaver: A brief introduction In this section a brief introduction to KiwiSaver is provided as a context for the evaluation and the findings reported subsequently. 2.2 The mechanics of KiwiSaver 2.1 The purpose of KiwiSaver The key mechanics, or design features, of KiwiSaver are described in Appendix 1. These include features related to KiwiSaver being a work-based scheme including the automatic-enrolment and opt-out features and the incentives to attract membership, such as the kick-start payment and member tax credit. KiwiSaver is a voluntary, work-based savings scheme, designed to help people prepare for their retirement. It does, however, make provisions for the self-employed and children. The primary legislative objectives of KiwiSaver are to:  encourage a long-term savings habit and asset accumulation by individuals who are not in a position to enjoy standards of retirement similar to those in pre-retirement  increase individuals’ well-being and financial independence, particularly in retirement, and to provide retirement benefits. Findings related to the effectiveness of these features as they have been implemented are provided in Section 6 of this report. The policy drivers for the implementation of KiwiSaver were the perceived low levels of private saving for retirement and a concern that middle-income New Zealanders, in particular, were at risk of experiencing a substantial drop in their living standards during 5 retirement. A Treasury study (Le, et al, 2007) had shown that 20% of the population needed to save more for retirement. There were also fears that younger workers could have lower standards of living than current retirees and those approaching retirement. This was due to higher levels of personal debt from student loans, having children later in life and potentially fewer 6 mortgage free homes (Kritszer, 2008) . However, the prevalence of these policy drivers was questioned at the time. P4 F P5F P P The extent to which these policy objectives have been met and whether KiwiSaver has provided value for money are discussed in the final sections of this report. This includes a consideration of the additionality of any savings in KiwiSaver accounts and the extent to which individuals within the target market have been reached. 5 Le, T, Scobie, G and Gibson, J (2007). Are Kiwis saving enough for retirement? Preliimary evidence from SOFIE. New Zealand Treasury. 6 Kritszer, B (2008). KiwiSaver: New Zealand’s new subsidized retirement savings plan. Social Security Bulletin, Vol 67, 4), pps 113– 119. 6 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report Section 3. The KiwiSaver evaluation 3.1 The evaluation strategy: Purpose, focus and scope 3.2 Evaluation objectives The objectives of the KiwiSaver evaluation are listed below with the relevant sections of the report identified in brackets. 7 The KiwiSaver Joint Evaluation Strategy (the Strategy) was developed to cover the period from 2006/2007 through to 2012/2013. The purpose of the Strategy was to guide evaluation activities over the lifetime of the evaluation and to provide a common point of reference for the agencies involved. P6F P 1. To assess the implementation and delivery of KiwiSaver to inform ongoing development and services delivery (Section 4) 2. To monitor KiwiSaver usage to understand the scale and pattern of uptake (Section 5) Inland Revenue, the Ministry of Business, Innovation and Employment (MBIE), The Treasury, the Commission for Financial Capability, the Financial Markets Authority, the Ministry of Social Development, Victoria University of Wellington and Statistics New Zealand provided input and/or were consulted during the development of the Strategy. They also provided representatives to the Steering Group, which was formed to oversee the comprehensive evaluation that was undertaken. 3. To assess whether the key features of KiwiSaver are generating expected outcomes (Section 6) 4. To determine the impact of KiwiSaver on individuals’ saving habits and asset accumulation (Section 7) 5. To determine the impact of KiwiSaver on superannuation markets and the financial sector (Section 8). Key questions and indicators for each of these objectives are provided in the appendices to the Strategy. The Strategy outlined the focus, purpose and scope of the evaluation. The focus of the evaluation was stated as being KiwiSaver, not the broader “savings” policy. The purpose of the evaluation was to establish whether and how KiwiSaver was addressing and meeting its policy objectives. The scope was defined by the statement of general policy and purpose in the KiwiSaver Act 2006 as well as methodological and practical considerations, such as a realistic timescale. It was recognised that it would be difficult to determine the extent to which KiwiSaver contributed directly to a range of indirect outcomes and to extend evaluation activities beyond the foreseeable future to look at longer term economic or financial outcomes. For these reasons the initial scope was:  to assess all process aspects of the implementation of KiwiSaver  to measure whether there had been a change in savings behaviour as a result of KiwiSaver. 7 3.3 Evaluation design The overall evaluation approach was to, where possible, embed the collection and extraction of evaluation information into the design of KiwiSaver. This was to be augmented with additional research activities as necessary. The rationale for this was to: minimise research costs  minimise the imposition of survey work on provider, employers and individuals  establish long-term data collection . The evaluation of KiwiSaver commenced in the 2007/2008 financial year, with the final evaluation activities being undertaken at the end of the 2013/2014 financial year, one year later than anticipated. The initial six-year time frame for the evaluation was to take into account the early implementation of KiwiSaver, the rollout of the home ownership assistance and the expectation that longer term trends and broader The strategy can be accessed from: http://www.ird.govt.nz/aboutir/reports/research/reportks/research-ks-joint-eval-strategy.html 7 Prepared by: National Research and Evaluation Unit  impacts were unlikely to emerge until the end of the evaluation period. More specifically the key limitations include the following: The evaluation was coordinated by Inland Revenue, but implemented by Inland Revenue, MBIE and the Treasury. A memorandum of understanding, regarding the use of KiwiSaver data, was signed by these agencies in 2011. 1. At the time of this report KiwiSaver had only been operational since 2007 (seven years). This is problematic in terms of measuring long-term economic impacts or the impact of KiwiSaver on retirement well-being. 2. Much of the evidence centres on analysis of the SoFIE–Inland Revenue linked data set. These data were collected in the first 3.5 years of KiwiSaver. This also coincided with the first 3.5 years of the global financial crisis. 3. In 2013/2014, the first retirees to withdraw funds will only have been in the scheme for five years; those using it for first home purchases will only have had three years of savings. Neither of these groups could be seen as representative of the longterm impact of KiwiSaver on asset accumulation, or decumulation. Their experiences and behaviour will be influenced by the short time they’ve had to save. 4. The KiwiSaver policy objectives state that KiwiSaver is aimed at individuals who are not in a position to enjoy standards of retirement similar to those in pre-retirement. This group is frequently referred to as the target market. However, no explicit, demographic definition of this group was provided at the beginning of the evaluation. As a result, evaluating the effectiveness of KiwiSaver in reaching the intended market has always been problematic. 5. Critical elements of the value for money analyses, related to the flow measurements, rely on respondents’ perceptions and self-reported data. These include the measure of additionality and the extent to which the target market has been reached and any subsequent leakage. 6. A number of factors influence well-being in retirement and the ability of individuals to save for their retirement. This makes it difficult to attribute any causality to KiwiSaver or to ascribe definitive reasons for respondent behaviours. 3.4 Evaluation method Data for the evaluation were collected through four different sources/methods: 1. The evaluation relied heavily on administrative data collected by Inland Revenue for monitoring membership and contributions trends. These data have been reported in the annual reports to the 8 Minister . P7 F 2. 3. P 9 Data from the SoFIE –Inland Revenue linked data set (the linked data set) were used to gather more detailed information about membership demographics and the accumulation of net wealth. P8F A number of research activities were undertaken at different stages of the implementation of KiwiSaver. These included individual and panel surveys (faceto-face and telephone). Participant groups in these surveys included providers, employers and individuals (members and non-members). Standalone reports were developed and reported for 10 much of this work , which was generally contracted out to a research provider. These studies are listed in Appendix 2. P9 F 4. P P Specific analyses were undertaken of the wider data set by individual agencies, including MBIE, the Treasury and Inland Revenue. While MBIE focused on the housing incentives and supply side, the Treasury and Inland Revenue focused on value for money questions, additionality and the economic impacts of KiwiSaver. 3.5 Limitations to the evaluation KiwiSaver also has significant impacts on, and is influenced by, other outcomes that are of policy and public interest. These impacts include fiscal effects, effects on levels and forms of saving and investment, and on the financial sector, and redistributive effects that arise because the benefits and costs do not fall equally on all individuals. This broader context for KiwiSaver makes it difficult to evaluate its overall impact. Focusing too narrowly on its primary objectives can, potentially, ignore some of the other impacts of KiwiSaver. However, focusing too broadly will minimise the validity of any evaluation. There are a number of limitations to the KiwiSaver evaluation to date, which means some caution is needed when reading this report. These limitations are primarily related to the timing of the evaluation activities and the relatively short time period that KiwiSaver has been in existence. Many of these were noted when the evaluation was first planned. They have also been referred to in key reports. 8 These reports can be accessed at http://www.ird.govt.nz/aboutir/reports/research/report-ks/ 9 Survey of Family, Income and Employment 10 These reports can be accessed at http://www.ird.govt.nz/aboutir/reports/research/report-ks/ 8 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report Section 4. The implementation and delivery of KiwiSaver  4.1 Introduction This section considers the first objective of the KiwiSaver evaluation, which was to: the 2014 qualitative research for which 35 participants were interviewed (15 individuals, 15 employers and 5 providers). This is a very small sample size. Key findings from across these research activities have been collated under the respondent categories. assess the implementation and delivery of KiwiSaver to inform ongoing development and services delivery. 4.2 Employers It summarises information related to the experiences of providers, employers and employees (KiwiSaver members) with regard to the implementation of KiwiSaver. Much of the information reported in this section was used to inform the ongoing development and delivery of KiwiSaver prior to 2010. As such its value today is limited to historical record. 4.2.1 Support for KiwiSaver Overall, the employers interviewed in 2007 were supportive of KiwiSaver—believing a retirement savings plan was a good idea. However, across the 15 employers in the 2014 qualitative research a range of attitudes and behaviours towards contributing to KiwiSaver were found. These ranged from those that were openly supportive through to those who expressed a grudging acceptance that they had to comply. A key determinant appears to have been the perceived affordability of the contributions and the ability of their business to absorb any costs. More recent evidence relates to the attitudes and behaviours of different stakeholders, in particular, members. Key findings that are pertinent in 2014 have been included in the separate summary report. Evidence for this section was collated from:  Inland Revenue administrative data  panel interviews in 2007 and 2008 of employers (n=63) and their employees (n=34)  interviews with 18 providers in 2007  a survey of 1,728 small and medium enterprise (SME) employers in 2009 about KiwiSaver compliance costs. This was repeated in 2013  a supplementary survey of 640 tax advisors was also undertaken regarding compliance costs  in 2010, a further 253 SME employers were surveyed as a follow-up to the 2009 survey. These were employers who had changed their approach to remuneration due to KiwiSaver and/or had a workplace superannuation scheme prior to 2007  the 2010 survey of 750 individuals comprising both KiwiSaver members and non-members  the SoFIE–Inland Revenue linked data set (the linked data set)  the 2013 provider surveys (n=23) and interviews (n=11) undertaken as part of the early withdrawal research 4.2.2 Impact on workload In 2007, employers on the panel reported that the impact on their workloads had been minimal and less than anticipated. The implementation had been straightforward for most and they reported understanding the employer guide and set-up processes. Larger enterprises reported more difficulties and spending more time on initial implementation than SMEs. This was seen as being due to their more complex human resources processes and payroll schemes. The business owner or a payroll/accounts administrator was generally responsible for KiwiSaver in SMEs, while larger enterprises reported employing a team of people to implement KiwiSaver. 4.2.3 Inland Revenue communications Inland Revenue’s employer-specific communications were seen as central to the successful implementation of KiwiSaver. The campaign had reached all respondent employers on the 2007 panel and there were very high levels of awareness around most of the communication materials. The majority of employers interviewed as part of the panel (81%) reported that Inland Revenue material was straightforward and easy to understand. A 9 Prepared by: National Research and Evaluation Unit similar proportion (82%) reported they had received the material in time. auto-enrol new employees as they had not realised it was mandatory to do so. 4.2.4 Awareness of obligations The same group of participants was again interviewed in 2008. At that time they seemed more certain about how to meet their obligations and less concerned about operational issues. KiwiSaver had, reportedly, become part of their usual staff induction and payroll processes. They had also become more familiar with how KiwiSaver worked and most found it easy to include the compulsory employer contribution and employer tax credit in the PAYE process. In 2007, there was some variation in the extent to which employers reported being aware of their obligations. Most knew they had to check if new employees were eligible to join KiwiSaver and distribute KiwiSaver packs to both new and existing eligible employees. They also knew they had to deduct contributions and forward them to Inland Revenue with PAYE payments. However, there were lower levels of reported awareness with regard to sending details of new employees who had been automatically enrolled, acting upon an employee’s contributions holiday notice or the need to provide employees with an investment statement if they had an approved alternative scheme. 4.2.5 Compliance costs Compliance costs for KiwiSaver include the cost of administering KiwiSaver, the use of a tax advisor, learning about KiwiSaver and dealing with Inland Revenue. Both internal and external compliance costs were considered in detail in both 2009 and 2013 through the surveys of SMEs (Table 1). There was also some confusion over the processes around auto-enrolment and opting out. This confusion had the potential to create conflict between employers and employees. In some instances, employers did not Table 1: Overall compliance costs for SMEs by business size in 2009 and 2013 2009 2013 Micro businesses (1–5 employees) $623 $654 Small businesses (6–19 employees) $743 $724 Medium businesses (20+ employees) $721 $645 Average cost regardless of size $770 $661 Source: 2009, 2014 survey data Other findings from these surveys are presented in Table 2. Of note is the marked drop in the average external costs. This is likely due to the embedding of KiwiSaver into payroll practices in 2013, meaning there is minimal requirement for a tax advisor to be involved. Table 2: KiwiSaver compliance costs 2009 2013 Average hours spent on internal processes 14.5 14.9 Percentage of time spent by business owners 70% 52% 40–50% 50–60% Percentage of SMEs using a tax advisor for KiwiSaver 68% 20% Average external cost $49 $12 Percentage increase in payroll compliance costs Source: 2009, 2013 survey data 10 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report In the 2014 qualitative research, it was reported that larger organisations tended to absorb the administration costs into their overall compliance activities. Smaller businesses reported that as long as their employee turnover was low and/or or the pay systems straightforward, the administrative costs were also low. This was because KiwiSaver could then be absorbed into overall PAYE administration. 4.2.8 Existing superannuation schemes In 2009, very few respondent businesses (2%) reported having an existing superannuation scheme prior to the introduction of KiwiSaver. For those who had schemes, the 2010 research findings suggested that KiwiSaver was complementing, rather than replacing or adversely affecting, existing registered workplace superannuation schemes. 4.2.6 Remuneration approaches Of the 84 businesses, in the 2010 study, that had a scheme prior to 2007, most (71%) were still operating the scheme independently of KiwiSaver. In many instances, this was because the scheme offered employees unique benefits. These included higher employer contributions and being able to get a lump sum when they left. Consideration of remuneration approaches was considered in a number of separate studies:  In 2007, the employers on the panel wondered about the implications of employer contributions on their future cost structures and the ability of their companies to offer pay rises.  In 2009, it was reported from the survey of employers that the introduction of KiwiSaver had led to changes in the approach to remuneration for only 18% of the businesses.  Responses were similar in the 2010 follow-up survey, when most employers (81%) reported not changing their remuneration practices.  Employers, in the 2010 survey of SMEs, who had made changes most commonly applied a “salary sacrifice”, which tended to negatively impact on the uptake of KiwiSaver. That is, they paid those on KiwiSaver less than those who were not. This was, reportedly, done for reasons of fairness.  4.2.9 Informing employee decisions Most employers (62%) surveyed in 2010 had not taken up the option of nominating a KiwiSaver scheme for employees to join if they did not choose their own. There was concern expressed by some employers that they would be morally responsible for the scheme’s performance if they did so. Similarly, the majority of employers surveyed (78%) did not make general information about managing money available to employees. However, when this information was available, in combination with extra contributions and no existing scheme, the uptake of KiwiSaver was greater than when it was not. Individual members interviewed in the 2014 qualitative research, tended to view their KiwiSaver contributions as “extra” money, not part of their salary or wage package. Many also said they had not explicitly negotiated their salary or wage taking KiwiSaver into account, although some were aware that recovery of costs is achieved by some employers through such discussions. This suggests that when employers are actively involved in promoting the retirement savings of their employees through KiwiSaver there is greater uptake. The concern for employers would seem to be the moral responsibility they then assume for their employees’ savings. 4.3 Providers 4.2.7 Employer contributions Findings related to employer contributions in the 2010 survey of SMEs included the following:  Most respondent employers (89%) reporting being able to absorb their contributions as an extra cost to the business.  The majority (81%) were contributing the minimum default rate of 2%.  The most common additional rate, for those who were contributing more than the default rate, was 2%.  These respondents did not report changing their additional contributions when the employer tax credit was removed. 4.3.1 KiwiSaver schemes Since 2008, the number of KiwiSaver schemes has ranged from 54 in 2009 to 45 in 2013 (Figure 1), one less than in 2008. Of the 45 schemes in 2013, four had assets of over $1 billion at 30 June, with four having assets of less than $1 million. Just over half of the schemes (n=25) had assets of more than $50 million. 11 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report Figure 1: Number of schemes over time 60 55 50 45 40 35 30 25 20 15 10 5 0 2008 2009 2010 2011 2012 2013 Source: Inland Revenue administrative data 4.3.2 Engagement with Inland Revenue reported using social media, text and face-to-face meetings. Providers interviewed in 2007 (n=18) were asked about the engagement model implemented by Inland Revenue during the early implementation of KiwiSaver. Key findings include the following:  The default providers appeared to initiate communication with their members less frequently than others.  Overall, they considered the model had been successful.  The majority of providers directly inform their members of their account balance yearly.  They viewed it as a new and innovative development on the part of Inland Revenue.   Larger providers reported being more satisfied than smaller ones. All providers were aware of the need to better target their communications and products to eligible members in the future as KiwiSaver balances grow.  Inland Revenue had been more flexible and consultative than they had expected.  They were particularly appreciative of the relationship manager positions and the forums, which were seen as informative and timely.   4.3.4 Processes for withdrawal of funds As part of the survey (23 respondents) the 2013 provider research also considered the processes for withdrawing funds. Key findings include the following: While there were some weaknesses identified, these were perceived as being due to the tight timeframes for implementation and a lack of relationship managers given the number of providers. It was generally accepted that the engagement with the default providers was prioritised.  Most providers reported flexibility is allowed in the types and frequencies of withdrawals eligible members can make.  All providers allow members to withdraw their savings either partially or wholly as lump sums.  Nearly all (only four do not) also allow them to make regular withdrawals.  Most providers do not place a limit on the number of times members can make a partial lump sum withdrawal.  Most providers allow members to withdraw fortnightly.  The majority of providers allowing regular withdrawals allow this to occur monthly (17 out of 19). 4.3.3 Communication with their members In 2013, Inland Revenue interviewed 11 providers as part of a study of the first withdrawals for retirement purposes. Amongst other things the providers were asked about communications with their members. Key findings include the following:  The providers reported using a range of methods including letters, email and telephone. Some also 12 Prepared by: National Research and Evaluation Unit  provider and what information they should receive from their provider. Most providers place a limit on the minimum amount that can be withdrawn. More consideration was given in 2010 to the efficacy of the information strategies used to inform the general public about KiwiSaver and retirement savings. 4.4 Employees (individuals) 4.4.1 The efficacy of the design features in promoting membership The early panel research, in 2007 and 2008, considered the extent to which a number of design features were working as intended. Later research in this area is reported in Section 6. Key findings included the following:  KiwiSaver members were, firstly, asked whether they felt they had enough information to help them decide whether or not to join KiwiSaver. The majority of KiwiSaver members (80%) reported that they had.  Secondly, they were shown a card with visual 12 illustrations of Sorted information sources and asked if they had ever used Sorted information (such as the Sorted website, booklet and seminars) to help them make a decision about KiwiSaver. Around one in five (19%) KiwiSaver members used Sorted information to make a decision about KiwiSaver. 11 F  Employees were reportedly encouraged to join KiwiSaver because of the ease with which they could save and because they felt like they were getting “free money”.  However, those features designed to keep savings locked in until retirement had also, reportedly, proven to be a challenge for others. This was seen as being due to the apparent inflexibility of the scheme. This was particularly true in 2007, but less evident in 2008.  The extent to which these respondents were actually knowledgeable about KiwiSaver is reported in section 5.5.2, which discusses the findings from the construction of a knowledge index. In this work it was found that 53% of KiwiSaver members had excellent levels of 13 knowledge . Further, a concern that future governments might change, or discontinue, KiwiSaver had, reportedly, led some existing employees to postpone joining the scheme. This was true in both 2007 and 2008.  In 2008, the 4% default contribution reportedly remained the main feature discouraging individuals from joining.  These employees did not believe they could afford this level of contribution. What is not clear is whether this was due to limited incomes, other competing life-stage priorities or other investments. This is worth noting, particularly if consideration is ever given to increasing contribution rates. P12 F 4.4.3 Making decisions about their investments In the 2010 survey of individuals, KiwiSaver members were asked about their provider and investment fund decisions. Both these actions were included in the construction of an engagement index, which is discussed in detail in section 5.5.3. 4.4.3.1 Choosing a provider Key findings from the 2010 survey of individuals related to choosing a provider include: While the features have encouraged people to join, there is less evidence to suggest they have promoted long-term savings behaviours or led to substantial additional savings (as discussed in subsequent sections). 4.4.2 Learning about KiwiSaver features  50% of members surveyed had chosen their own provider  27% had joined their employer’s chosen provider  27% had considered more than one provider. Of those who had chosen their own provider, 40% said they had considered what other providers could offer. This was a higher proportion than of those who had been allocated a default provider (17%) or chosen their employer’s provider (13%). This suggests there is a distinct delineation between active engagement and passive involvement through the default features. In 2007 and 2008 members were, reportedly, more aware of the general features of KiwiSaver than the home ownership features. This was to be expected, as there had been no public awareness campaign about 11 either the mortgage diversion option or the first home ownership incentives. P10F P The research in 2008 highlighted that there was still a need for financial education for both members and potential members. Some panel participants wanted guidance for members on how to choose a scheme The most commonly reported determinants of choice, when selecting a provider, were related to perceptions of the financial security of the provider. These were 12 Sorted is an independent money guide website, providing a range of resources for New Zealanders. https://www.sorted.org.nz/?gclid=COyH1Zvkt78CFYQIvAodGK4ACw 13 Note this is a comparative measure not an exact one. 11 The mortgage diversion option was only available prior to 2009 and only where a provider agreed. 13 Prepared by: National Research and Evaluation Unit P mentioned by 62% of KiwiSaver members who had been actively engaged in the process of choosing a provider. More specifically, KiwiSaver members reported looking for a:  provider with a good reputation, which may have included factors other than financial security  financially stable organisation  provider that offers the level of risk they are willing to take. familiarity with the provider  perceptions of financial advantage  recommendations by others  characteristics of the investment funds offered.  Māori and Pasifika people  single people  those with a combined income of up to $30,000.  had a combined annual income of over $80,000  had postgraduate qualifications  had a greater knowledge of KiwiSaver. KiwiSaver members, in the 2010 survey, reported choosing different types of investment funds for a variety of reasons:  Of respondent members 68% gave a reason related to the expected risk and financial return of the investment.  The most important individual driver, and the factor that best helped people differentiate between funds, was getting the best long-term returns—33% said this was important. 14 Prepared by: National Research and Evaluation Unit Almost all members (96%) were aware of their eligibility to withdraw funds.  This awareness came primarily from personal knowledge when they joined (55%) and through communications from their provider (35%).  Most (75%) found the withdrawal process easy. Key findings related to employers’ experiences of KiwiSaver during its initial implementation include the following: Of those who knew what type of investment fund they had, 39% had considered more than one type. This group had relatively high proportions of those who were: employed full time  The purpose of this section was to assess the implementation and delivery of KiwiSaver. Much of the data reported was collected prior to 2010 and informed the early implementation of KiwiSaver. Its value can be seen in the early changes made to both processes and the scheme design features. In this group of characteristics, lower incomes are often associated with the other three descriptors, suggesting income may be the key determinant. This could be related to the amount of money they have invested, or are able to invest.  Not surprisingly, an individual’s attitude to risk was a key influencing factor on their choice of fund. 4.5 Summary of key points When asked about their investment fund, 40% of KiwiSaver members in the 2010 survey of individuals did not know what type of investment fund they had. This group included relatively high proportions of: young people  The 2013 early retirement survey asked respondents about their experiences in withdrawing funds. Key findings from the study include the following: 4.4.3.2 Choosing an investment fund  Intolerance for short-term fluctuations was also evident, with 27% saying that lower risks of ups and downs in the short term was an important factor. 4.4.4 The withdrawal experience Other reported determinants included:    Overall, employers interviewed in 2007 were supportive of KiwiSaver although various views were expressed in 2014. The key determinant of employer attitude appears to be the perceived affordability of the contributions.  In 2007, employers reported the impact on their workload had been minimal and less than anticipated.  Inland Revenue employer specific communications were seen as central to the successful implementation of KiwiSaver  By 2008, employers seemed largely aware of their obligations and not overly concerned about operational issues.  Across SMEs the average compliance cost in 2009 was $705. This decreased to $661 in 2013. External costs seem to have been the major area of reduction.  In 2010, 81% of employers interviewed said they had not changed their remuneration practices as a result of KiwiSaver. Those who had made changes most commonly reported applying a “salary sacrifice”.  In 2001, most employers (89%) reporting being able to absorb their employer contributions.  In 2009, very few respondent businesses (2%) reported having an existing superannuation scheme prior to KiwiSaver.  Concern was expressed amongst employers about either promoting a scheme with their employees or providing information. They appear to have been worried about being morally responsible for the performance of any scheme if they did so. Key findings related to providers’ experiences of KiwiSaver during its initial implementation include the following:  The number of providers dropped from 54 in 2009 to 45 in 2013.  In 2013, four schemes had assets of over $3 billion at 30 June. Just over half of the schemes had assets of more than $50 million.  Providers were satisfied with the early engagement with Inland Revenue, considering the communications strategy to have been successful.  A number of methods were used by providers to communicate with their members. All providers interviewed in 2013 reported a need to better target communications and products in the future. Key findings related to employee/individual experiences of KiwiSaver during its initial implementation include the following:  A key deterrent for individuals to join was a concern that future governments might change or discontinue KiwiSaver.  The 4% default contribution rate in 2008 was seen as unaffordable for a number of individuals and was reportedly the main reason they did not join.  The apparent inflexibility of the scheme also seen as a difficulty.  The main reason for joining KiwiSaver appears to be that it is seen as an easy way to save.  In 2010, the majority of KiwiSaver members (80%) reported they had enough information to help them decide whether to join or not.  Around 19% reported using Sorted. 15 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report Section 5. KiwiSaver usage: Scale and patterns of uptake 5.1 Introduction This section considers the second evaluation objective: to monitor KiwiSaver usage to understand the scale and pattern of uptake. To meet this objective, trend data related to membership and contributions are reported. These data provide evidence about the demographics of KiwiSaver members, the increase in membership over time and the amount of money being contributed by members. Contributions by the Crown and employers are discussed in Section 6 of the report, as they are key incentives to drive membership. There were three primary sources for data related to member and non-member demographics and characteristics. These were:  the 2010 survey of individuals (the 2010 survey)  Inland Revenue administrative data  the SoFIE–Inland Revenue linked data set (linked 14 data set) using waves 6 to 8 of the SoFIE study from October 2007 to September 2010. 13 F The linked data set composes administrative data and data from a longitudinal study. As such it has been reported wherever possible, rather than the 2010 survey data. The sample sizes (weighted and unweighted) for the linked data set are provided in Table 3. These figures relate to all data reported in this section and referred to as being from the linked data set as well as the tables in Appendices 4, 5 and 6. 14 October 2007 to September 2010 16 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report Table 3: Sample sizes for the linked data set Unweighted (n) Weighted (n) KiwiSaver members 6,889 908,962 Never been a member 10,982 1,224,802 All persons 17,871 2,133,764 Source: 2009, 2014 survey data 5.2 KiwiSaver membership 5.3 Enrolment trends By June 2013, the cumulative KiwiSaver membership had reached 2.15 million (as reported in administrative data). This was a net increase of over 1.4 million (200% increase) from the 716,637 members in June 2008. Although membership numbers continue to grow, as shown in Figure 2, the year-on-year increase in net membership has been steadily declining, with only a 9% increase in 2013 (Table 4), possibly due to near saturation. The June 2013 number of members equated to 53% of the eligible population including:  32% of eligible children  72% of people aged between 18 and 24. Figure 2: Total net enrolments (cumulative) over time 2,500,000 2,250,000 2,000,000 1,750,000 1,500,000 1,250,000 1,000,000 750,000 500,000 250,000 0 2008 2009 2010 2011 2012 2013 Source: Inland Revenue administrative data Table 4: Year-on-year increases in membership Cumulative total Number change Percentage change 2008 2009 2010 2011 2012 2013 716,637 1,100,540 1,459,942 1,755,932 1,966,444 2,146,843 383,903 359,402 295,990 210,512 180,399 54% 33% 20% 12% 9% Source: Inland Revenue administrative data 17 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report Members can join KiwiSaver in three ways:  opting in through a provider  opting in through their employer  being auto-enrolled.  the proportion of members opting in has remained relatively consistent over time (between 61% and 63%).  the proportion of members opting in through a provider has increased (38% in 2008 compared with 49% in 2013; reaching 50% in 2011 and 2012). Analysis of membership enrolments through administrative data over time (Figure 3) shows that: Figure 3: Proportion of members joining KiwiSaver by enrolment method 100% 90% 80% 70% 60% Auto enrolled 50% Opt in via employer 40% Opt in via provider 30% 20% 10% 0% 2008 2009 2010 2011 2012 2013 Source: Inland Revenue administrative data The number of individuals being auto-enrolled is influenced by the employment market, the number of people joining that market for the first time and the extent to which the market is fluid (ie, people are changing jobs). 5.4 Individual member demographics A substantial amount of work was undertaken to profile KiwiSaver members through the linked data set. These analyses are too detailed to include in any depth in this report. However, key tables from the report have been included in the appendices as a reference for those interested in considering this further. The number of individuals opting in is more difficult to explain, particularly given the consistency of the proportion over time. In 2008, one could have argued that the initial high proportion of opt-ins, relative to those auto-enrolling, would decrease. This is because individuals will continue to be automatically enrolled over time, either through being newly employed or changing employers for the first time. However, opting in is an active choice and the fact that KiwiSaver continues to attract additional members through this method is interesting. Appendix 4 is a detailed summary of membership and non-membership profiles.  Appendix 5 summarises demographic profiles by membership status and enrolment type. The analyses included a comparison between KiwiSaver members and those who had never enrolled. Certain characteristics were found to be more common amongst members than amongst those who had never been enrolled. Those in the following demographic groups were more likely to be KiwiSaver members than not: Given KiwiSaver has only been operational since July 2007, it is possible there remain significant numbers of people who have either not changed employers or did so when KiwiSaver was still new and, at that stage, opted-out. Based on the figures reported in section 6.2.1 about income sources, it appears that members who have opted-in are predominantly wage or salary earners who have not changed jobs, or initially opted-out than decided to opt-in. 18 Prepared by: National Research and Evaluation Unit   18–24 years old  university educated  partnered  working in lower white-collar positions  working in retail, manufacturing, rental and health care industries  earning $30,000 or less. as a safe option through an erroneous belief that it was 15 government guaranteed. P14 F 5.5.2 Knowledge of KiwiSaver features As discussed subsequently, there are a number of features of the KiwiSaver policy that are designed to either make it easy to join or attractive to do so through incentives. For these to promote membership, individuals need to be aware of them. The following groups were more likely to never have joined KiwiSaver:  33–55 years old  no formal qualification  working in upper blue-collar positions  no employment income  ethnic minority groups with high earnings ($80,000 and above). P In the 2010 survey of individuals, respondents were asked about their awareness of the different features. Key findings related to levels of awareness across all respondents were:  92% knew about the kick-start  58% knew about the members tax credit 5.5 Behavioural characteristics of members  81% knew about the employer contributions  64% of potential new homebuyers knew about the first home withdrawal and 34% about the first home subsidy In addition to demographic characteristics the behavioural characteristics of members were also considered in the 2010 survey of individuals. These were their:  71% of respondents were aware of nine out of sixteen listed features, a “good level of knowledge”  KiwiSaver members were more aware than nonmembers of the different features.  attitude to financial risk  knowledge/awareness of KiwiSaver  engagement with KiwiSaver. In addition to their knowledge of individual features, each respondent was assigned a KiwiSaver “knowledge score”, ranging from 0 to 16. This score was based on the number of KiwiSaver features that each respondent reported being aware of. For analysis purposes, respondents were grouped into four segments based on their knowledge score. The question of what really constitutes a “good” level of knowledge was not determined. 5.5.1 Risk taking Attitude towards risk was not found to be a defining characteristic for whether or not a person joined KiwiSaver. Survey respondents, whether KiwiSaver members or not, did not see themselves as risk takers:  47% of all respondents said they were average risk takers  43% of all respondents reported being low risk takers. Using this index, it was found that KiwiSaver members’ knowledge of KiwiSaver is significantly higher than that of non-KiwiSaver members. The only level where the difference between members and non-members was not statistically significant was for those with “good” levels of knowledge. Amongst both members and non-members, responses varied in terms of their willingness to accept more risk to achieve possible higher returns:  44% of members and 43% of non-members either disagreed or strongly disagreed that they were willing to accept more risk  non-members may be a little more risk averse, with members 12% strongly disagreeing compared to 7% of members. When this survey was undertaken, confidence in investment funds was generally low with many people having been financially hurt by the global financial crisis. This may have prevented some from joining KiwiSaver and influenced the extent to which they were willing to take risk. Conversely, KiwiSaver was perceived by many 15 A UMR national survey showed that 48% of KiwiSaver members surveyed thought their scheme had a government guarantee. This survey was undertaken in March 2010 and reported in The Press on 15 March. 19 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report Table 5: Percentage of each respondent group by level of reported knowledge Members Non-members Total sample Poor (0–4) 2 14 9 Reasonable (5–8) 13 23 19 Good (9–12) 32 33 32 Excellent (13–16) 53 31 39 Source: KiwiSaver Evaluation: Survey of Individuals, http://www.ird.govt.nz/resources/0/3/03e46600437177c5a25eb24e9c145ab7/ks-evaluation-individuals.pdf 37TU While the direction of causality cannot be determined— does membership drive knowledge or knowledge drive membership—there is an obvious correlation between the two. The profiles of these different knowledge groups were also considered. Higher knowledge of KiwiSaver tended to be associated with: U37T  those who don’t own a home  those with less positive health  those with lower combined annual incomes  those with lower net worth  those with lower educational qualifications  those with lower levels of participation in paid employment.  larger proportions of 50–65 year olds  New Zealand Europeans  people with partners  homeowners  people with excellent or very good health  those with higher combined annual incomes and higher net worth  people with higher educational qualifications 5.5.3 Engagement with KiwiSaver  those in full-time employment. An engagement score from 0 to 6 was also constructed. Points were allocated based on the extent to which members had made an active or passive decision across different stages of the enrolment process and as members. These included, for example, how many providers they had considered, how many investment funds they had considered and whether they had changed their contribution rate. It appears that knowledge of KiwiSaver is related to income and the accumulation of assets. The direction of causality is difficult to determine and is likely to be twoway. The more money people earn and the more assets they are likely to have accumulated the greater the concern they may have about both the security of those assets and how they can be grown. Conversely, lower knowledge of KiwiSaver tends to be associated with larger proportions of:  18–29 year olds  Māori, Pacific and Asian peoples  single people Table 6: Percentage of KiwiSaver members at each level of engagement Engagement Members % 0 (not engaged) 15 1 18 2 7 3 17 4 11 5 12 6 (highly engaged) Source: Ibid 21 20 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report As is shown in Table 6, 21% of members were found to be highly engaged. This group made informed decisions about whether to join, their provider and their investment type. The 15% with a score of 0 made no decisions about their KiwiSaver. They will have been automatically enrolled and stayed with a default scheme. 5.5.4 Attitudes of members towards savings In the 2014 qualitative research, where 10 KiwiSaver members and 5 non-members were interviewed, it was suggested that a variety of attitudes were driving behaviour. Some respondents were largely passive members of KiwiSaver making few, if any, decisions regarding their KiwiSaver accounts. Others were highly self-directed, confident and able to make their own decisions. High levels of engagement with KiwiSaver tended to be associated with:  50 to 65 year olds  New Zealand Europeans  people with a partner  not being convinced that it was the right thing to do  current homeowners   higher combined annual incomes (over $80,000) not having sufficient, uncommitted, resources to save for their retirement  higher net worth  prioritising spending over saving.  higher educational qualifications (postgraduate university qualification)  self-employed people. In the same research, it was suggested that life-stage and financial drivers are likely to “marry with KiwiSaver triggers and types of funds” and that that there is a potential for individuals to migrate from nonmembership to membership. It was also suggested that attitudes and behaviours could change over time. These changes were seen as being driven by changes to an individual’s life-stage and/or circumstances, increased knowledge and/or confidence in either KiwiSaver or their own decision-making, or as the amount invested grows. A number of reasons were suggested in the report for why individuals might not have joined KiwiSaver. These included: Conversely, lower levels of engagement with KiwiSaver tended to be associated with:  higher proportions of 18–29 year olds  single people  those who do not own a home  lower combined annual incomes (especially under $30,000)  lower net worth (especially negative net worth)  lower educational qualifications (especially school qualifications). 5.6 Summary of key points The purpose of this section was to report information related to the scale and pattern of uptake of KiwiSaver. Key findings include the following: These findings suggest that, overall, levels of engagement with KiwiSaver were not high in 2010. At that time, these levels of engagement were seen as potentially being a consequence of the design features of KiwiSaver (auto enrolment, default providers). The thought then was that as savings grew, engagement would also increase. Evidence reported subsequently with regard to choosing and/or changing providers and schemes suggests this may be happening.  By June 2013, the cumulative KiwiSaver membership had reached 2.15 million, equating to 53% of the eligible population. This was a net increase of over 1.4 million from the 716,637 members in June 2008.  Although membership continues to grow, the yearon-year increase in net membership has been steadily declining with only a 9% increase in 2013 compared with 54% in 2009. There is a strong positive correlation between levels of engagement with KiwiSaver and knowledge of KiwiSaver. Only 33% of those with an engagement score of 0 or 1 have “excellent” knowledge of KiwiSaver compared to 54% of those with an engagement score of 2 to 4 and 72% with an engagement score of 5 or 6. This is also reflected in the similar demographics for these groups.  The proportion of members opting in has remained consistent over time (between 61% and 63%). Given the high proportion of members on wages or salaries it is likely these are individuals who have not changed jobs or who initially opted-out and then changed their minds.  From the 2010 survey it was found that attitude towards financial risk was not a defining characteristic for whether or not an individual joined KiwiSaver. Nearly half of members (44%) and non-members (43%) reported they either disagreed or strongly disagreed with the statement that they 21 Prepared by: National Research and Evaluation Unit were willing to accept more risk to achieve possible higher returns.  Of respondents to the 2010 survey 71%were aware of 9 out of 16 KiwiSaver features. It appears that KiwiSaver knowledge is associated with income and the accumulation of assets. The more assets an individual has the higher their level of knowledge about KiwiSaver.  Of respondents to the 2010 survey 21% were highly engaged with KiwiSaver. That is they will have made informed decisions about whether to join, their provider and their investment type. There was a very strong correlation between levels of engagement and knowledge. 22 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report Section 6. The features of Kiwi Saver A number of changes have been made to these KiwiSaver features since it was first introduced. These are summarised in chronological order in Appendix 3. 6.1 Introduction This section considers the third evaluation objective: 6.2 Making it easy to join to assess whether the key features of KiwiSaver are generating expected outcomes. 6.2.1 Kiwi-Saver as a work-based scheme A brief description of each feature is included at the start of most sections. Appendix 1 also contains a description of each. Information presented in this section was drawn from: As a work-based scheme KiwiSaver is designed to enable members to make contributions through salary or wage deductions, minimising the effort required. It is also designed to include employer contributions. This means:  administrative data   the 2010 survey of individuals (the 2010 survey) the scheme is essentially administered by employers  the SoFIE–Inland Revenue linked data set (the linked data set).  individuals are automatically enrolled when they begin any new employment (unless already members)  deductions are made at source. The features have been grouped into the following three categories: 1. The impact of KiwiSaver being a work-based scheme can 16 be seen in the income sources of members (Figure 4): Those that make it easy to join KiwiSaver: P 2. 3.  being administered by employers  automatic enrolment when changing employers  deductions made at source  default contribution rate  default KiwiSaver scheme Those that make it hard to get savings out of KiwiSaver before the age of 65:  limited opportunity to opt-out  locked-in savings  the majority of KiwiSaver members (55%) earn a salary and/or wage only  a further 19% earn a salary and/or wage plus other income  for up to 56% of KiwiSaver members their deductions will be made at source  only 4% of members’ incomes come solely from non-wage or salary sources Those with no income sources are likely to be predominantly children or young adults, signed up by parents or other family. Those that make it attractive (incentives):  kick-start payment  member tax credit  ability to withdraw savings towards a first home mortgage and first home deposit subsidy  contributions holidays  employer contribution. 16 23 Prepared by: National Research and Evaluation Unit 15 F These figures are from Inland Revenue administrative data KiwiSaver evaluation: Final summary report Figure 4: Number of members by source of income (2013 annual report) 1,400,000 1,200,000 1,000,000 800,000 600,000 400,000 200,000 0 Salary/wage only Salary/wage + other Other only No income Source: Inland Revenue administrative data 6.2.2 Automatic enrolment  self-employed All individuals who start a new job, whether a first job or through changing employers, are automatically enrolled in KiwiSaver. If they do not wish to belong they must opt-out between two and eight weeks after their autoenrolment.  in receipt of interest and investment  beneficiaries  unemployed. Those who opted-in through the employer have higher proportions of being: In June 2013, 39% of members had been auto-enrolled compared with 61% who had opted-in through a provider or their employer (Figure 3). This has remained consistent since 2009 with the number auto-enrolled each year ranging from 37% to 39% of the cumulative total for the year. The linked data set was used to compare the demographic characteristics dependent on how individuals had enrolled. single and young (18–24 years old)  Māori and Pacific islander  secondary school educated  in lower white-collar positions  in employment earning $30,000 dollars or less. 55 plus  Asian  middle-aged female European  university educated  in upper white-collar positions  in employment earning between $30,000 and $80,000. Administrative data were used to directly compare whether members have made an active choice, or were allocated to a scheme by either Inland Revenue or their employer over time (Figure 5). 24 Prepared by: National Research and Evaluation Unit  Members can choose their own KiwiSaver provider, join one that has been nominated by their employer, or be allocated to a default scheme by Inland Revenue. As with auto-enrolment, the default scheme feature is designed to make it easy for members. It removes the necessity to make decisions unless they wish to. It also minimises the need for financial knowledge. Both of these can have unintended consequences in terms of the levels of engagement and knowledge of individuals with regard to their KiwiSaver accounts. KiwiSaver members who opted in through a provider have relatively higher proportions of being:  35–55 years old 6.2.3 Choosing a KiwiSaver scheme Compared to other enrolment methods, KiwiSaver members who were automatically enrolled have relatively high proportions of being:   Highlighted is the extent to which members have chosen their own scheme. Points to note include:  those in an employer-nominated scheme decreased slightly from 13% in 2008 to 9% since 2011  the proportion of those choosing their own scheme has been gradually increasing from 49% in 2008 to 67% in 2013. In 2008, 38% of members were in a default scheme compared with 25% in 2013. Figure 5: Percentage of all members by means of choosing a scheme 100% 90% 80% 70% 60% Active choice 50% Employer 40% Default 30% 20% 10% 0% 2008 2009 2010 2011 2012 2013 Source: Inland Revenue administrative data Analysis of the linked data set showed that, in 2010, the method by which an individual was enrolled largely determined the choice of scheme (Table 7). In the next section the extent to which members change schemes once enrolled is discussed. Table 7: Enrolment method and scheme choice (percentage) Enrolment type % Method of choosing scheme % Active Default Employer Opt-in via provider 39.4 98.2 1.0 0.2 Opt-in via employer 22.1 10.2 70.1 19.5 Auto-enrolled 38.6 1.7 74.2 23.9 Source: Inland Revenue administrative data 6.2.4 Changing schemes (providers) question was considered in the analysis of the linked data set. Members who have been automatically enrolled can transfer schemes (providers) within a three-month provisional holding period. Members can also elect to change schemes at any point during their membership. The majority of member respondents to the 2010 survey (90%) had not changed their provider. Those who had changed provider were more likely to have chosen their provider in the first place; again suggesting there is a group of members who are very actively engaged, compared with some who do not take a strong interest In the 2010 survey KiwiSaver members were asked whether they had ever changed providers. The same 25 Prepared by: National Research and Evaluation Unit in their funds. Further detail is given on levels of engagement in section 5.5.3.  This was an increase of 24% (26,117 members) over 2012 numbers. This was after a decrease in the number of transfers since 2010. Analysis of the linked data set also showed that, in 2010, only a few members had changed schemes. The demographic profile of KiwiSaver members who had changed schemes was considered in the analysis of the linked data set. Appendix 6 provides details of the results of this analysis.  With the exception of 2008, the majority of transfers occur outside the holding period.  The number of transfers peak in March and May each year. This could indicate a link to the tax year; people are likely to be engaged in financial decisions at this time of the year when completing tax returns. More recent administrative data shows that increasing numbers of members are choosing to change schemes (Figure 6). This suggests increasing engagement with retirement savings by individuals. Key points to note include the following:  There is a marked increase between 2009 and 2010. This is followed by a decline for three years before a noticeable increase between 2012 and 2013. In 2013, the number of members who had transferred schemes was 136,167. Figure 6: Number of members transferring schemes 160,000 140,000 120,000 100,000 Holding period 80,000 Standard 60,000 40,000 20,000 0 2008 2009 2010 2011 2012 2013 Source: Inland Revenue administrative data 6.2.5 Default contribution rates As Figure 7 clearly shows, the default, or minimum rate, has been the most common, over time, whether 2% or 3%: There is a minimum default rate for KiwiSaver. This is the rate individual contributions are set at when they are deducted at source. Individuals can also choose to set their deductions at a higher rate. The contribution rate is an area of KiwiSaver that has undergone multiple changes. When KiwiSaver was first introduced, the minimum contribution from salary or wages was 4%, this decreased to 2% in 2009 before being raised to 3% in April 2013. This is the default rate. 41% of all members were paying the default rate in 2010, when it was 2%  58% are paying the default rate of 3% in 2013  the proportion of members contributing 4% has decreased from 55% in 2010 to 36% in 2013. The decrease in those paying 4% does not appear related to the change to 3% as the main decrease was between 2010 and 2011. Rather it may be related to concerns over the economy at the time and individual perceptions of what was affordable. Currently, members on wages or salaries can also choose to contribute 4% or 8%. Other options such as lump sum payments or regular payments for those not on a wage or salary will be determined by their scheme provider. 26 Prepared by: National Research and Evaluation Unit  KiwiSaver evaluation: Final summary report Figure 7: Proportion of members contributing each rate by year 100% 80% 8% 60% 4% 40% 3% 20% 2% 0% 2010 2011 2012 2013 Source: Inland Revenue administrative data In the 2010 survey only 15% of KiwiSaver members who made contributions directly from a salary or wage reported making a change to their contribution rate since joining KiwiSaver:  3% had increased their rate  12% had decreased their rate  3% had made additional contributions. During the analysis of the linked data set, cumulative contributions over time were analysed. Appendix 7 summarises the mean and median levels of cumulated contributions for a cohort of members who joined KiwiSaver during wave 6 through to the end of wave 17 8 . P16F Key findings from this analysis were as follows: Of the 1,000 early retirement withdrawal respondents surveyed in 2013:  81% contributed at the minimum 4% rate when they joined  6% reported they decreased their contributions to 2% in 2009 when the minimum rate dropped. P This supports the view that most members, deducting at source, remain on the minimum default rate. It also explains why there was still a relatively large proportion contributing at 4% in 2010. However, it must be remembered that this group would all have been approaching 60 when they joined and are likely to be an outlier group based on a very bounded set of demographic characteristics.  Members who opted-in through an employer had the largest mean value of total contributions.  There is a direct correlation between income and contributions.  Males generally have a higher mean value of contribution than females.  The value of contributions peaks at 45–55 years of age.  Compared to contributions made by other ethnic groups Māori cumulate the least.  Members who earn salaries and wages contribute the most. There is a group of KiwiSaver members who are not contributing to their accounts (Table 8). This group is larger than that on contribution holidays alone and, as reported subsequently (section 6.4.3), many of those on contributions holidays do make voluntary contributions. Overall, the findings reported here highlight the importance of the default rate. To achieve the goals of the policy regarding providing for people in retirement the default rate needs to be matched to member’s needs and aspirations for retirement, without being so high that they find it too difficult to maintain, or too low for them to achieve their goals. These members could be children who were joined by their parents or self-employed individuals. They could also be members who were in a salary or wage position but are no longer employed in that manner. 6.2.6 Contribution amounts While the minimum contribution rate is clearly an important factor in determining the amount of savings achieved through KiwiSaver it is also important to consider how much has been saved. 17 27 Prepared by: National Research and Evaluation Unit October 2007 to September 2010 KiwiSaver evaluation: Final summary report Table 8: Percentage of KiwiSaver members not contributing to their accounts Year Proportion of KiwiSaver members not contributing (all ages) Proportion of Kiwisaver members not contributing (18+) 30 June 2010 35% 22% 30 June 2011 37% 24% 30 June 2012 37% 26% 30 June 2013 38% 27% Source: Inland Revenue administrative data 6.3 Locked-in features  18–34 years old  university educated Two design features of KiwiSaver are aimed at locking in members’ savings once they have been enrolled. These are a limited opportunity to opt-out and the limited opportunities to withdraw savings until members are 65. As discussed in the earlier research, in the past these features have acted as difficulties for some.  earning $30,000 or less in employment  working in lower level positions. 6.3.1 Optingout of KiwiSaver In the 2010 survey of individuals, the main reasons respondents gave for opting out were: This is very similar to the demographic of those who are auto-enrolled. The first of the features aimed at making it difficult for individuals to get savings out of KiwiSaver is the need to actively opt-out within eight weeks of starting a new job. This feature ensures people join KiwiSaver, ie, membership is the default position. Their contributions, and those of their employer, begin the day they start work. Choosing to opt-out is an active process.  affordability (32%)  the belief that there were better financial alternatives (30%). These were similar reasons to those provided in the early employer panel research in 2007/2008. Appendix 5 provides demographic information about members who have opted-out. It shows that compared to those who had never been KiwiSaver members, those who had opted-out have relatively high proportions of individuals who are: Over time, those opting out have fallen as a proportion of those being auto-enrolled (Figure 8). Figure 8: Percentage of those auto-enrolled that opt out 100% 80% 60% 40% 20% 0% 2008 2009 2010 2011 2012 Source: Inland Revenue administrative data 28 Prepared by: National Research and Evaluation Unit 2013 KiwiSaver evaluation: Final summary report In 2013, the cumulative number of individuals who had opted-out of KiwiSaver, and remained out, was 249,872. This was the first time the cumulative number who had opted-out had decreased after four years of rises. However, the percentage increase had been declining over time (Table 9). Table 9: Year-on-year increases in numbers opting out Cumulative total 2008 2009 2010 2011 2012 2013 137,762 221,045 245,898 249,549 255,935 249,872 83,283 24,853 3,651 6,386 –6,063 60% 11% 1% 3% -2% Number change Percentage change Source: Inland Revenue administrative data Whatever the underlying reason for the decreases reported here, one could argue the auto-enrolled/optout feature is now more effective than it was in the early days of KiwiSaver. The one word of caution here is the increase in those taking longer holidays and those not making contributions. This could well be affordability and short-term prioritising of funds. 6.3.2 Locked-in savings KiwiSaver contributions are generally locked-in until members are 65. Provided they have been a member for a minimum of five years they can then withdraw all or part of their savings. If the provider agrees, they can also make regular withdrawals at this time. Members can only make early withdrawals for the following reasons: 1. buying their first home provided he or she has been a member for three years and the home is for them to live in. The first withdrawals under this facility started in 2010 2. significant financial hardship 3. moving overseas permanently (excluding 18 emigration to Australia) . P17F 4. P serious illness or permanent disability affecting their ability to work. In June 2013, the cumulative number of individuals who had closed their accounts was 71,720. This was more than double the 2012 number (32,227). This was largely due to the first KiwiSaver members being able to close their accounts for retirement purposes (Figure 9). 18 KiwiSaver members emigrating to Australia can elect to have their funds (including their member tax credits) transferred to an Australian Superannuation Scheme. They cannot withdraw their funds. 29 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report Figure 9: Cumulative number of individuals closing their accounts 80000 70000 60000 50000 Retirement withdrawls 40000 Other withdrawls 30000 20000 10000 0 2008 2009 2010 2011 2012 2013 Source: Inland Revenue administrative data 6.3.3 Withdrawal for retirement and typically had small amounts accumulated in KiwiSaver. Their withdrawal behaviour may be quite different from retirees who, in the longer term will have larger amounts to withdraw. Those who joined KiwiSaver shortly before retirement may also be different from those who joined at much younger ages. The early retirees are likely to have other retirement plans and would have been nearing the end of their working careers. Joining KiwiSaver was likely to have been a deliberate decision to increase their retirement funds through the incentives. Between 1 July 2012 and 30 June 2013, 73,992 members became eligible to withdraw their savings for retirement purposes. This was the first cohort of members able to do so. Of these 28,549 (39%) did so, with the majority making full withdrawals. In the first year of KiwiSaver a larger proportion of those aged 60 plus joined compared to later years. Consequently, there were over 25,000 members eligible to withdraw their funds in July/August 2012 compared to the 6,000 in July/August 2013 (Table 10). 6.3.4 The funds available to early retirees In 2013, two pieces of research were undertaken regarding the withdrawal of KiwiSaver funds for retirement using this group of early retirees:  a telephone survey of 1,000 eligible KiwiSaver members.  an online survey (n=25) and interviews (n=11) with KiwiSaver providers. As evidenced in Table 10, the first groups able to withdraw funds from KiwiSaver had limited savings: In considering the findings from these pieces of research, it is important to remember their generalisability is limited. The early retirees are essentially outliers as they are a very specific group of people. They only had a limited period of membership 30 Prepared by: National Research and Evaluation Unit  in 2012 66% had less than $15,000 saved  in 2013 69% had less than $15,000 saved  in both years just over a third had between $10,000 and $15,000 saved (39% and 35%). KiwiSaver evaluation: Final summary report Table 10: Members eligible to withdraw funds in July/August of each year by estimated total contribution funds Estimated savings ($) 2012 2013 n % n % 117 <1 180 3 1001–5000 1,507 6 771 12 5001–10,000 3,694 21 1,185 19 10,001–15,000 11,359 39 2,209 35 15,001–20,000 4,278 17 834 13 20,001–25,000 2,102 8 492 8 25,001–30,000 1,021 4 233 4 30,001–35,000 551 2 139 2 35,000+ 843 3 249 4 25,472 100 6,292 100 1,000 Total Source: Inland Revenue administrative data 6.3.5 The withdrawal behaviour of the early retirees This withdrawal behaviour was found to be strongly correlated with the value of their KiwiSaver accounts (Table 11): The withdrawal behaviour of this early group of retirees was studied in the 2013 early withdrawal research. Key findings include:  36% of the sample had withdrawn all of their savings  57% had not withdrawn any of their savings  7% had made a partial withdrawal.  Those with balances up to $15,000 were likely to have withdrawn their entire savings.  Those with balances over $15,000 were more likely to have withdrawn none of their savings. The group most likely to have withdrawn some of their savings were those with more than $30,000. Table 11: Percentage of respondents withdrawing proportion of funds by value of KiwiSaver savings All Some None Less than 10,000 45 8 47 10,001 to 15,000 50 4 46 15,001 to 20,000 25 9 66 20,001 to 30,000 22 9 69 More than 30,000 18 14 68 Source: Inland Revenue administrative data 31 Prepared by: National Research and Evaluation Unit Respondents to the 2013 survey were asked about either their actual or intended behaviour about withdrawing funds: withdrawals and even fewer were making regular withdrawals. The providers did suggest that this was likely to change over time as KiwiSaver balances grew.  81% had already/intended to only withdraw one or more lump sums  4% had already/intended to only make regular withdrawals  Respondents to the 2013 survey were also asked when they intended to have fully withdrawn their KiwiSaver funds. As Table 12shows: 14% were unsure of their future intentions. These findings were supported by the survey of providers, undertaken at the same time, where it was reported that only a few members were making partial  36% had already done so  36% intended to do so within the next five years  11% intended to do so in more than five years. Table 12: Percentage of respondents by intention to withdraw all of their funds % of respondents Already withdrawn 36 In the next year 11 2–3 years 12 4–5 years 13 6–10 years 7 More than 10 years 4 Don’t know 15 No plans to ever do so 3 Source: Inland Revenue administrative data 6.3.6 Withdrawal for reasons outside retirement If one looks only at those withdrawing their funds for reasons other than retirement the rate of increase has slowed markedly in 2012 and 2013 compared with 2010 and 2011 (Table 13). Table 13: Year-on-year increases in the proportion of individuals withdrawing funds for reasons other than retirement 2008 2009 2010 2011 2012 2013 1,044 8,240 13,656 25,559 32,337 43,171 Additional number 7,196 5,416 11,903 6,668 10,944 Percentage increase 689% 66% 87% 26% 34% Cumulative number Source: Inland Revenue administrative data 32 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report 6.4 Making KiwiSaver attractive  In the 2012 year, $455 million in MTC payments was made to 1,232,453 members.  Less than half of these members (45%, n=556,652) received the maximum amount. There are a number of features built into KiwiSaver to make it more attractive for individuals to join. These include the contributions from the Crown and employers. Two other features designed to make KiwiSaver more attractive, and lessen the severity of the locked-in savings, are the ability to withdraw funds for a first home and the contributions holiday. A further incentive is the First Home Deposit Subsidy (FDS).  This proportion remained stable from 2009 through to 2012.  The median annual contribution rate for members earning solely wages or salary was $859 in 2013.  For those earning only non-salary or wage income the median contribution was $1040. The explanation for this low percentage of eligible members receiving the full tax credit appears to be related to the default contribution rate. A full-time worker on the minimum wage would have contributed approximately $936. This would be insufficient to receive the full credit. Part-time workers would be even less likely to reach the threshold. 6.4.1 Employer contributions One of the incentives of KiwiSaver, and a feature of it being a work-based scheme, is the employer contribution. Employer contributions were one of the reasons given for joining KiwiSaver by 56% of respondent members to the 2010 survey of individuals; it was the most important reason for 5%. That the median contribution for non-salary or wage members is close to the amount required to achieve the full tax credit suggests many of these members are contributing to maximise the value of the MTC they receive. Key findings related to this feature include the following:  Over the year to June 2013, Inland Revenue passed $833 million in employer contributions to scheme providers for investment in their employees’ accounts.  Over the first five years of KiwiSaver employers contributed $3.5 billion to KiwiSaver accounts.  Most employers (81%) contribute the minimum 3% of the salary or wages of their employees. 6.4.3 Contributions holidays Another feature designed to make KiwiSaver more attractive to potential members is the opportunity to take contributions holidays. Members who have been making contributions for 12 months or more can take a contributions holiday of between three months and five years. Early holidays (before 12 months) are considered only in the case of financial hardship. 6.4.2 Crown contributions The Crown contributes to KiwiSaver accounts in two ways:  the kick-start payment of $1,000  the annual member tax credit. In the 2010 survey, contributions holidays were a reason for joining KiwiSaver for 12% of members, with 9% of respondents saying they were the most important reason. An annual member tax credit (MTC) is paid to members 18 years or older until they are eligible to withdraw their savings for retirement purposes. To receive the maximum annual MTC payment members must have been a member for a full 12-month period and contributed at least $1042.86 to their account. The maximum amount of the MTC was decreased in 2011 from $1042.86 to $521.43. Employer contributions and government contributions do not count towards the eligibility for this credit. As at June 2013, 127,974 members had taken a contributions holiday since the initiative began. Of these 101,145 were on holiday at that time. Of note from Table 14 and Table 15: Along with the employer contributions these are the two main incentives for individuals to join KiwiSaver. In the 2010 survey, 67% of respondents said these incentives were a reason for joining; with 19% saying they were the most important reason. Key findings related to these incentives include the following: 33 Prepared by: National Research and Evaluation Unit  the number on a hardship holiday has steadily declined  the number on an ordinary holiday has steadily increased. This is likely due to the cumulative effect of those on holidays for more than one year  the annual percentage increase is actually decreasing. This suggests fewer new people taking a holiday over time  most have only taken one holiday. KiwiSaver evaluation: Final summary report Table 14: Number of members on a contributions holiday (years ended 30 June) 2008 Ordinary 2009 2010 2011 2012 2013 25,112 45,069 63,324 83,151 101,141 Hardship 3,820 813 494 383 219 274 Total 3,280 25,935 45,563 63,707 83,370 101,145 Number increase 22,655 19,628 18,144 19,663 17,775 Percentage increase 691% 76% 40% 31% 21% Source: Inland Revenue administrative data Table 15: Number of contributions holidays taken by members (June 2013) Holidays (n) Number % 1 111,172 87 2 10,374 8 3 3,756 3 4+ 2,132 2 127,974 100 Total Source: Inland Revenue administrative data Administrative data show that the number of people on long holidays (five years) continues to increase, while the number on short holidays (up to one year) has decreased. Key findings from the 2010 survey with regard to those on contributions holidays include the following:  46% of those who had taken a holiday did so due to financial hardship related to employment factors In the analysis of the linked data set consideration was given to the demographics of those on contributions holidays at the time (Appendix 8). The analysis showed that:  75% of those who had taken a holiday had experienced a major life change in the last 12 months compared to only 57% who had not taken a holiday  the income distribution of those on contributions holidays was generally consistent with the income distribution of KiwiSaver members  20% of those who had taken a holiday described their health as only “fair” or “poor” compared with 5% of those who had not.  the gender distribution was also consistent with that of KiwiSaver members  76% of those on a contributions holiday were aged between 18 and 44, with many of them aged between 25 and 34. 6.4.4 Purchasing a home KiwiSaver, while essentially a retirement income scheme, also aims to enable New Zealanders to accumulate more assets. Given the extent to which home ownership can be seen as not only an investment for retirement but also important for wellbeing, there are two homeowner features built into the design of KiwiSaver. These are: Members on a holiday can make voluntary contributions. Administrative data, reported in June 2011, showed that most had made contributions during that time. This would suggest that financial reasons are driving the decision to take a holiday. It may be these members do not feel they can afford the regular contributions from their salaries or wages. A contributions holiday would be one way of controlling their contributions. 1. 34 Prepared by: National Research and Evaluation Unit the first home deposit subsidy (FDS). The FDS is $1,000 for each year an individual has been contributing to KiwiSaver, up to a maximum of $5,000. Couples can receive a combined subsidy of $10,000. To receive the subsidy members must have been contributing at the default rate for at least three years. 2. 6.4.4.2 Uptake of the features A study of the uptake of the homeownership features, using administrative data, was completed in 2013 by the MBIE. It showed there has been an increase in the use of these features since 2010. This increase is most likely a reflection of the growing number of people who have become eligible to use these features. It may also reflect rising house prices and the, well documented, difficulties faced by first-home buyers in some regions. the first home withdrawal (FHW). Through the FHW members can withdraw all their savings (with the exception of the Crown contributions) to purchase their first home. Again they must have been a member for three years and be going to live in the purchased house. The objective of these features is to assist members to enter home ownership by helping them overcome the barrier of not having sufficient funds to purchase a first home. Both became operational in July 2010 when the first members were eligible to withdraw funds. The study showed that, as at December 2012: In the 2010 survey, it was found that these incentives were a reason for joining for 24% of members; with 10% saying they were the most important reason. Housingspecific research undertaken by the Department of Building and Housing (now Ministry of Business, Innovation and Employment or MBIE) showed that a number of people joined KiwiSaver specifically for the first-home buyers package.  more than 11,000 withdrawals for first homes had been made  there were more than 5,800 recipients of the 19 deposit subsidy 18 F  the highest users of both components were younger people (under 35 years)  the proportion of residential sales drawing on FDS support increased from 0.1% in July 2010 to 5.3% in December 2012. In the study it was found that: Housing subsidies could have an important impact on Crown costs moving forward. As the proportion of younger members increases through automatic enrolment it is likely the use of these subsidies will also increase. In addition, from 1 April 2015 the subsidy available for first home buyers who purchase newly built homes will double.  most applications for the FDS were from two or more people and were to purchase a first home  three quarters of these sought the minimum subsidy of three years  half of the applications were for houses under $300,000  of the applicants, almost three-quarters were New Zealand European  most applicants had an annual income of between $30,000 and $60,000  72% of KiwiSaver members who were aware of the feature said they were either “very likely” or “quite likely” to apply for a subsidy. the highest proportion of approvals was for the Auckland region  over a quarter (28%) intended to apply within two years of its introduction prior to applying, most had already been saving for a deposit and they applied because they needed extra funds  for over half of those who received a deposit subsidy the extra funds were essential to being able to purchase a home. 6.4.4.1 Intentions to use the features In the 2010 survey of individuals, KiwiSaver members who were aware of the deposit subsidy were asked how likely they would be to apply for a home deposit subsidy. Their responses suggest usage will increase over time. Key findings include the following:    four in 10 (42%) intended to apply within five years and 24% sometime after 2015  it was also found that 56% of those who were aware of the first home deposit subsidy were likely to buy a home in the next five years compared to 38% of those who were not aware of this feature. Reasons for disinterest in using either feature primarily related to greater importance being placed on retirement savings (28%) and financial constraints (22%) than on home ownership. 19 Not everyone who withdraws qualifies for a subsidy as well as there are both income limits for the applicants and price caps on the home. 35 Prepared by: National Research and Evaluation Unit 6.5 The overall efficacy of KiwiSaver features 6.6 Summary of key points The purpose of this section was to consider the extent to which the key features of KiwiSaver were generating expected outcomes. Points to note from this section include the following: In the 2010 survey, KiwiSaver members were asked why they had joined KiwiSaver. They could give more than one reason but also had to indicate the most important reason. Their responses provide some indication of the efficacy of these incentives. A similar question was asked in the 2013 survey about the early withdrawals for retirement.  The majority of KiwiSaver contributions are made by employees. For up to 56% of KiwiSaver members their deductions are being made at source from salary or wages.  Interestingly, the proportion of members who are auto-enrolled is comparatively low (39% in 2013) compared with those who opt-in (61%). This proportion has remained consistent. Given how many members earn salaries or wages it would seem that a relatively high proportion of people either have not changed jobs, or initially opted-out and have since changed their minds about membership.  There is evidence to suggest increased engagement with KiwiSaver accounts. The proportion of those choosing their own scheme has been gradually increasing from 49% in 2008 to 67% in 2013. In addition, in 2008, 38% of members were in a default scheme compared with 25% in 2013. Finally, there has been a marked increase in the number of members transferring schemes. In 2013, 136,167 members did so. This was 24% more than in 2012.  The default contribution rate has remained the most common over time. In 2010, when it was 2%, 41% of members were contributing it. In 2013, 58% of members who had KiwiSaver deductions taken from their salary and wages were contributing the default rate of 3%. Of note is that the initial 4% default rate was reportedly considered unaffordable by some and a challenge to joining. This highlights the importance of the default rate and setting it at a rate that enables sufficient savings without being so high that many perceive it as unaffordable.  There is a large group of KiwiSaver members who are not making contributions in 2013 (38% of members). This is larger than can be accounted for by contributions holidays, particularly when one considers that many on holidays do make voluntary contributions.  It appears the auto-enrol/opt-out feature has become increasingly successful in promoting membership. In 2013, for the first time, the cumulative number who had opted-out had decreased after four years of rises. It should be noted that the percentage change had been decreasing over that time.  The first group able to withdraw funds from KiwiSaver had limited funds available to them due to the short time they had been members. How Findings from the 2010 survey include the following:  The most commonly reported reasons for joining were the Crown and employer contributions. Combined, these were mentioned by 77% of respondents.  A similar proportion (75%) reported that KiwiSaver was a means of saving for retirement.  Saving for retirement was the most commonly reported ”most important reason”; designated as such by 46% of respondents compared with 29% of respondents who said the contributions were. What is important to note here is that most respondents reported joining KiwiSaver because it was a good way to save. The incentives played a part in their decisions but were not the most important reason they joined. A word of caution though—it is not possible to determine the extent to which the incentives are part of why KiwiSaver is a good way to save. There could well be an interaction effect and the incentives are the primary reason. In the 2010 survey, the main reasons given for not joining KiwiSaver were:  affordability (32% of non-member respondents)  other investment opportunities being better (30%)  they had not got around to it (28%). Respondents to the 2013 survey were also asked why they had joined KiwiSaver. In considering these responses it is important to remember the very limited demographic of this group, particularly in reference to any comparison with the 2010 survey results. The reasons they gave were as follows:  47% reported the contributions were the main reason for joining  40% said KiwiSaver was a good and easy way to save  31% said it was a way to secure income for retirement. 36 Prepared by: National Research and Evaluation Unit they withdrew/intended to withdraw their funds (lump sums or regular payments) was strongly correlated with the value of their account. Those with smaller balances were more likely to have withdrawn all their funds.  The rate of increase of those withdrawing their funds for reasons other than retirement has slowed markedly in 2012/13 compared with 2010/11.  Over the first five years of KiwiSaver, employers contributed $3.5 billion to KiwiSaver accounts. In the year to June 2013 they contributed $833 million.  In the 2012 year, $455 million in member tax credits was paid to 1.232,453 members. Less than half these members received the full tax credit.  For those earning solely wages or salary the median contribution in 2013 was $859; for those earning only non-salary or wage income the median was $1040.  Administrative data shows that the number of members on long holidays (five years) continues to grow while the number on shorter holidays has decreased. There also appears to be a decrease in the number of members commencing a holiday.  The cost of the home ownership features is likely to increase as more members begin to access them. The proportion of residential sales drawing on FDS support increased from 0.1% in July 2010 to 5.3% in December 2012. While the KiwiSaver scheme has promoted KiwiSaver membership, evidence reported subsequently suggests it has been less effective in promoting additional savings. There are also considerable numbers who are on contributions holidays or not contributing for other reasons. 37 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report Section 7. Individuals’ savings habits and asset accumulation The additionality research repeated the questions from the 2010 survey with some minor modifications to clarify responses. The 2011 TWP also undertook a separate analysis of the 2010 survey data. 7.1 Introduction This section looks at the fourth of the evaluation objectives: In both surveys, members were asked to distribute 10 points across various categories, some of which were related to saving and debt reduction, while others related to consumption. to determine the impact of KiwiSaver on individual’s savings habits and asset accumulation. Data for this section have been drawn from analyses of:  the 2010 survey of individuals (the 2010 survey)  analysis of the 2010 survey data in the 2011 TWP  the SoFIE–Inland Revenue linked data set (the 21 linked data set) in the 2014 TWP For savings to be deemed additional they had to be related to consumption only. Substitution was defined as any contributions sourced from any other form of savings including debt reduction. 20 19 F 7.2.1 Additionality in 2010 20F  the 2013 KiwiSaver early retirement withdrawal survey (individuals) (the 2013 withdrawal survey)  KiwiSaver early retirement withdrawal research with KiwiSaver providers (provider’s research)  the 2014 additionality and substitution research (additionality research). As Table 16 shows, in 2010 KiwiSaver member respondents reported that 36% (3.64 points) of their KiwiSaver contributions would have been spent in the absence of KiwiSaver. This is the amount of savings viewed as additional with 64% (6.36 points) seen as substitution. 7.2 Additional savings and retirement savings substitution KiwiSaver is intended to foster additional savings amongst a specific group of individuals, namely those who are not in a position to enjoy standards of retirement similar to those in pre-retirement. However, subsidised schemes are also likely to promote substitution,ie, the diversion of funds from other forms of saving that do not offer the same level of subsidy or incentives. The question of additionality was examined in both the 2010 survey and in a 2014 survey of KiwiSaver members using flow measures of savings. (This was supplemented with an examination of KiwiSaver on stock measures of savings as discussed in section 7.5 below.) 20 Law, D., Meehan, L. & Scobie, G., (2011) KiwiSaver: An Initial Evaluation of the Impact on Retirement Saving, Wellington, The Treasury, Treasury Working Paper 11/04 21 Law, D. & Scobie, G. (2014). KiwiSaver and the Accumulation of Net Wealth, Wellington, The Treasury, Treasury Working Paper 38 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report Table 16: Alternative use of KiwiSaver contributions (average points allocated) Overall Home owners Not home owners Spent on consumption 3.64 2.94 4.44 Daily activities and normal outgoings 3.58 2.87 4.36 Other 0.07 0.07 0.08 Would have been saved or used to reduce debt 6.36 7.06 5.56 Superannuation scheme 0.85 0.96 0.73 Other savings or investment for retirement 1.85 2.20 1.46 Saving or investment other than for retirement 1.29 0.89 1.73 Pay off mortgage or other debt 2.36 3.01 1.64 Source: Additionality and substitution: A national survey of 506 KiwiSaver members (to be published) As part of the analysis, the impact of homeownership was also considered. Respondents who owned their home were found to have allocated around 15% more of their KiwiSaver contributions to other forms of saving, or paying down debt than those who did not own a home. This difference was not solely due to mortgage repayment. Homeowners would have also allocated more of their contributions to both superannuation schemes and other savings or investments for retirement than non-homeowners.  Over 40% of homeowners would have saved all their contributions compared to only 20% of nonhome owners.  females also tended to spend more than males (7 percentage points).  with household incomes of $30,000 or less  aged under 50  with only a vague idea or no idea of their likely retirement income. This supports the notion that the value of assets and savings could drive knowledge and awareness, which in turn could promote membership and savings. This would explain why homeowners were more likely to have diverted other savings and/or debt reduction into KiwiSaver—they are already focused on some form of asset accumulation. Key findings include the following: From the total sample of respondents, 47% had a score of 8 representing high levels of substitution. those in part-time employment tended to spend more of their contributions than those working full time (12 percentage points) In the original analysis of the 2010 survey, it was reported that 45% of respondents were not likely to have saved for their retirement in the absence of KiwiSaver. This group contained relatively high proportions of individuals: In addition, respondents were assigned a score of between 0 and 10 representing the sum of the points they allocated to the saving and debt reduction categories. Essentially this score represented the extent to which their KiwiSaver contributions were substitutes for other forms of saving—the higher the score, the greater the substitution.   7.2.2 Additionality in 2014 compared with 2010 These scores were then used to develop a regression model to consider the impact of a wider range of factors on the reported levels of substitution. The results indicated that:   In 2014 there was no additional analysis of the survey data collected. In Table 17 the results from the 2014 survey are directly compared with those in 2010 as reported in the original survey report. those owning a home would have saved 12 percentage points more than non-homeowners in the absence of KiwiSaver Of note:  respondents with higher levels of education would have saved 4 percentage points more of their contributions for each additional year of education 39 Prepared by: National Research and Evaluation Unit The level of additionality, defined as savings that would otherwise have been spent on consumption, has decreased from 36% to 31%. The source of KiwiSaver contributions appears to have been increasingly derived from funds that would otherwise have been used for paying off mortgages or other debt (24% in 2010, compared with 32% in 2014). Table 17: How KiwiSaver contributions would be spent in the absence of KiwiSaver (percentage) 2010 % 2014 % Spend on daily activities and normal outgoings 36 31 Pay off mortgage and other debt 24 32 Other savings investment for retirement 18 15 Superannuation 9 8 Savings or investment for something other than retirement 13 14 Source: Ibid In 2014, consideration was given to how additionality varied by income and age (Table 18). The proportion of KiwiSaver funds that are considered additional were higher for those on lower incomes and for those who were younger. This is not surprising as they would be the least likely to already have other savings and investments. They will also have significant working lives to save and as such their balances on retirement are likely to be greater than those currently being reported. Table 18: Proportion of funds that are additional by income and age Additional % Personal income Up to and including $30,000 39 $30,001 to $60,000 31 Over $60,000 17 Age 18–34 35 35 –54 30 55–64 29 Source: Ibid 40 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report 7.3 Attitudes and behaviours towards retirement Respondents to the 2010 survey were asked how much financial planning they had conducted for their retirement. Key findings were as follows: half of the respondents said they had planned “a lot” (20%) or a “fair amount” (30%)  a further 32% said they had planned “a little”  17% had “not planned at all”. 23% had a vague idea  41% had no idea or had not thought about financial planning at all. homeowners  high-income groups  those with high net worth.  33% expected it to be less than that (0–19 years)  17% expected it to be longer (30+ years)  KiwiSaver members did not differ from nonmembers. 7.4.1 Sources of income All respondents to the 2010 survey were asked what sources of income they expected to use in retirement and what the main source would be. Appendix 9 summarises the findings from this question and compares the responses of KiwiSaver members and nonmembers. The main income sources for KiwiSaver households are: Further analysis undertaken by the Treasury (Law, et al, 2011) of these data showed two factors significantly reduced the likelihood of a respondent having undertaken financial planning. They found that those who reported poorer health and those who expected NZ Super to be their main source of income were less likely to engage in financial planning for their retirement. This is likely to be linked to perceptions of the value of doing so. 1. NZ Super (44%) 2. KiwiSaver (25%) 3. income from the sale or rent of property (12%) 4. stocks and shares or savings accounts (11%) 5. superannuation or work-based savings schemes (11%) 6. paid work (9%). The main income sources for non-KiwiSaver households are: 7.3.2 Expected standard of living in retirement All respondents to the 2010 survey were also asked how they thought their standard of living would change after retirement: 53% expected it to stay the same 41 Prepared by: National Research and Evaluation Unit the most common length of time was 20–29 years (28%) 7.4 Income in retirement There is obviously a relationship between savings and assets value and the amount of planning and knowledge. The direction of this relationship is not known.   The length of time of retirement did not influence other variables. However, those who had thought through their likely income in retirement were more likely to expect to retire before 65. Membership of KiwiSaver or not did not make a significant difference to either the extent of financial planning or their level of knowledge of their expected retirement income. Those who had thought about financial planning either “a lot” or “fair amount”, and those with higher levels of knowledge, contained high proportions of:  18% expected it to increase. All respondents to the 2010 survey were asked about the length of time they expected to be retired for. The length of time was calculated by asking respondents the age they expected to live to and when they expected to retire. The difference between these was the length of time they would be retired for. Key findings included the following: 36% had a good (8%) or reasonable (28%) idea of their retirement income   7.3.3 Expected length of retirement They were also asked what level of knowledge they had about their (and their partner’s) expected retirement income. Overall their reported levels of knowledge were not high:  27% expected it to decrease However, 33% of KiwiSaver members surveyed reported that their standard of living would decrease in retirement. This was significantly higher than the equivalent proportion for non-members (23%). 7.3.1 Financial planning for retirement   1. NZ Super (38%) 2. income from the sale or rent of property (22%) 3. stocks and shares or savings accounts (14%) 4. superannuation or work-based savings schemes (10%) 5. paid work (9%) 6. selling a business (7%). non-home owners  those under 50  those with a combined income under $50,000. In the 2013 survey of early withdrawers it was found that only 16% of respondents did not have other savings and investments and 42% had savings and investments of more than $100,000 (Table 19). Those expecting to use KiwiSaver as their main source of income contained relatively high proportions of:   those with a low net worth Table 19: Percentage of all respondents by value of other savings and investments Total % (=1000) None 16 Less than 20,000 12 $20,001–$50,000 12 $50,001–$100,000 12 More than $100,000 42 Source: KiwiSaver early retirement withdrawal survey. Results of a telephone survey of KiwiSaver members eligible to withdraw savings for retirement purposes http://www.ird.govt.nz/resources/d/a/da38fa80400b580d8174ef5d802abedf/ks-report-withdrawal-withmembers.pdf 37TU U37T  The more KiwiSaver savings a member had, the more likely they were to have other savings and investments (Table 20). Of particular note:  Members with more than $30,000 in KiwiSaver are much more likely to have more than $100,000 in other savings and investments. Members with up to $10,000 in KiwiSaver are much more likely to have less than $20,000 in other savings and investments. Table 20: Value of other savings and investments by value of KiwiSaver savings (percentage) Other savings (non-KiwiSaver) KiwiSaver values Less than 20,000 $20,001– $100,000 More than $100,000 Less than $5,000 62 14 22 $5,000–$10,000 38 24 31 $10,001–$15,000 23 24 46 $15,001–$20,000 24 31 40 $20,001–$25,000 25 22 51 $25,001–$30,000 40 20 38 More than $30,000 22 18 60 Source: Ibid 42 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report The respondents to this survey would have been close to retirement age when they joined KiwiSaver so it is, perhaps, not surprising they already had some other form of savings or investments. At the same time it underlines the fact that some are clearly planning for retirement even in the absence of KiwiSaver. 7.4.2 The perceived adequacy of retirement income The adequacy of the expected income in retirement of respondents to the 2010 survey was analysed in the 22 2011 TWP. The findings from this analysis (Table 21), for any expected shortfall or excess in retirement income, showed that: P21F P  78% of all respondents reported an expected excess retirement income for meeting needs  50% of all respondents reported an expected excess for being comfortable in retirement. For those reporting shortfalls, the median shortfall based on meeting basic needs was $6,800 while it was $12,000 for being comfortable. For those reporting an excess, the median excess was $10,800 based on basic needs compared with $5,200 for being comfortable. 22 Law, D., Meehan, L. & Scobie, G., (2011), KiwiSaver: An Initial Evaluation of the Impact on Retirement Saving. Wellington, The Treasury, Treasury Working Paper 11/04 43 Prepared by: National Research & Evaluation Unit KiwiSaver evaluation: Final summary report Table 21: Extent and size of any shortfall or excess in expected retirement incomes Basic needs Those reporting a shortfall Those reporting an excess Being comfortable Share of total (%) 22 50 Mean shortfall ($) –12,600 –15,100 Median shortfall ($) –6,800 –12,000 Average income ($) 39,400 45,700 Share of total (%) 78 50 Mean excess ($) +10,800 +5,200 Median excess ($) +6,000 0 Average income ($) $54,500 $56,400 Source: KiwiSaver: An initial evaluation of the impact on retirement saving http://www.treasury.govt.nz/publications/research-policy/wp/2011/11-04/twp11-04.pdf 37TU U37T Of the 1,000 respondents to the 2013 survey of early withdrawals for retirement, 73% believed their retirement income was either adequate or more than adequate. This is very similar to the results from the 2010 survey. KiwiSaver members accumulated less wealth over the period compared to non-KiwiSaver members). This potentially paints a bleaker picture than the flow-based additionality work described in section 7.2, which concluded that KiwiSaver did result in additional savings. A similar proportion of respondents had other savings and investments to rely on beyond KiwiSaver. Those with less in other savings and investments, single people and renters, were the groups most likely to report their savings were inadequate to some extent, for meeting their living costs in the immediate future and throughout retirement. At the same time, the Treasury paper points out that there is reason for caution in interpreting the findings of their study because of significant measurement error in key variables and wide change in net wealth. Efforts to correct for these are only likely to be partially successful. Also data is available for only a relatively short time. An 23 independent review of the study also discussed some potential limitations and recommended caution in drawing the conclusion that KiwiSaver had had a negative impact on the accumulation of net wealth. P22F 7.5 The accumulation of net wealth 7.6 Using KiwiSaver in retirement The SoFIE–Inland Revenue data set was used to determine the extent to which KiwiSaver has resulted in greater accumulations of net wealth amongst its members. The final sample size employed in the study was approximately 10,000 individuals, of whom 39% were KiwiSaver members. The 2013 early withdrawal survey was the first opportunity to consider the use of KiwiSaver funds in retirement. This is because they were the first cohort of members who were ready to retire. Again though, caution is needed in generalising these findings beyond this group. Many had other sources of incomes and their balances in KiwiSaver were not large. Two approaches were employed in the analysis. The first was to compare wealth changes over the period between KiwiSaver and non-KiwiSaver members (known as difference in differences analysis). The second was various panel regression techniques, which allow for other factors that might affect changes in wealth to be taken into account. One of the questions asked was how they had used, or intended to use their KiwiSaver funds. Respondents could give more than one answer (Table 22). Key findings include the following: The overall conclusion of this work is that it is not possible to reject the hypothesis that KiwiSaver has had no effect on net wealth accumulation (indeed, the only regression result with statistical power indicates that 23 Guest, R, MacDonald, K, Drew, M & Bianchi, R (2014). KiwiSaver Conclusions Report. Griffith University: Australia, p 14 44 Prepared by: National Research & Evaluation Unit P  Over half (58%) plan to put the money into other savings or investments.  Half (50%) intended to spend or had already spent it.  Of those who had withdrawn all their KiwiSaver funds most were likely to reinvest it (69%). Those who had made a partial withdrawal were most likely to have spent or be intending to spend the money (67%). Table 22: Percentage of respondents reporting different uses for funds by proportion withdrawn Total (990) Withdrawal of KiwiSaver All (446) Some (61) None (481) Put into other savings or investments 58 69 30 56 Spent/spend the money 50 40 67 54 Use(d) to pay off mortgage or debt 15 18 34 11 Other 4 3 6 4 Source: KiwiSaver early retirement withdrawal survey. Results of a telephone survey of KiwiSaver members eligible to withdraw savings for retirement purposes http://www.ird.govt.nz/resources/d/a/da38fa80400b580d8174ef5d802abedf/ks-report-withdrawal-withmembers.pdf 37TU U37T Of note is that 58% reported they were going to put money into other savings or investments. The main reasons for doing so were related to a perception of better return or security. For others, a driver was to improve ease of access to their funds. to use the money to pay off debt and were least likely to put the money into other savings or investments. The value of other savings or investments they had did influence their use of their KiwiSaver funds. It is likely this is because of the total retirement package including KiwiSaver, rather than these funds alone. As is shown in Table 23: Those who reported spending or intending to spend their funds (50%) were asked about their expenditure. The most commonly mentioned expenditure was holidays (61%). This was followed by home improvements or a new home (37%), major purchases such as cars or boats (32%) and health and medical expenses (29%). Only 22% envisaged spending their savings on everyday living expenses such as food and bills.  the more money saved or invested the more likely they were to reinvest their KiwiSaver funds. It could be they saw KiwiSaver as a short-term, relatively high return investment given the other contributions  those with no other investments were the most likely to pay off their mortgage or other debt. The amount respondents had saved in KiwiSaver did not seem to affect their intentions with one exception. Those with less than $5,000 were the group most likely Table 23: Percentage of respondents reporting different uses for funds by value of other savings and investments None (157) Below 20,000 (118) 20,001– 50,000 (122) 50,001– 100,000 (113) Over 100,000 (422) Spend the money 51 60 50 50 44 Pay off mortgage or debt 40 19 17 11 6 Other savings or investments 35 54 58 61 69 Source: Ibid 45 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report  7.7 Summary of key points This section of the report considered the influence of KiwiSaver on individual’s savings habits and asset accumulation. Key findings from this section include:  In 2010, using a flow measure of savings it was estimated that 36% of KiwiSaver contributions represented additional savings, with 64% substitution from other forms of savings. In 2013, the estimated level of additionality had fallen to 31%.  Half of the respondents to the 2010 survey reported they had planned either “a lot” or “a fair amount” for their retirement. 36% said they had a “good” or “reasonable” idea of their retirement income. Those who had poor health and those who expected NZ Super to be their main source of income were less likely to engage in financial planning. Higher levels of knowledge and more planning were associated with higher incomes and higher net worth.  Of all respondents to the 2010 survey 53% thought their standard of living in retirement would stay the same, 27% expected it to decrease and 18% expected it to increase. KiwiSaver members were significantly more likely to report their standard of living would decrease than non-members.  The main retirement income source for both KiwiSaver households (44%) and non-KiwiSaver households (38%) was NZ Super. The next most commonly reported main source for KiwiSaver households was KiwiSaver (25%) compared with income from the sale or rent of property for nonKiwiSaver households (22%).  The analysis in the 2011 TWP of the 2010 survey showed that 78% of respondents had an expected excess retirement income for meeting needs; 50% had an excess for being comfortable. This analysis found no evidence that KiwiSaver membership was associated with any expected shortfall in retirement.  In the 2013 withdrawal survey of early retirees, 73% believed that their retirement income was either adequate or more than adequate. Only 16% did not have other savings and investments. For 42% of respondents these exceeded $100,000. The more savings they had in KiwiSaver the more they were likely to also have other savings and investments. This suggests KiwiSaver was a means of broadening their portfolio, although the financial incentives could have equally played a role. The value of their other savings and investments influenced their use of their KiwiSaver funds. The more money they had saved or invested the more likely they were to reinvest their KiwiSaver money once they were able to withdraw it. 46 Prepared by: National Research & Evaluation Unit The analysis in the 2014 TWP of the linked data set, with regard to asset accumulation, found no evidence of KiwiSaver leading to the accumulation of net wealth. Indeed, panel regressions found a negative effect for KiwiSaver membership. This was only statistically significant in one of six regressions. KiwiSaver evaluation: Final summary report Section 8. Superannuation markets and the financial sector 8.1 Introduction desktop research. In addition Inland Revenue administrative data are reported. This section of the report considers the final objective of the evaluation: 8.2 KiwiSaver scheme fund balances To determine the impact of KiwiSaver on superannuation markets and the financial sector As of 31 March 2013, scheme providers held $16.6 billion in KiwiSaver schemes—an estimated 19% of the superannuation and managed funds market. This had grown from an initial $701 million in 2008 (Table 24). The total comprised member contributions, employer contributions, Crown payments and investment returns. The results of a study undertaken in 2010 by the Ministry of Economic Development (now MBIE) were used to consider this question. Data were collected through surveys and interviews with providers, and Table 24: Year-on-year increases in assets in KiwiSaver schemes 2008 2009 2010 2011 2012 2013 701 2,660 5,851 9,187 12,735 16,565 Number increase ($m) 1,959 3,191 3,336 3,548 3,830 Percentage increase 279% 120% 57% 39% 30% Cumulative total ($m) Source: Financial Markets Authority KiwiSaver reports http://www.fma.govt.nz/help-mecomply/kiwisaver/monitoring-and-surveillance/kiwisaver-reports/ 37 TU U37T a medium/long-term growth opportunity. Building the necessary infrastructure to support the product had made it marginally profitable in most cases. 8.3 Supply side impacts Supply side impacts have been evaluated twice with the 24 latest report completed in July 2010 . The specific focus of the 2010 evaluation was the effect of KiwiSaver on the superannuation and managed funds market and the wider financial services sector. This evaluation used two sets of indicators: P23F  The global financial crisis and resulting volatility of investment returns had not impeded KiwiSaver’s growth significantly, but had reduced its profitability for providers because the negative investment returns reduced the funds under management. This in turn, drives the level of income managers receive.  The number of providers was currently stable but more consolidation occurring would be consistent with a healthy industry.  To date the funds flow had had only a small positive impact on New Zealand capital markets.  KiwiSaver funds had a marginally higher weighting to domestic assets than other superannuation schemes or managed funds. P  the firm dynamics indicators: to show the health of the KiwiSaver market  the market dynamics indicators: to show how the KiwiSaver market relates to other superannuation and managed funds more generally. The conclusions drawn from the study were as follows:  24 KiwiSaver was providing a stimulus to the New Zealand financial sector. Providers saw KiwiSaver as MBIE report on KiwiSaver supply side evaluation 47 Prepared by: National Research & Evaluation Unit 8.4 Summary of key points In summary, the supply side study of the period up to 2010 concluded that KiwiSaver was providing a stimulus to the New Zealand financial sector and that it was a medium to long-term growth opportunity for the sector. 25 Since then more recent analysis has shown that a significant proportion of KiwiSaver funds are invested overseas and relatively conservatively. The impact on the capital markets remains small. P24F P 25 Retirement Policy and Research Centre. (2014). Observations on Reserve Bank’s Household Financial Assets 2003–2013. Pension Briefing 2014-1. University of Auckland Business School 48 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report Section 9. Value for money of KiwiSaver  9.1 Introduction The 2011 TWP reported that the strong growth in KiwiSaver membership and the private benefits from increased retirement savings had come at a significant public cost. Given the contributions were being funded through increases in public debt they argued that the impact of the scheme needed to be evaluated in light of this cost—essentially a consideration of the value for money KiwiSaver represents. 9.3 The costs to the Crown Direct KiwiSaver costs to the Crown come from the $1,000 kick-start to new members and the annual member tax credit. Until 2009 there was also an employer tax credit and a fee subsidy of $40 per member. This section considers the broader picture of the value for money of KiwiSaver. It looks at Crown contributions, the level of additionality achieved, the extent to which the target market was met, the impact on national savings and the cumulative costs and benefits. In 2013, KiwiSaver cost the Crown $677 million in contribution payments to members. This was a decrease from $1.5 billion in 2012 when the member tax credit was halved. Overall, in the first five years of KiwiSaver, payments to members and employers have cost the Crown $5.3 billion. Evidence for this section has come from a range of data sources including:  administrative data  the 2010 survey  the 2013 study of additionality and substitution  the SoFIE–Inland Revenue linked data set. As a proportion of the total value of funds passed to providers for investment in members’ accounts, the contribution from the Crown is declining over time. 9.4 Effectiveness of KiwiSaver in meeting the target market 9.2 Considering value for money The policy objective of KiwiSaver is to encourage a savings habit and asset accumulation amongst individuals who may not be in a position to enjoy standards of living in retirement similar to those in preretirement. This suggests a “target population”, which the scheme is intended to help. The extent to which KiwiSaver had reached this market was considered in the 2011 TWP, which used data from the 2010 survey. It is difficult to accurately define a “target group” for KiwiSaver having regard to the legislative purpose and applicable to the data set available. Other approaches to defining a “target group” and a more sophisticated data set could result in very different value for money estimates. A study of the value for money of KiwiSaver was completed in 2014. A relatively simple approach was taken in this study, combining direct costs with the results of earlier studies on the level of additionality and leakage of KiwiSaver benefits to those outside its target. There are uncertainties regarding the influence of some factors on KiwiSaver although these are unlikely to have a substantial impact on the value for money conclusions. The factors driving this uncertainty included:  the timeframe since the introduction of KiwiSaver  the impact of KiwiSaver on employment costs  the changes made to KiwiSaver over the time period of the evaluation The 2011 TWP, which analysed the 2010 survey data, used a specific definition for the target population, which was defined by two characteristics:  49 Prepared by: National Research and Evaluation Unit the impact of economic cycles and the markedly volatile period within which KiwiSaver has been implemented. An individual expected to have income in retirement that was less than what they self- identified to meet their basic needs or to live comfortably.  9.5 Cumulative costs and benefits If an individual were to join KiwiSaver at least 30% of their contributions would have to represent new savings. How the total costs of KiwiSaver relate to the contributions that have been made was considered through an analysis of Inland Revenue administration data combined with the results of studies on the extent KiwiSaver contributions represent new savings (additionality) and the extent KiwiSaver benefits have accrued to those outside the target population (leakage). This allowed a comparison between the additional savings of those in the KiwiSaver target population and the costs incurred by government. Using these criteria the 2011 TWP concluded that very few KiwiSaver members were part of the target population definition. If the target population was taken to be those expecting insufficient income in retirement to live comfortably, the study found that 22% of KiwiSaver members fit within the target. This fell to only 7% if the target were instead defined as those who did not expect to be able to meet their basic needs in retirement. The cost to the government was defined as: This implied that the ongoing cost of the scheme for each target member, based on living comfortably, was around $4,000 per year. The cost based on those in need was even higher at $13,000 per year. The study also found that at March 2010, 47% (based on living comfortably) or 33% (based on meeting basic needs) of the target population had joined KiwiSaver.  the cost of all incentives available to KiwiSaver members over the review period  the administration costs faced by Inland Revenue and Housing New Zealand (now MBIE) in administering KiwiSaver. This analysis (Table 25) showed that the government is spending $1 to get $0.38 of individual savings, factoring in additionality and leakage under a definition of KiwiSaver’s target being those individuals expecting to be unable to live comfortably in retirement. Table 25: Overall costs and benefits 2008/09 2010/11 2012/13 $0.20 $0.22 $0.38 Percentage growth in membership +33.4% +17.8% +9.1% Percentage growth in members’ mean annual contribution –6.2% –0.8% +9.5% Additional savings member contributions made by KiwiSaver target per $ of government cost Source: Government retirement savings incentives in New Zealand: do they provide value for money? A value for money analysis of the KiwiSaver programme from its inception to 30 June 2013 (to be published)  There are three main drivers for the improvement in the cost ratio in 2012/13 when for every government dollar, additional savings among the target rose to $0.38. These were:  the declining growth rate of new members significantly reduced the cost of the kick-start incentive from $471 million in 2008/2009 to $217 million in 2012/2013  the reduced member tax credit from July 2011 reduced that cost from $664 million in 2010/11 to $455 million in 2012/2013 It is too soon to tell whether this slight improvement in value for money will be sustained over time. Factors to be considered include whether the growth in membership continues to slow, impacting on the cost of the kick-start. Also to be considered is whether the firsthome deposit subsidy becomes a significant element of cost going forward. It is probably too early to assume 50 Prepared by: National Research and Evaluation Unit in 2012/2013 members’ mean contributions increased faster than membership. The increase in contribution rates from 2% to 3% will have been an important factor in driving this. that the policy framework is stable, ie, that further changes will not be made to KiwiSaver. The analysis of the distribution of incentives relative to contributions across three income bands and four age groups shows that the actual incidence of incentives is not disproportionately favouring the highest income group. The assumptions on which this analysis was based are explained in the detailed report. The additionality figures used are those from the 2010 and 2014 surveys of individuals reported earlier. 9.6 Summary The following are the key findings from this section of the report:  In 2013, KiwiSaver cost the Crown $677 million in contribution payments to members. This was a decrease from $1.5 billion in 2012 when the member tax credit was halved. Overall, in the first five years payments to members and employers have cost the Crown $5.3 billion.  In 2010, the estimated level of additionality was 36%, this decreased to 31% in 2013. In 2010, homeowners were found to have allocated around 15% more of their KiwiSaver contributions to other forms of saving or paying off debt than those who did not own a home.  In 2013, increases in KiwiSaver contributions appear to have been increasingly related to paying off mortgages or other debt.  The 2011 Treasury Working Paper (Law, et al, 2011) reported that in 2010 the KiwiSaver scheme appeared to be reaching about only one third of the target population with leakage estimated to be as high as 93%. This implied that the ongoing cost of the scheme for each estimated target member was around $4,000 per year.  A costs and benefit analysis shows that for the period 2007/08 to 2013/13, the additional savings amongst the estimated target group for each $ of government spending ranged from $0.20 to $0.38 as the level of government contributions dropped with less new enrolments and policy changes. The analysis also used a narrower target group which produced a lower value for money. Given the importance of the assumptions used, this analysis could produce different results (better value for money) if a wider definition of the target group were used. 51 Prepared by: National Research and Evaluation Unit  The costs to the Crown are reducing over time as membership growth slows. The reduction to the member tax credit and the increase in the default rate have also had an effect.  The actual incidence of incentives is not disproportionately favouring the highest income group. The trend in the mean subsidy paid within income bands is evening out. Further, the proportion of subsidies paid to the highest income group is less than 25% of the total while they contribute more than 45% of the savings. KiwiSaver evaluation: Final summary report Section 10. Conclusions Essentially, KiwiSaver needs to be measured against its policy objectives, which were:  to encourage a long-term savings habit and asset accumulation by individuals who are not in a position to enjoy standards of retirement similar to those in pre-retirement  to increase individual’s well-being and financial independence, particularly in retirement and to provide retirement benefits. It would also appear that KiwiSaver has not been successful in improving the accumulation of net wealth for its members. Rather, over the period under review (up to 2010) KiwiSaver appears to have had a negative impact when compared with non-members. Again this was over a relatively short time period and it is possible this will change over time. It could be, for example, that the default schemes are too conservative for short-term wealth accumulation, but will be more secure in the long term. There are also concerns about the extent to which KiwiSaver has been effective in meeting the target market. Based on the perceptions of individuals with regard to their retirement income and needs, there appears to have been significant leakage, ie, KiwiSaver is reaching a large number of individuals outside the estimated target group. Further, it appears to have reached only about one-third of the target market, as defined for this evaluation. This report has drawn together seven years of evaluative activity under the key objectives of the evaluation. It also includes a 2014 Value for Money study. Based on the evidence collected across the evaluation the success of KiwiSaver in achieving its policy goals in the shortterm appears to be marginal at best. There is an argument that KiwiSaver is promoting saving for retirement, ie, more individuals now belong to a retirement scheme from which it is very difficult to withdraw funds prior to retirement. Further, they are locked into this scheme. As such, one could argue the features of KiwiSaver designed to attract members have worked. The proportion of new members who have opted-in has continued to remain high (over 60%), suggesting KiwiSaver remains an attractive option. Further, fewer individuals are opting out. Finally, the value for money study showed that KiwiSaver is currently expensive in terms of the cumulative costs and benefits. These findings need to be considered in light of the limitations and concerns expressed throughout this report and in the feedback from the independent reviewers. They are a point in time analysis of the early effectiveness of KiwiSaver. The extent to which they indicate its success in five or 10 years cannot be readily extrapolated from these data given the numerous factors that can influence these findings. Whether this leads to long-term benefits and the impact of KiwiSaver on the retirement lives of the 18–25 year old members, who will have had 30 plus years to contribute, cannot be extrapolated from the experiences of the early retirees who had only five years of contributions. However, there is concern over the amount individuals are saving, in particular the extent to which this is additional saving. The level of additionality has remained relatively constant between 2010 and 2013 and suggests only about a third of KiwiSaver contributions represent new savings. There is a considerable proportion of members not contributing or on contributions holidays. Both would seem to be a way of individuals managing the locked-in features of KiwiSaver. The reasons for this behaviour would seem to be related to perceived affordability and major life events, which can alter priorities. 52 Prepared by: National Research & Evaluation Unit KiwiSaver evaluation: Final summary report Section 11. The KiwiSaver evaluation programme In this section the effectiveness of the KiwiSaver evaluation programme is considered along with recommendations for the future, should an evaluation be required. To date this has been a very comprehensive and extensive evaluation. Substantial amounts of data have been collated and analysed across many different activities. The richness and breadth of the story this evaluation tells should not be underestimated. Nor should its potential value to a number of agencies and for a number of different purposes. It has been praised internationally as the most comprehensive evaluation of 26 a retirement scheme undertaken. 11.1.2 Survey of Steering Group members The members of the Steering Group were asked to complete a short survey regarding the effectiveness and future of the KiwiSaver evaluation. There were nine respondents to the survey from across six agencies. A full report from that survey is provided in Appendix 10. Respondents were asked to indicate the level of effectiveness of the evaluation in meeting each of its objectives on a scale of 1–6. P25F 11.1 The effectiveness of the evaluation The effectiveness of the evaluation should be measured primarily against its ability to judge the success of the policy against its objectives. It should also be measured by its usefulness to key stakeholders or owners of the evaluation. The lowest reported mean level of effectiveness was 3.1 (slightly good), reported for determining the impact of KiwiSaver on superannuation markets and the financial sector.  The highest mean level of effectiveness was 4.5 (moderately/very good), given for understanding the scale and pattern of uptake of KiwiSaver. They were also asked to rate the evaluation against some general criteria: 11.1.1 Evaluating against policy objectives Its ability to achieve this first indicator has been discussed already. The evaluation has essentially met this indicator given the constraints of the time-period within which it was implemented. This is particularly true if one considers this to be a process and early outcomes evaluation.  The lowest rating of 3.7 was for the value for money of the information provided.  The highest rating of 5.1 (very good) was given for the overall model of a lead agency with others doing their own research. 11.2 The future of the evaluation programme Any programme of evaluation should be open to criticism. There has perhaps been a tendency to do discrete pieces of work rather than building up a systematic research programme, allowing for ready comparisons over time. However, there is certainly a rich and comprehensive evidence base regarding KiwiSaver across the first years of its implementation. Much of the potential usefulness of this work will depend on the extent to which this work is used as a baseline in the future, and to which the discrete pieces of work to date are built into a longitudinal research programme. In the survey the Steering Group was also asked about the future of the evaluation programme. The majority view was that Inland Revenue should continue to lead the work. However, there was little consensus about the evaluation model. Respondents were asked to rank five different options from most preferred (1) to least preferred (5). No one option dominated, with all five having an average rating of 3. This reflects the divergent perspectives of this group and one of the confounding factors for this evaluation.  26 Collard, S, & Moore, N (2009). Review of International Pension Reform. Research Report No 663. Department for Work and Pensions. Available from https://www.gov.uk/government/uploads/system/uploads/attachmen t_data/file/214434/rrep663.pdf 53 Prepared by: National Research & Evaluation Unit  The least preferred option (average of 3.3) was the development of another stand-alone comprehensive monitoring, research and evaluation framework covering a 10-year period.  3. Regular reviews of KiwiSaver, including some research and the use of administrative data, had an average rating of 2.6, and was the most preferred option. This does suggest an appetite for more work but not on a large, extended timeframe. The issue is that without a framework the potential value of research and monitoring is reduced. Discrete pieces of work do not have the same overall usefulness over time. Respondents were also asked what they thought the key areas for evaluation/research should be moving forward. Again there was a range of responses reflecting their different policy areas, such as:  4. revisiting the key areas of research related to outcomes for members and the value for money at a later stage to compare against established baselines  comparisons with international models  exploring the full range of decumulation issues  confirming the additionality assumptions  the impact of housing incentives on the housing market  implications for social equity  analysis of membership from an affordability perspective 5. Attitudes and behaviours  Are attitudes towards retirement savings changing over time? Repeat the knowledge, risk and engagement questions from the 2010 survey to enable a comparison?  To what extent do the views of employers impact on the decisions individuals make?  What impact has the increased marketing and information available had on the attitudes and behaviours of individuals? Contribution rates  What determines how much an individual contributes?  Is there an “ideal” contribution rate? How does this differ across different demographics? What is considered “fair and affordable”?  What are the spending priorities of individuals and how do these change over time?  Why are some members not contributing? When do they intend to do so? Why did they join in the first place? Longitudinal questions  Additionality and substitution  Meeting the target market  whether KiwiSaver targeting the right members  The cost to the Crown  what affects changes in contributions.  Accumulation of net wealth  Impact on other superannuation and retirement schemes (the supply side)  Impact on retirement income, behaviour and experiences across different cohorts over time. 11.3 Research ideas from the evaluation In completing this report a number of potential research areas were highlighted. Some of these have been listed here, grouped under thematic headings. 1. Membership  2. 11.4 Discussion / recommendations This work has occurred across the first seven years of KiwiSaver. This is a very short time period for an initiative aimed at changing long-term behaviours and improving well-being in retirement. Further, the policy framework of KiwiSaver has undergone substantial changes over this period. As such all findings need to be treated with caution and viewed as early outcomes only. Other limitations, as indicated at the start of this report include the difficulty of defining the target market, the difficulty of ascribing causality in a complex environment and the difficulty of measuring additionality. Why are individuals continuing to opt-in? What is driving their decision to actively join KiwiSaver? What has changed for them? Employers  How are they funding their contributions to KiwiSaver, eg, salary sacrifice, increased charges to customers?  What does the remuneration package look like for their employees?  What are the costs/benefits to employers of KiwiSaver? Are there any other costs outside of compliance? Do they see any benefits? These limitations notwithstanding, the evaluation provides a very detailed snapshot of the retirement attitudes, behaviours and expectations of individuals at this time. It also provides a rich data set regarding the 54 Prepared by: National Research and Evaluation Unit implementation of KiwiSaver, including trend data related to membership and contributions. The early outcomes analyses against the objectives of KiwiSaver are also of value in informing future policy decisions and as a benchmark for considering the future performance of KiwiSaver. Whether a further comprehensive evaluation is needed is a moot point. The evaluation has highlighted a number of potential research areas for the future (listed above). The findings from the Steering Group survey suggest there is an appetite to continue the work in some form. But evaluations cost money and this needs to be assessed against competing spending priorities. Should there be changes to the features of KiwiSaver (eg, contribution rates) there would be some value in considering the impact of these, using this body of work as a benchmark. There is also value in considering repeating the core pieces of work in five years time. These are the studies related to additionality, meeting the target market, accumulation of net wealth and cumulative cost-benefit. Changes in key behaviours and attitudes as well as the impact on retirement are also worth considering across different cohorts. This could be achieved through longitudinal research designs and very focused pieces of work. There would also be some value in developing an agreed framework for research in areas that can help inform both strategic and operational policy regarding the wellbeing of New Zealanders in retirement, the effectiveness of retirement policy and the attitudes and behaviours of New Zealanders. Such research need not be expensive if well managed and focused. If further research were considered a priority, the recommendation is that the Steering Group consider the development of a systematic research, evaluation and monitoring framework that includes discrete pieces of work building towards a second evaluation of the key policy objectives in approximately five years time. This should build on the current work and enable comparisons over time. 55 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report Appendix 1: KiwiSaver features and glossary See also Appendix 3 for changes to KiwiSaverfeatures since its introduction in 2007. Features Work-based scheme KiwiSaver is a work-based initiative meaning that information is provided by employers, and employee’s contributions are deducted directly from their pay. Eligibility requirements To be eligible to join KiwiSaver, individuals must be a New Zealand citizen (or entitled to live in New Zealand indefinitely), and personally present or normally personally present in New Zealand, and under the age of eligibility for NZ Super (currently 65). Default providers Those individuals who are automatically enrolled will be allocated to one of six government selected default schemes. These members have three months to choose their own scheme and if they do not they will be enrolled in the default scheme. Opt-in Individuals can elect to join KiwiSaver directly, either through their employer or by contacting a scheme provider. Automatic enrolment and optout Provided they are eligible for KiwiSaver, all new employees (ie, those starting a new job) will be automatically enrolled in KiwiSaver unless they meet one of the conditions for exemption. Individuals may choose to opt-out within eight weeks (but not within the first two weeks) of starting the new job. Scheme and contributions choice Originally, KiwiSaver members could choose from a range of superannuation schemes and providers. They could also elect to contribute either 4% or 8% of their gross income (changed). Savings locked in Savings cannot be withdrawn until the age of eligibility for NZ Super or until they have been a member for five years, whichever is the later. Self- and nonemployed Those who are self-employed or not-employed can opt-in to KiwiSaver and can determine their level of contributions. Employer tax credit An employer tax credit of up to a maximum of $20 per week for each employee to reimburse employers for their contributions to employees’ accounts (ceased). Incentives Kick-start payment Initial payment of $1,000 to an individual’s account upon joining. Member tax credit Members were entitled to a tax credit of up to $1,042.86 per year ($20 per week) from 1 July 2007. The tax credit is paid automatically to an individual's KiwiSaver account, based on the level of contribution made (changed). 56 Prepared by: National Research & Evaluation Unit Employer contributions From 1 April 2008, employers were required to match individuals’ contributions up to 1% from 1 April 2008, rising 1% a year to 4% from 1 April 2011 (changed). Contributions holiday Members are able to take a contributions holiday of between three months and five years after contributing for an initial 12-month period. Individuals can apply for a holiday within the first 12 months for reasons of financial hardship. First home deposit subsidy After three years of contributions, an individual can access a conditional grant towards buying a home, equal to $1,000 per year of contribution up to $5,000. First home withdrawal After being a member for three years, individuals can withdraw all or part of their savings to contribute to a deposit on a first home. Mortgage diversion facility After being a member for 12 months an individual can split their contributions between their KiwiSaver account and their mortgage payments (ceased). Fee subsidy Members entitled to a $40 provider fee subsidy each year (ceased). 57 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report Appendix 2: Research included in this synthesis For more detail on each of these research activities please refer to the research report available on Inland Revenue’s website or from the lead agency. Agency Topic Method Date Treasury Accumulation of net wealth SoFIE–Inland Revenue linked data set 2014 Inland Revenue Value for money of incentives Administrative and survey data 2014 Inland Revenue Additionality and substitution research Repeat of relevant questions from 2010 survey (506) 2014 Inland Revenue Value for money – qualitative 35 interviews providers (5), employers (15), members (10), non-members (5) 2014 MBIE (Internal) Uptake of first-home ownership package Analysis of administrative data 2013 Inland Revenue KiwiSaver early retirement withdrawal Telephone survey of 1,000 eligible members 2013 Inland Revenue KiwiSaver early retirement withdrawal Survey (n=25) and interviews (n=11) with scheme providers 2013 MBIE Use and experiences of the KiwiSaver home ownership package Survey of 402 people and analysis of available data 2012 MBIE (MED) Report on KiwiSaver Supply Side Evaluation Provider survey (24) and interviews (7); desk top 2010 Treasury An initial evaluation of the impact on retirement savings Analysis of the 2010 Colmar Brunton survey of individuals 2011 Inland Revenue Opting out and taking contributions holidays Administrative data 2011 Inland Revenue Survey of Individuals (members and non-members) Survey of 825 individuals 2010 Inland Revenue SME compliance costs Survey measuring 1,728 SMEs tax compliance costs 2009 2010 58 Prepared by: National Research & Evaluation Unit Agency Topic Method Date Inland Revenue Automatic enrolment process Qualitative research with employers (20 interviews) and employees (50 telephone interviews) 2007 Inland Revenue Implementation in the workplace Employer panel – interviews with 34 employers and 63 employees 2007 Inland Revenue Employer’s awareness, understandings and reactions to KiwiSaver Employer survey – 500 telephone interviews of PAYE registered employers 2007 Inland Revenue The engagement model Scheme provider interviews (n=18) 2007 59 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report Appendix 3: Changes to KiwiSaver features over time See also Appendix 1 to compare against original KiwiSaver features. Date change effective Feature affected Nature of change April 2013 Employee contributions Minimum employee contribution rate rose from 2% to 3% for all members 3% also became the default contribution rate for all new employees April 2013 Employer contributions Compulsory employer contributions rose from 2% to 3% April 2013 Disclosure rules KiwiSaver fund managers required to report performance and returns, fees and costs, assets and portfolio holdings, liquidity and liabilities, and key personnel along with any conflicts of interest in a standardised format April 2012 Employer superannuation contribution tax All employer contributions were made subject to this tax, with the removal of the exemption for contributions at 2% of the wage or salary of the employee. The tax is applied at a rate equivalent to an employee’s marginal tax rate July 2011 Member tax credit (MTC) The maximum annual MTC payment reduced from $1,042.86 to $521.43. The total contribution required to receive this stayed the same ($1,042.86) April 2009 Member contributions Minimum rate decreased from 4% to 2% 60 Prepared by: National Research & Evaluation Unit KiwiSaver evaluation: Final summary report Appendix 4: Membership profiles from linked data set as at September 2010 KiwiSaver members % Never been member % % difference Proportion of eligible sample % 18–24 years 60.9 39.1 21.8 10.5 25–34 years 42.6 57.4 –14.8 17.2 35–44 years 37.0 63.0 –25.9 22.3 45–55 years 38.2 61.8 –23.5 24.9 55–65 years 44.3 55.7 –11.4 25.2 Male 40.5 59.5 –19.0 48.8 Female 44.6 55.4 –10.7 51.3 NZ European 42.4 57.6 –15.1 71.0 Māori 41.9 58.1 –16.1 13.7 Pacific Island 46.5 53.5 –7.0 5.6 Asian 40.8 59.2 –18.3 7.8 Other ethnicity 50.4 49.6 0.8 1.9 No qualification 37.3 62.7 –25.4 15.5 School Certificate/NCEA 43.7 56.3 –12.6 26.1 Trade/Diploma 41.4 58.6 –17.1 36.7 University 47.8 52.2 –4.4 13.8 Postgraduate 46.0 54.0 –8.1 7.9 Single, no resident children 38.4 61.6 –23.1 11.2 Single, with resident children 41.5 58.5 –17.1 8.5 Partnered, no resident children 45.0 55.0 –10.1 24.1 Partnered, with resident children 40.5 59.6 –19.1 44.1 Partnered, with others 38.5 61.5 –23.0 0.7 Other household situation 13.6 9.6 4.0 11.3 Upper white collar 44.2 55.8 –11.6 48.6 Lower white collar 50.4 49.6 0.9 23.9 Upper blue collar 33.4 66.6 –33.2 14.6 Age group Gender Ethnicity Highest qualification Household type Occupation 61 Prepared by: National Research & Evaluation Unit KiwiSaver members % Never been member % % difference Proportion of eligible sample % Lower blue collar 45.5 54.5 –9.0 12.6 Unidentifiable 34.1 65.9 –31.8 0.3 Salary and wages 47.7 52.3 –4.7 63.7 Self-employed 31.8 68.2 –36.5 12.2 NZ Super 29.4 70.6 –41.1 1.7 Government benefits 35.2 64.9 –29.7 12.8 Interest and investment 36.2 63.8 –27.6 3.9 Other sources 42.3 57.7 –15.4 3.3 No income source 25.5 74.5 –49.0 2.5 Loss 0.5 1.4 –0.9 1.0 Zero Income 1.5 3.3 –1.7 2.5 $30,000 or less 36.2 34.0 2.2 35.0 Over $30,000 to $50,000 29.0 25.7 3.3 27.1 Over $50,000 to $80,000 21.0 22.2 –1.2 21.7 Over $80,000 to $120,000 7.3 7.9 –0.6 7.6 Over $120,000 4.5 5.6 –1.1 5.1 Main income source Employment income 62 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report Appendix 5: Demographic profiles by membership status and enrolment type Current members Non-members Autoenrolled % Opted-in through employer % Enrolled through a provider % Opted-out % Never been a member % 18–24 years 32.9 9.2 8.7 23.0 11.2 25–34 years 21.3 18.5 14.1 27.8 18.5 35–44 years 19.4 20.9 19.0 24.8 25.0 45–55 years 17.3 26.4 24.8 17.3 25.7 55–65 years 9.1 25.0 33.5 7.2 19.7 Male 47.9 48.7 43.0 46.5 49.9 Female 52.1 51.3 57.0 54.5 50.1 NZ European 65.6 76.8 71.6 70.0 70.5 Maori 18.5 10.8 11.0 15.9 14.3 Pacific Island 7.6 3.9 6.3 2.4 5.5 Asian 6.2 6.5 8.9 9.8 8.0 Other ethnicity 2.1 1.9 2.2 1.9 1.8 No qualification 13.6 12.0 14.9 10.5 16.3 School Certificate/NCEA 32.9 25.4 26.9 29.4 27.5 Trade/Diploma 35.3 35.2 35.5 38.1 36.6 University 13.2 16.7 14.1 16.6 12.5 Postgraduate 5.1 10.8 8.5 5.4 7.0 Single – live alone 12.0 8.7 8.9 8.1 10.9 Single parent 10.4 6.9 7.1 8.8 8.5 Couple children at home 44.5 43.2 42.1 47.9 46.9 Couple no children at home 14.9 26.5 28.9 20.1 20.9 Couple with others 1.0 1.2 0.5 0.2 0.9 Other household situation 17.3 13.6 12.4 14.9 11.9 38.1 56.4 51.9 45.8 47.5 Age band Gender Prioritised ethnicity Highest qualification Household type Occupation Upper white collar 63 Prepared by: National Research & Evaluation Unit Current members Non-members Autoenrolled % Opted-in through employer % Enrolled through a provider % Opted-out % Never been a member % Lower white collar 33.3 24.4 22.0 26.4 22.6 Upper blue collar 11.8 8.7 13.4 14.8 16.7 Lower blue collar 16.7 10.3 12.5 13.1 12.9 Unidentifiable 0.2 0.3 0.2 0.0 0.3 Salary and wages 76.5 89.4 54.9 82.1 61.3 Self-employed 4.6 3.4 16.6 3.4 13.4 Superannuation 0.2 0.0 1.2 0.0 0.8 Government benefits 12.2 2.3 14.1 9.0 14.1 Interest and investment 1.6 1.2 6.1 1.3 3.9 Other sources 2.8 2.1 5.2 2.9 3.1 No income source 2.1 1.6 2.0 1.3 3.3 Agriculture, forestry and fishing 5.0 3.9 7.2 6.6 8.9 Mining 0.7 0.2 0.0 0.2 0.3 Manufacturing 11.5 13.0 11.0 12.4 11.9 Electricity, gas, water and waste services 0.4 0.5 0.6 0.9 0.5 Construction 7.1 5.1 7.2 9.1 8.5 Wholesale Trade 5.6 7.8 4.5 5.4 5.6 Retail trade and accommodation 20.6 11.9 12.3 18.4 11.6 Transport, postal and warehousing 4.2 4.5 5.3 3.5 3.5 Communication services 1.4 1.2 1.6 1.4 1.6 Financial and insurance services 1.2 4.3 3.3 1.5 2.5 Rental, hiring and real estate services 14.5 13.5 15.4 14.2 14.0 Government administration 4.2 5.2 4.6 4.0 5.5 Education and training 5.7 8.3 9.2 3.0 10.2 Health care and social assistance 10.7 13.6 10.8 12.3 7.9 Arts, recreation and other services 7.0 6.9 6.8 7.3 7.6 Not elsewhere 0.0 0.0 0.1 0.0 0.0 Loss 0.3 0.1 0.5 0.4 10.0 Zero income 2.1 1.6 2.0 1.3 3.3 $30,000 or less 52.9 22.0 38.7 42.0 36.3 Over $30,000 to $50,000 25.7 34.7 24.6 29.9 26.2 Main income source Industry sectors Personal income 64 Prepared by: National Research and Evaluation Unit Current members Non-members Autoenrolled % Opted-in through employer % Enrolled through a provider % Opted-out % Never been a member % Over $50,000 to $80,000 13.3 24.7 19.9 18.0 21.0 Over $80,000 to $120,000 3.6 10.3 8.6 5.3 7.0 Over $120,000 2.0 6.5 5.7 3.2 5.1 Zero Income 11.1 5.2 37.6 7.3 32.2 $30,000 or less 53.2 25.9 23.6 44.8 22.4 Over $30,000 to $50,000 21.4 35.5 17.6 27.2 21.7 Over $50,000 to $80,000 11.0 21.1 15.3 16.5 17.0 Over $80,000 to $120,000 2.4 7.9 3.9 3.3 4.6 Over $120,000 0.9 4.4 1.9 1.0 2.2 Employment income 65 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report Appendix 6: Change of scheme demographic characteristics at September 2010 No change of scheme % Change of scheme % All persons % 18–24 years 25.3 25.3 25.3 25–34 years 21.3 18.1 20.2 35–44 years 20.8 16.8 19.4 45–54 years 19.5 21.9 20.4 55–65 years 13.2 17.8 14.8 NZ European 69.6 69.0 69.4 Maori 16.3 14.9 15.8 Pacific Island 5.8 7.1 6.3 Asian 6.3 6.5 6.4 Other ethnicity 2.0 2.5 2.2 No qualification 12.1 15.1 13.2 School Certificate/NCEA 31.4 28.3 30.3 Trade/Diploma 35.9 34.0 34.0 University 14.4 14.2 14.2 Postgraduate 6.2 8.4 8.4 Upper white collar 45.6 44.6 45.3 Lower white collar 29.8 30.6 30.1 Upper blue collar 10.8 10.0 10.5 Lower blue collar 13.6 14.7 14.1 Unidentifiable 0.2 0.2 0.1 Salary and wages 79.5 82.8 80.7 Self-employed 4.5 3.6 4.2 Government benefits 9.5 8.3 9.1 Interest and investment 1.8 1.0 1.5 Other sources 2.7 2.5 2.6 No income source 1.8 1.9 1.9 Zero income 10.1 7.5 9.1 $30,000 or less 45.7 42.9 44.7 Age group Ethnicity Highest qualification Occupation Main income source Average employment income 66 Prepared by: National Research & Evaluation Unit No change of scheme % Change of scheme % All persons % Over $30,000 to $50,000 24.2 29.0 25.9 Over $50,000 to $80,000 14.3 13.6 14.0 Over $80,000 to $120,000 4.2 4.3 4.3 Over $120,000 1.5 2.8 2.0 Loss 0.2 0.4 0.26 Zero income 1.8 1.9 1.8 $30,000 or less 44.3 41.0 43.1 Over $30,000 to $50,000 27.7 30.9 28.8 Over $50,000 to $80,000 17.4 15.7 16.8 Over $80,000 to $120,000 5.8 5.6 5.7 Over $120,000 2.9 4.6 3.5 Average personal income 67 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report Appendix 7: Mean and median for cumulative member contribution value ($) from October 2007 to September 2010 Automatically enrolled Opted-in through employer Enrolled directly through a provider Mean Median Mean Median Mean Median 18–24 years 1,358 870 2,417 2,368 1,757 1,366 25–34 years 2,479 2,427 4,307 4,103 2,883 2,703 35–44 years 2,884 2,444 5,887 5,244 3,676 2,856 45–55 years 3,915 3,086 6,137 4,960 4,926 3,245 55–65 years 3,254 2,861 5,973 4,821 4,077 3,130 Male 3,113 2,620 6,693 5,308 5,001 3,425 Female 2,118 1,714 4,470 3,834 3,215 2,856 Salary and wages 2,943 2,546 5,832 4,783 5,213 4,008 Self-employed 2,374 2,241 5,541 3,828 2,858 2,697 Superannuation 2,065 2,059 3,181 2,543 2,084 2,262 Government benefits 708 667 1,366 910 1,028 451 Interest and investment 874 409 3,058 2,576 2,623 2,780 Other sources 1,198 524 4,234 3,911 2,391 2,773 No income source 1,533 767 2,523 1,031 2,415 2,686 No qualification 2,452 2,231 4,167 3,684 2,727 2,820 School Certificate/NCEA 2,091 1,774 4,864 4,395 3,036 2,897 Trade/Diploma 2,715 2,194 4,826 4,221 4,127 3,130 University 2,785 1,701 6,780 5,265 5,129 3,365 Postgraduate 3,958 3,050 4,826 4,221 6,062 4,039 NZ European 2,732 2,126 5,905 4,821 4,264 3,188 Māori 2,370 1,974 3,802 3,466 3,627 2,200 Pacific Island 2,124 1,280 4,271 4,273 1,910 2,135 Asian 2,259 2,097 5,411 4,264 2,948 2,800 Other ethnicity 2,297 2,853 5,311 5,029 2,148 1,320 Zero income 1,533 767 2,523 1,031 2,314 2,610 $30,000 or less 1,149 843 2,472 2,299 2,059 2,119 $31,000 to $50,000 2,675 2,911 4,198 4,134 3,296 3,092 Age band Gender Main income source Highest qualification Ethnicity Personal income 68 Prepared by: National Research and Evaluation Unit Automatically enrolled Opted-in through employer Enrolled directly through a provider Mean Median Mean Median Mean Median $51,000 to $80,000 4,562 4,715 6,067 5,979 4,611 4,050 $81,000 to $120,000 6,175 6,683 7,821 7,722 6,588 7,112 Over $120,000 10,679 10,762 14,950 13,991 11,541 11,054 69 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report Appendix 8: Demographic characteristics of those on a contributions holiday Demographic characteristics Taken a holiday All members Number % % 18–24 years 9,923 25.3 13.6 25–34 years 10,643 27.1 15.8 35–44 years 9,289 23.7 18.6 45–55 years 5,772 14.7 22.7 55–65 years 3,596 9.2 29.3 Male 18,387 46.9 46.1 Female 20,836 53.1 53.9 NZ European 28,144 71.8 74.0 Maori 6,122 15.6 11.3 Pacific Island 1,615 4.1 5.3 Asian 2,138 5.5 7.3 Other ethnicity 1,204 3.1 2.3 No qualification 3,858 9.8 12.9 School Certificate/NCEA 10,457 26.7 27.8 Trade/Diploma 15,675 40.0 34.7 University 5,387 13.7 15.2 Postgraduate 3,846 9.8 9.4 Upper white collar 15,897 46.5 48.0 Lower white collar 10,235 30.0 28.9 Upper blue collar 3,567 10.4 10.9 Lower blue collar 4,470 13.1 11.9 34,183 87.2 74.5 446 1.1 9.1 -- -- -- 2,855 7.3 8.0 Interest and investment 670 1.7 3.3 Other sources 545 1.4 1.4 No income source 525 1.3 0.7 Age band Gender Ethnicity Highest qualification Occupation Income source Salary and wages Self-employed Superannuation Government benefits 70 Prepared by: National Research and Evaluation Unit Demographic characteristics Taken a holiday All members Number % % Zero income 2,512 6.4 18.3 $30,000 or less 13,113 33.4 29.0 $31,000 to $50,000 12,966 33.1 26.5 $51,000 to $80,000 7,749 19.8 18.1 $81,000 to $120,000 1,990 5.1 5.0 892 2.3 3.2 Employment income Over $120,000 71 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report Appendix 9: Sources of income of respondents to 2010 survey (percentages) Total (687) Members (427) Non-members 256) All Main All Main All Main NZ Super 78 41 83 44 74 38 KiwiSaver 49 13 91 25 10 2 Savings account 56 9 54 7 58 10 Paid work 54 9 57 9 51 9 Rent from property 27 9 23 5 Other work-based schemes 26 7 25 7 26 7 Selling investment properties 15 5 11 4 18 6 Stocks, shares, bonds, unit trusts 22 4 23 4 22 4 Selling a business 12 4 7 1 16 7 Other superannuation plans 13 3 11 3 15 3 Downsizing house 18 2 22 3 16 2 Inheritance 22 2 22 1 23 4 Selling home and renting 5 1 5 1 5 1 Re-mortgaging home 2 1 2 - 2 1 Endowment/life assurance 13 - 13 - 13 - Other insurance policy 8 - 8 1 8 - Income from family 4 - 5 - 3 1 Selling possessions 5 - 6 - 4 - Don’t know 1 1 - - - - Other 2 - 6 1 2 2 72 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report Appendix 10: Survey of Steering Committee As part of this synthesis an online survey of the Steering Committee was undertaken. The purpose of this survey was to inform the final section of the report, which considered the effectiveness of the evaluation to date, and made recommendations for future activities. This appendix provides a summary of the data collected from this survey and is the basis for the comments made in the body of the report. A10.1 Survey respondents In total, 9 members of the Steering Committee responded to the survey. Table 26 summarises the number of respondents per agency. Seven agencies were represented. There were no respondents from Treasury or Statistics New Zealand. (Although one person from each of these agencies did access the survey they did not complete any of the questions beyond their name and agency identification.) Table 26: Respondents by agency Agency Number Inland Revenue 3 Financial Markets Authority 1 Commission for Financial Capability 1 Ministry of Business, Innovation and Employment 2 Ministry of Social Development 1 Victoria University 1 Total 9 The Steering Committee membership has changed over time. Respondents were asked how long they had been on the committee to provide some context to their answers with respect to their involvement in the evaluation. Two of the respondents had been members for more than six years and as such involved in the initial development of the evaluation plan. The other seven had all been members for less than three years, with two being on the committee for less than a year. respondents reported their knowledge was moderately good or higher, 3 reported it was slightly good and one commented that it was moderately poor. A10.1.1 The effectiveness of the evaluation Three questions asked respondents to rate the effectiveness of KiwiSaver. The scale used for these questions was: 1 = moderately poor; 2 = slightly poor; 3 = slightly good; 4 = moderately good; 5 = very good; 6 = extremely good. Respondents were also asked how they would rate their personal expertise and knowledge with regard to their ability to answer the survey questions. The scale used was from 1 = moderately poor through to 6 = extremely good. The average level of knowledge reported was 4, which equates to moderately good. Overall five A positively packed scale was used based on the assumption that respondents were more likely to have a positive view of the evaluation and this enabled a wider range of responses. 73 Prepared by: National Research and Evaluation Unit Respondents were firstly asked how they would rate the effectiveness of the evaluation in meeting its objectives. Given that some key pieces of work were still to be completed there were two separate questions as shown in the following tables. The respondent who reported a moderately poor level of knowledge did not answer these questions. The range of means across both questions was from 3.1 (slightly good) through to 4.5 (moderately/very good). The highest level of effectiveness was reported for understanding the scale and uptake of KiwiSaver, the lowest for its impact on superannuation markets and the financial sector. Table 27: Mean levels of effectiveness of the evaluation Mean Informing the ongoing development and delivery of KiwiSaver 4.1 Understanding the scale and pattern of uptake of KiwiSaver 4.5 Assessing whether the KiwiSaver features are generating expected outcomes 3.9 Determining the impact of KiwiSaver on superannuation markets and the financial sector 3.1 Table 28: Mean levels of effectiveness of the evaluation to date Mean The impact of KiwiSaver on individual’s savings habits and asset accumulation 3.8 The impact of KiwiSaver on national savings 3.5 The value for money of KiwiSaver as a policy 3.4 Respondents were also asked to rate the evaluation in terms of a number of general criteria related to its implementation and usefulness based on the work undertaken to date (Table 29). The mean levels of effectiveness regarding the information provided are remarkably consistent and all equate to it being moderately good. The highest rating was given for the usefulness of the information ( xˉ = 4.4); the lowest for the value for money of the information provided ( xˉ = 3.7). A E A A E A The overall model had a mean level of effectiveness of 5.1, which equates to very good and suggests that people are, generally satisfied with how this works. 74 Prepared by: National Research and Evaluation Unit KiwiSaver evaluation: Final summary report Table 29: Effectiveness of the evaluation in terms of its implementation and usefulness Mean The timeliness of the information provided 4.0 The usefulness of the information provided 4.4 The usability of the information provided 4.1 The relevance of the information provided 4.1 The depth of the information provided 4.1 The breadth of the information provided 4.0 The value for money of the information provided 3.7 The overall model (lead agency with others doing individual research) 5.1 In addition to these three questions respondents were also asked two open-ended questions regarding the effectiveness of the evaluation.  Continued review, and inter-agency collaboration.  I have been on the Steering Group panel for a relatively short period of time, approximately 8 months. As such, I cannot comment in detail on the pieces of work that were delivered and discussed prior to this time. However, I have found the findings from the qualitative research involving indepth interviews with KiwiSaver members and nonmembers, employers and scheme providers interesting and insightful—painting a picture of the attitudes and behaviour of individuals and employers towards KiwiSaver. The discussions I have been involved in re: the scope for the value of money research was a useful cross-agency collaboration and I would expect this to produce some useful and relevant information.  The (lack of) additionality result, for its policy implications. For me the most important result will be value for money and that is not finished yet. First, they were asked what they considered to be the most useful aspects of and/or findings from the evaluation to date. Their responses were as follows:      The true success of KiwiSaver policy will be measured over decades more than the initial few years. That said, the KiwiSaver evaluation has enabled very useful initial research and the establishment of baselines against which future research can be compared. Secondly, the KiwiSaver evaluation has driven excellent inter-agency engagement and communication for the initial life of KiwiSaver. While the focus of the Steering Committee has been on KiwiSaver evaluation, it has also driven better cross-agency understanding and alignment. Excellent understanding of uptake, member demographics, response to the different features of KiwiSaver. They were also asked if they had any general comments to make. Four respondents made further comments to those above: The most useful aspect from my perspective, which is a by-product, has been the inter-agency contact and being linked into the research that has and is being done. I think the most useful thing has been bringing a range of government departments together to collaborate on a policy change that has a wide range of impacts. Having a summary provided annually, which has assisted in identifying trends in the first few years of KiwiSaver. Providing a forum to ask the questions around the effectiveness of certain KiwiSaver features. 75 Prepared by: National Research and Evaluation Unit  An excellent example of different parts of the sector coming together to work on a common purpose and achieving far more than any one agency could have achieved on its own.  I think it is difficult to quantify the value of the work done to date given KiwiSaver has only been in existence for just under 7 years and we had the GSC in 2008. However the work will be useful in my view from a benchmarking perspective.  It was a good exercise, but KiwiSaver has now become “Business as Usual” and except for completing the SoFIE work, there is probably no longer a justification for a specific evaluation project.  Firstly, they were asked to rank five different options for the evaluation with a sixth option of “other”. The option with the lowest average, and therefore the most preferred overall, was for regular reviews of KiwiSaver. The least preferred was the development of a comprehensive framework to cover the next 10 years. Generally the evaluation has been a model for large and costly programmes. A10.2 Future evaluation activities The second part of the survey asked respondents to consider the future of the evaluation, beyond June 2014. Table 30: Average ranking for each option for the evaluation Mean Regular reviews of KiwiSaver including some research and the use of the administrative data 2.6 The development of another stand-alone, comprehensive monitoring, research and evaluation framework to cover the next five years 3.0 Regular monitoring of KiwiSaver through administrative data 3.0 Incorporating KiwiSaver into relevant agency monitoring and evaluation frameworks rather than a specific KiwiSaver evaluation 3.1 The development of another stand-alone, comprehensive monitoring, research and evaluation framework to cover the next 10 years 3.3 As the average scores show, the respondents all had very differing views on what is needed. For example, four different items were selected as the most preferred option by two people. Only one person ranked regular reviews of KiwiSaver above a 3. It was also the only item that was not ranked a 5 or 6 by any respondent. For both items suggesting comprehensive evaluation frameworks five respondents ranked them 4 or 5. Two respondents offered a different option to those listed in the survey. In both instances they ranked this th 6 in order of preference. These options were for regular monitoring of KiwiSaver for the next five years before developing another comprehensive evaluation plan for the next five years. The other option was 27 related to the decumulation problem. P I think the key thing is to recognise that much of the research conducted to date has only been done at the initial stages of KiwiSaver. Areas of research such as impacts on scheme providers and employers, outcomes for members, value for money all need to be revisited to compare against the established baselines and to ensure that KiwiSaver is on targeted to achieve its intended outcomes. This further research would ensure that the Government’s ongoing significant investment in KiwiSaver is achieving the desired outcomes and it also helps to inform future policy development.  It would need to explore a full range of decumulation issues. Confirmation of additionality assumptions.  Non-contributors not on an s64 holiday. Member ethnicity. Contributory members’ analysis by Stats NZ industry classification. Analysis of membership from an affordability perspective.  My answer to 10 and 11 really depends on what the angle of the evaluation is. Either IR or Tsy would be interested from an overall financial/savings perspective. From an MBIE perspective we would really only focus on the housing market aspect of KS, which is quite small. But we would certainly be interested in understanding how the first home P P26F P Respondents were also asked which agency they believed should be the lead if an evaluation plan was to be developed. Of the eight respondents to this question, six suggested Inland Revenue while two suggested Treasury. Finally respondents were asked what they thought the key interests of any research and evaluation should be moving forward: 27  Decumulation is the “spending” of one’s assets to fund retirement 76 Prepared by: National Research and Evaluation Unit package has affected the Housing market generally, especially in relation to house price inflation—and the interplay with the LVR restrictions.  Principal objectives continue to be impact of KS on personal savings; Impact of KS on national savings; Impact of KS on financial sector productivity  The extent to which KiwiSaver is used versus alternative saving options. What affects changes in contributions? Is KiwiSaver targeting the right people? Comparison with international models. Issues around employer contributions and firm size  Implications for social equity; the decumulation problem. 77 Prepared by: National Research and Evaluation Unit