Assessing alternative approaches to design tax and financial incentives for retirement savings

Transcription

1 Organisation for Economic Co-operation and Development DAF/AS/PEN/WD(2017)11 English - Or. English DIRECTORATE FOR FINANCIAL AND ENTERPRISE AFFAIRS INSURANCE AND PRIVATE PENSIONS COMMITTEE 10 November 2017 Working Party on Private Pensions Assessing alternative approaches to design tax and financial incentives for retirement savings 4-5 December 2017 This document is circulated for discussion under the agenda of the WPPP meeting. For further information, please contact Ms. Stéphanie Payet [Tel: ; stephanie.payet@oecd.org] or Mr. Pablo Antolin [Tel: ; Pablo.antolin@oecd.org] JT This document, as well as any data and map included herein, are without prejudice to the status of or sovereignty over any territory, to the delimitation of international frontiers and boundaries and to the name of any territory, city or area.

2 2 DAF/AS/PEN/WD(2017)11 Assessing alternative approaches to design tax and financial incentives for retirement savings 1. The project on Financial Incentives and Retirement Savings assesses tax and other financial incentives to encourage saving for retirement in funded private pension arrangements. The project (DAF/AS/PEN/WD(2014)11), endorsed by the OECD Working Party on Private Pensions (WPPP) in 2014, examines firstly, the different tax and financial incentives that exist and the tax advantage they provide; secondly, whether those incentives are cost efficient in terms of increasing contributions into funded pension arrangements and, ultimately, contributing to adequate overall retirement incomes; and thirdly, other alternative approaches to encourage retirement savings that may be more efficient. 2. The WPPP has already discussed a number of documents examining the different tax and financial incentives that exist, the tax advantage they provide, their fiscal cost and their impact on retirement savings. DAF/AS/PEN/WD(2015)3 and DAF/AS/PEN/WD(2015)3/ADD1, discussed in June 2015, assess the differences across various tax regimes that exist for private pension plans and compile country profiles with detailed information on the tax treatment of funded private pension plans. DAF/AS/PEN/WD(2014)14, discussed in December 2014, presents the methodology to calculate the tax advantage that individuals saving into private pension plans may enjoy over their lifetime. DAF/AS/PEN/WD(2015)11, discussed in December 2015, examines whether the tax treatment of private pension plans in different OECD countries provides a lifetime advantage when people save for retirement. DAF/AS/PEN/WD(2016)14/REV1 and DAF/AS/PEN/WD(2016)7, discussed in June 2017, assess the fiscal cost of the tax treatment of private pensions, accounting for personal income tax but also corporate income tax and the impact of new savings. Finally, DAF/AS/PEN/WD(2017)2, also discussed in June 2017, reviews the empirical literature on the impact of tax incentives and other policies on individuals' behaviour with respect to retirement savings. 3. This document covers the next step of the project. It provides a theoretical assessment of alternative approaches to design tax and financial incentives. In addition, DAF/AS/PEN/WD(2017)12 examines the actual cost effectiveness of tax and financial incentives for OECD and EU countries. The project will end with a set of policy guidelines and effective approaches on how to design tax and financial incentives to promote retirement savings. 4. The landscape of incentives for retirement savings is changing. Historically, tax incentives were the most common type of incentive for retirement savings, in particular the one deferring taxation to the retirement phase, taxing individuals only on their pension benefits. This is still the main approach currently implemented, but recent developments, such as government matching contributions, government flat-rate subsidies, or providing tax credits rather than tax deductions are changing the landscape of incentives for retirement savings. These alternative approaches to design incentives may alter the tax advantages that individuals get when saving in funded pension arrangements instead of in traditional savings vehicles, the relationship between tax advantage and income, and the fiscal cost for the treasury. 5. This document assesses alternative approaches to design incentives for retirement savings to see whether they would be an improvement on the current main approach.

3 DAF/AS/PEN/WD(2017)11 3 Improvement is measured through the tax advantage that people may get over their lifetime when saving for retirement, and through the fiscal cost for the treasury, relative to the tax advantage and fiscal cost of the current main approach based on taxing only pension benefits and exempting contributions and returns on investment. This document does not cover other potential ways of assessing improvements on the current approach to provide incentives, such as measuring whether alternative approaches achieve better coverage, higher contribution levels, or higher national savings. DAF/AS/PEN/WD(2017)12 discusses potential links between incentives and coverage and contribution levels The best approach to design incentives for retirement savings depends on the objective policy makers and regulators have in mind. In order to increase the attractiveness of saving for retirement for middle to low-income groups, tax credits, tax deductions at fixed rate and matching contributions combined with tax incentives can smooth out the tax advantage across income groups. Tax credits expressed as a flat-rate amount, matching contributions, and flat-rate subsidies can also be used to target the tax advantage at low-income earners. Unfortunately, tax credits suffer from important relative drawbacks as they lead to lower pension benefits and higher fiscal costs for the treasury. When focusing on reducing the fiscal cost of tax incentives to the treasury, the assessment needs to differentiate between the short term and the long term impacts, as some approaches to provide incentives may temporarily reduce the fiscal cost but end up costing more in the long term than the current main approach of only taxing withdrawals. Tax incentives in which pension benefits are taxed tend to achieve comparatively lower fiscal costs than other approaches in the long term but not necessarily in the short term. Finally, there are no alternative approaches to design incentives for retirement savings that can improve on the current main approach of taxing only pension benefits by simultaneously providing higher tax advantages to individuals at a lower long-term cost for the treasury. 7. The document proceeds as follows. Section 1 describes the most common approached used currently by countries to promote retirement savings, what tax advantages it provides to individuals and at what cost for the treasury. Section 2 lists the alternative approaches to design tax and financial incentives and builds illustrative cases representing the most relevant approaches on which the comparative analysis is based in subsequent sections. Section 3 compares the tax advantage provided by alternative approaches, while section 4 examines the evolution of the fiscal cost of a transition from the current approach to alternative approaches to design incentives for retirement savings. Section 5 includes a policy discussion and concludes. 8. Delegates are invited to address the following questions: Could you suggest other alternative approaches to design tax and financial incentives for retirement savings? Which additional policy implications of using different alternative approaches to design tax and financial incentives for retirement savings could be worth looking at? 1 DAF/AS/PEN/WD(2017)2 examines the literature review on the impact of tax incentives and other policies on retirement savings.

4 4 DAF/AS/PEN/WD(2017)11 1. Taxing pension benefits while exempting contributions and returns is currently the most common approach to design incentives for retirement savings 9. This section describes the most common approach used currently by countries to promote retirement savings, what tax advantages this approach provides to individuals and at what cost for the treasury. 10. Tax incentives tend to be the main type of incentive provided by governments to promote savings in funded pension arrangements. Tax incentives come from a differential tax treatment applied to funded pension arrangements as compared to other savings vehicles. In most countries, the "Taxed-Taxed-Exempt" ("TTE") tax regime applies to traditional savings vehicles, meaning that money saved come from after-tax income, returns on investment are taxed as income and withdrawals are tax exempt. When the tax regime applied to funded pension arrangements deviates from the "TTE" tax regime, this could lead to a reduction in taxes paid over the individual's life time. 11. The most common tax treatment of retirement savings exempts contributions (i.e. they are deductible from taxable income) and returns on investment from taxation while it taxes pension benefits and withdrawals as income. About half of OECD and EU countries apply a variant of this "Exempt-Exempt-Taxed" ("EET") tax regime to retirement savings (Figure 1). Figure 1. Tax treatment of retirement savings in funded pension arrangements OECD and EU countries EEE Bulgaria, Slovak Republic TET Austria, Belgium, France, Israel, Korea, Malta, Portugal TEE Czech Republic, Hungary, Lithuania, Luxembourg, Mexico EET Canada, Chile, Croatia, Estonia, Finland, Germany, Greece, Iceland, Ireland, Japan, Latvia, Netherlands, Norway, Poland, Romania, Slovenia, Spain, Switzerland, United Kingdom, United States ETT Denmark, Italy, Sweden TTE Australia, New Zealand, Turkey Source: OECD (2015 [1] ). 12. This tax regime provides a tax advantage to the individual. Figure 2 shows the overall tax advantage (solid line) broken down by components (tax advantage on contributions, tax advantage on returns and tax advantage on withdrawals) for different levels of income. The overall tax advantage represents the amount saved in taxes by the

5 DAF/AS/PEN/WD(2017)11 5 individual over his/her working and retirement years when contributing the same amount (before tax) to a pension plan instead of to a benchmark savings vehicle. It is expressed as a percentage of the present value of pre-tax contributions. The overall tax advantage provided by the "EET" tax regime is positive for all income groups, as the preferential tax treatment that contributions and returns on investment enjoy (represented by positive blue and grey bars respectively in the chart) outweighs the taxation of withdrawals (represented by negative light-blue bars in the chart). Figure 2. Overall tax advantage for an "EET" tax regime, by income level and components 80% Taxes saved over a lifetime, as a percentage of the present value of contributions Contributions Returns Withdrawals Overall tax advantage 60% 40% 20% 0% -20% -40% -60% Income level (multiple of average earnings) Notes: The calculations assumes an individual contributing 10% of wages from age 20 to 64 and withdrawing benefits from age 65 to 84 as a fixed-payment annuity. The tax system is such that tax rates increase with taxable income. The analysis assumes an income tax system with the following tax rates and tax brackets: 0% for income below EUR 9 700; 14% for income [EUR EUR ]; 30% for income [EUR EUR ]; 41% for income [EUR EUR ]; 45% for income above EUR Average earnings are assumed to be EUR This is inspired from the French tax system 13. In addition, the tax advantage provided by the "EET" tax regime increases with income when tax rates increase with taxable income. 2 Indeed, contributions are deducted from taxable income at a rate that grows with the income level of the individual. In the same way, returns on investment are exempted from taxation at a higher rate for individuals with income reaching higher tax brackets. Therefore, the tax advantage on contributions and the tax advantage on returns increase with income. This is partially offset by the taxation of withdrawals, as the average tax rate on withdrawals increases with the level of income. For individuals with income more than four times average earnings, the overall tax advantage even slightly declines. 2 When all individuals have their entire income taxed at the same rate, the overall tax advantage remains constant across all income groups.

6 6 DAF/AS/PEN/WD(2017) It is also worth noting that, for an individual whose income remains subject to the same marginal income tax rate during retirement as while working, the overall tax advantage is equal to the tax advantage on returns, because the taxation of withdrawals actually compensates for the tax advantage on contributions (for example, for the average earner in Figure 2, the tax advantage on contributions is equal to 30% and the tax advantage on withdrawals is equal to -30%). Similarly, if an individual's pension income is taxed at a lower rate than past work income, the overall tax advantage will be higher than just the tax advantage on returns, because the tax advantage on contributions will not be fully compensated by the taxation of withdrawals (for example, for an individual with income equal to twice the average earnings in Figure 2, the tax advantage on contributions is equal to 41% and the tax advantage on withdrawals is equal to -36%). 15. Finally, countries where the tax treatment of retirement savings follows the "EET" tax regime can expect a long-term fiscal cost. Figure 3 presents projections of the net tax expenditure level of a funded pension arrangement with an "EET" tax regime since its introduction. The net tax expenditure incurred by an "EET" arrangement, with respect to a "TTE" benchmark, is the sum of tax revenue foregone on contributions, tax revenue foregone on returns and tax collected on withdrawals. The lag in the growth of benefits behind that of assets and investment returns is what creates the temporary increase in the net tax expenditure level. As the density of contributions progresses, asset and investment return levels grow and so, with a lag, do benefit levels. Once the system has reached maturity, i.e. all retirees draw their pension based on a full career and constant contribution rules, the net tax expenditure stabilises at its steady-state level. That level is positive, which implies a long-term fiscal cost for the treasury. Figure 3. Net tax expenditure for a maturing "EET" pension system, by components As a percentage of GDP 1.6% 1.4% 1.2% 0.8% 0.6% 0.4% 0.2% Revenue foregone on contributions Revenue collected on withdrawals Revenue foregone on returns Total Number of years since introduction Note: The calculations assume that individuals save from age 20 to 64 and draw pension benefits from age 65 to 84; contributions represent 3% of GDP; the number of people in each single-year age cohort is equal; the same average tax rate (30%) applies to all sources of income; a nominal rate of return of 5.06% (3% real return plus 2% inflation); and GDP growth at 3.53% (1.5% real growth plus 2% inflation). 16. Tax revenue collected on withdrawals more than offsets tax revenue foregone on contributions, reducing significantly the fiscal cost related to an "EET" tax regime when the system is mature. As opposed to the tax advantage which is calculated over the lifetime of an individual, the fiscal cost is calculated over the total population at different

7 DAF/AS/PEN/WD(2017)11 7 points in time. On one hand, at the level of the individual, the present value of the sum of all taxes collected on withdrawals during retirement is equal to the present value of the sum of all taxes saved on contributions while working, as long as the same tax rate applies during working and retirement years. On the other hand, the calculation of the net tax expenditure sums up revenues foregone on contributions and on returns from working individuals with revenues collected on withdrawals from retirees. The size of withdrawals in a given year when the system is mature is more important than the size of contributions, as withdrawals are the result of several years of contributions accumulating with compound interests. Figure 3 shows that the tax collected on withdrawals more than compensates for tax revenue foregone on contributions once the system has reached maturity. 2. What are the alternative approaches to design incentives for retirement savings? 17. This section identifies alternative approaches to design incentives for retirement savings and builds illustrative cases that will help compare different approaches in the subsequent sections. 18. There are other tax regimes than the one exempting contributions and returns while taxing withdrawals ("EET") in OECD and EU countries to promote retirement savings. Figure 1 shows that a broad range of tax regimes apply to retirement savings, from the "Exempt-Exempt-Exempt" ("EEE") tax regime where contributions, returns on investment and pension income are all tax-exempt, to regimes where two out of three flows of income are taxed. 19. The tax treatment of the different flows (contributions, returns on investment and withdrawals) can be somewhere in between the two opposite cases of full taxation and full exemption. Three examples taken from Canada, Australia and France illustrate this for each of the respective flows. In Canada, pension contributions are deductible from income, but only up to a certain nominal ceiling, which means that contributions above the ceiling are not tax deductible. In Australia, returns on investment are taxed at the fixed rate of 15%, but this rate is lower than the marginal income tax rate for most individuals. This means that, for most people, the tax rate applied to returns in superannuation funds is lower than the one applying to returns generated in other savings vehicles. Finally, in France, pension benefits paid as annuities are treated as taxable income after a 10% deduction. Therefore, partial tax relief and partial taxation are possible for the three types of flows. 20. In particular, partial tax relief on contributions is widespread and can take several different forms: Partial tax deduction: Only a portion of the contributions is tax deductible. For example, in Austria, 25% of an individual's contributions to pension companies and occupational group insurance are tax deductible. Like with the "EET" tax treatment, the partial tax deductibility reduces an individual's taxable income. The reduction therefore applies at the individual's marginal income tax rate. Ceiling on deductible contributions: Contributions are tax deductible up to a nominal ceiling. Contributions above the ceiling are not tax deductible and are therefore taxed at the individual's marginal income tax rate. Most countries have ceilings on tax-deductible contributions. Tax deduction at a fixed rate: Contributions are tax deductible at the same fixed rate for everyone, independently of the individual's marginal income tax rate. This

8 8 DAF/AS/PEN/WD(2017)11 approach is discussed as an alternative to tax deduction at the individual's marginal income tax rate in some countries, like the United Kingdom, but is not implemented in any country yet. Non-refundable tax credit: Contributions are not tax deductible but the individual receives a tax credit which reduces the amount of personal income tax due. The value of the tax credit can be expressed as a percentage of contributions (e.g. Belgium, Estonia, Finland, Israel and Korea) or as a flat-rate amount (this approach is common for mortgages but it is not implemented in any country yet for retirement savings). The fact that the tax credit is non-refundable means that the value of the tax credit cannot exceed the amount of personal income tax due. Rate relief or taxation at a fixed rate: Contributions are taxed at the same fixed rate for everyone. For example, in Australia, pension contributions are taxed at the fixed rate of 15%. 3 As this rate is usually lower than the marginal income tax rate, most individuals benefit from a rate relief on their contributions. 21. Financial incentives in the form of matching contributions or flat-rate subsidies are increasingly popular. Financial incentives are payments made by the government directly in the pension account of eligible individuals, thus increasing the assets accumulated to finance retirement. Research shows that understanding around tax relief is low among low-income individuals, who also tend to consider tax incentives as a secondary determinant in their decision to save (Sandler, 2002 [2] ; Clery, Humphrey and Bourne, 2010 [3] ; Kempson, McKay and Collard, 2003 [4] ). Policy makers in some countries therefore offer financial incentives to encourage low-to-middle income individuals to save for retirement. 22. Matching contributions are usually conditional on the individual contributing and correspond to a certain proportion of the individual's own contributions, up to a maximum amount. The match rate paid by the government varies greatly across countries, from 3% in Austria to 325% in Mexico (programme for civil servants). 4 A match rate of 50% can be found in Australia, where the matching contribution is targeted at low-income individuals, and in New Zealand, where it applies to all plan members. 23. Flat-rate subsidies are designed to attract low-income individuals as the fixed subsidy represents a higher share of their income. They can be found in Chile, Germany, Lithuania and Mexico. 24. Financial incentives may be used as a substitute or as a complement to tax incentives. For example, in Australia, the super co-contribution is a government matching contribution for low-income individuals making a voluntary contribution to their superannuation fund. This voluntary contribution is made from after-tax income, i.e. from money on which the individual has already been taxed at his/her marginal rate. In that case, the matching contribution comes as a substitute for any tax relief on contributions. On the contrary, in the case of Riester pension plans in Germany, high-income individuals can cumulate the flat-rate subsidy with tax deductions of contributions (although the amount that can be deducted is reduced by the value of the subsidy). There, both types of incentives complement each other. 3 A higher tax rate applies for very-higher earners, at 30%. 4 The Mexican government matches each peso contributed voluntarily by civil servants with 3.25 pesos. Individuals can contribute up to 2% of their income to this plan.

9 DAF/AS/PEN/WD(2017) The analysis in the next sections uses illustrative cases of the different alternative approaches to design incentives for retirement savings discussed above. The selected illustrative cases are presented in Box 1. They can be split in three groups: alternative tax incentives to the "EET" tax regime, financial incentives and different approaches to provide partial tax relief on contributions. For the alternative tax incentives, the analysis chooses a particular design for each of the general tax regimes presented in Figure 1. 5 Regarding financial incentives, the analysis assesses matching contributions and flat-rate subsidies as substitutes for tax incentives. Therefore, the underlying tax regime is the same one as for traditional savings vehicles. This allows examining the specificities of the financial incentives as compared to the tax incentives. Box 1. Selected illustrative cases of alternative approaches to design incentives for retirement savings Alternative tax incentives: "EEE": contributions, returns and withdrawals are tax free "ETE": contributions are tax deductible, returns are taxed at 15% and withdrawals are tax free "ETT": contributions are tax deductible, returns are taxed at 15% and withdrawals are taxed at the individual's marginal tax rate "TEE": contributions are taxed at the individual's marginal tax rate, returns and withdrawals are tax free "TET": 50% of contributions are tax deductible, returns are tax free and 50% of withdrawals are tax exempt "TTE": contributions and returns are taxed at 15% and withdrawals are tax free Financial incentives: Matching contribution: 50% match rate Flat-rate subsidy: 1% of the average earnings in the economy Approaches to provide partial tax relief on contributions: Partial tax deductibility Limited tax deductibility Tax deductibility at fixed rate Non-refundable tax credit (expressed as a percentage of contributions) Non-refundable tax credit (expressed as a flat-rate amount) Rate relief / Taxation at fixed rate Memo item: Tax treatment of traditional savings vehicles "TTE": contributions and returns are taxed at the individual's marginal tax rate and withdrawals are tax free 5 The tax regime that taxes only returns on investment is also considered ("ETE"), as it applies in some countries, for example to voluntary private pension plans in Latvia.

10 10 DAF/AS/PEN/WD(2017)11 3. How does the tax advantages provided by alternative approaches to design incentives for retirement savings compare with the one provided by the current "EET" approach? 26. This section compares the tax advantage provided by alternative approaches to design incentives for retirement savings to the one provided by the main current "EET" approach. In order to compare tax incentives and financial incentives, the overall tax advantage is also calculated in the case of flat-rate subsidies and matching contributions, by interpreting them as refundable tax credits paid into the pension account. The section starts by comparing different alternative tax and financial incentives and then compares different approaches to provide partial tax relief on contributions. Alternative tax and financial incentives 27. Alternative tax and financial incentives for retirement savings to the most commonly used approach of taxing only withdrawals ("EET"), provide different overall tax advantages. Figure 4 provides this comparison. The overall tax advantage results from comparing each tax and financial incentive separately to the tax treatment of a benchmark traditional savings vehicle where contributions and returns on investment are taxed at the individual's marginal tax rate and withdrawals are exempt. Approaches that tax favourably contributions and returns on investment (full or partial exemption) while taxing fully withdrawals ("EET", "TET" and "ETT" tax regimes) provide an overall tax advantage that nets outs a positive tax advantage on contributions, a positive tax advantage on returns and a negative tax advantage on withdrawals. Approaches that tax favourably contributions and returns on investment (full or partial exemption) while not taxing withdrawals provide an overall tax advantage that results from a positive tax advantage on contributions and a positive tax advantage on returns ("EEE", "ETE" and "TTE" tax regimes). 6 The tax advantage when retirement savings are taxed upfront (i.e. only contributions are taxed, "TEE") comes from the tax exemption of returns on investment. Finally, financial incentives provide an overall tax advantage that result from a positive tax advantage on contributions and a negative tax advantage on returns. This negative tax advantage on returns is due to the fact that financial incentives increase the amount contributed in the pension plan, thereby increasing investment income and the tax due on it, as compared to the benchmark traditional savings vehicle. 28. The ranking of the different types of tax and financial incentives on the overall tax advantage vary according to the parameters used. For example, a 50% matching contribution provides a larger tax advantage to the individual than the "EET" tax regime, but this is reversed if the match rate is only 20%. The breakeven point is for a match rate of 29.62%, which is slightly lower than the average earner's marginal tax rate of 30% assumed in the calculations. Finally, compared to the "EET" approach, removing the taxation of withdrawals ("EEE" tax regime) will always increase the tax advantage, while 6 For the approach that taxes only returns on investment at a fixed rate of 15% ("ETE" tax regime), the analysis shows that the tax advantage on returns is positive. If the tax rate increases, this tax advantage on returns diminishes and even turns negative for a tax rate above 20%. This stems from the fact that the amount contributed in the pension plan is higher than in the benchmark traditional savings vehicle (because of the tax deductibility of contributions), thereby increasing investment income and the tax due on it.

11 DAF/AS/PEN/WD(2017)11 11 starting to tax returns on investment ("ETT" tax regime) will always reduce the tax advantage. Figure 4. Overall tax advantage for the "EET" tax regime and alternative tax and financial incentives, by components, average earner Taxes saved over a lifetime, as a percentage of the present value of contributions Contributions Returns Withdrawals Overall tax advantage 60% 50% 40% 30% 20% 10% 0% -10% -20% -30% -40% EET EEE Matching ETE TET TTE TEE ETT Subsidy Notes: "E" stands for exempt and "T" for taxed. The calculations assumes an average earner contributing 10% of wages from age 20 to 64 and withdrawing benefits from age 65 to 84 as a fixed-payment annuity. See Box 1 for the detailed design of each incentive. 29. An individual whose income remains subject to the same marginal income tax rate during retirement as while working is neutral in terms of taxes saved between the two approaches that tax retirement savings upfront versus in retirement. Figure 4 shows that the "TEE" and the "EET" tax incentives provide the same overall tax advantage to the average earner. As explained before, with the "EET" tax regime, the tax advantage on contributions is compensated by the tax due on withdrawals when the individual's marginal income tax rate is the same during retirement years as it was during working years. The overall tax advantage is therefore down to the tax advantage on returns, just as for the "TEE" tax regime. 30. Tax incentives provide an overall tax advantage that increases with income when tax rates increase with taxable income. Figure 5 illustrates how the overall tax advantage changes with the level of the income for the most common approach currently used to design incentives for retirement savings ("EET") and alternative tax and financial incentives. All the tax incentives provide a larger overall tax advantage to individuals with higher earnings. This is because of the underlying tax system, in which tax rates increase with income. This type of tax system implies that any tax relief is given at a higher rate for individuals whose income reaches higher tax brackets. 31. On the contrary, in tax systems where the same tax rate applies to the whole income of the individual, all tax incentives become income neutral, because any tax relief is given at the same rate for everyone.

12 12 DAF/AS/PEN/WD(2017)11 Figure 5. Overall tax advantage for the "EET" tax regime and alternative tax and financial incentives, by income level Taxes saved over a lifetime, as a percentage of the present value of contributions 70% 60% 50% 40% 30% 20% 10% 0% -10% EEE ETE TTE Matching TET EET TEE ETT Subsidy -20% Income level (multiple of average earnings) Notes: "E" stands for exempt and "T" for taxed. The calculations assumes an individual contributing 10% of wages from age 20 to 64 and withdrawing benefits from age 65 to 84 as a fixed-payment annuity. The tax system is such that tax rates increase with taxable income. See Box 1 for the detailed design of each incentive. 32. Financial incentives by themselves without any preferential tax treatment provide a higher overall tax advantage to low-income earners when tax rates increase with taxable income. This is not surprising for flat-rate subsidies, but less intuitive in the case of matching contributions. As the value of a flat-rate subsidy is fixed across the income scale, its relative value for low-income earners is higher. With matching contributions, the tax advantage on contributions is income neutral as everybody gets the same match rate, but the tax due on returns increases with income. Indeed, as returns on investment are taxable, the tax advantage on returns is negative because matching contributions increase the level of after-tax contributions and generate additional investment income, compared to a traditional savings vehicle. Therefore, the overall tax advantage provided by matching contributions declines with the individual's income level as returns get taxed at a higher marginal tax rate. 33. In tax systems where the same tax rate applies to the whole income of the individual, matching contributions become income neutral, but flat-rate subsidies remain more advantageous to low-income earners. Matching contributions provide the same advantage to all income groups in such tax systems because the match rate is the same for all and returns on investment are taxed at the same rate for everybody. Flat-rate subsidies provide a tax advantage that declines with the income level because the value of the subsidy is fixed. 34. Tax incentives make high-income earners better-off, while financial incentives make low-income earners better-off. High-income earners tend to be better-off with tax incentives especially those combining tax relief on contributions (full or partial) with tax exempt withdrawals, because they allow high-income earners to reduce their taxable

13 DAF/AS/PEN/WD(2017)11 13 income by deducting their contributions, and to have tax exempt pension benefits. Lowincome earners are better-off with financial incentives. Both matching contributions and flat-rate subsidies bring low-income earners more tax advantages than any tax incentive, even the one where none of the flows are taxed ("EEE" tax regime), as they provide a monetary advantage that is generally higher than the tax relief provided by tax incentives. Finally, middle-income earners would get the largest overall tax advantage from tax incentives combining tax-deductible contributions with tax exempt withdrawals ("EEE" and "ETE" tax regimes) or from matching contributions The two approaches that tax retirement savings upfront versus in retirement are not equivalent in terms of taxes saved for all income groups. This stems from the fact that some individuals face different tax rates over their lives in tax systems where tax rates increase with taxable income. Therefore, low-income earners tend to be better-off with upfront taxation (the "TEE" tax regime), as they pay little tax on their contributions and do not pay any further tax. Conversely, above-average earners are better-off with taxation in retirement (the "EET" tax regime) because they can reduce significantly their taxable income by deducting contributions, while not paying as much taxes once retired comparatively, if their retirement income is taxed at a lower average tax rate than their past work income. Approaches to provide partial tax relief on contributions 36. Different approaches can be used to provide partial tax relief on contributions and provide the same tax advantage to a given individual. Table 1 presents selected cases where different approaches to provide partial tax relief on contributions achieve the same level of overall tax advantage. For all the approaches, the analysis assumes that returns on investment are tax exempt and that withdrawals are considered as taxable income. Therefore, only the tax treatment of contributions differs between the different approaches. It is possible to design the different approaches in such a way that they will provide the same overall tax advantage to an average earner. For example, it is equivalent to deduct 50% of contributions; to deduct contributions up to a nominal ceiling equal to 6.57% of the average earnings in the economy; to deduct contributions at a rate equal to half the individual's marginal income tax rate; to provide a tax credit equal to 10.5% of pre-tax contributions; to provide a tax credit equal to a nominal amount corresponding to 1.38% of the average earnings in the economy; or to tax contributions at a rate equal to half the individual's marginal income tax rate. 7 As explained previously, these rankings also depend on the specification used in the analysis.

14 14 DAF/AS/PEN/WD(2017)11 Table 1. Selected cases where different approaches to provide partial tax relief on contributions achieve the same level of overall tax advantage Partial deduction Other approaches giving the same overall tax advantage Limited deductibility Deduction at fixed rate Tax credit (%) Tax credit (flat) Taxation at fixed rate Overall tax advantage (% of the present value of contributions) 0% of contributions deductible (=TET) No deduction of contributions No deduction of contributions No tax credit No tax credit Contributions taxed at the MTR 25% of contributions deductible at the MTR Contributions up to 3.29% of average earnings deductible at the MTR Contributions deductible at 25% of the MTR Credit equal to 5.25% of pre-tax contributions Credit equal to 0.69% of average earnings Contributions taxed at a rate equal to 75% of the MTR 50% of contributions deductible at the MTR Contributions up to 6.57% of average earnings deductible at the MTR Contributions deductible at half of the MTR Credit equal to 10.5% of pre-tax contributions Credit equal to 1.38% of average earnings Contributions taxed at a rate equal to half of the MTR 75% of contributions deductible at the MTR Contributions up to 9.86% of average earnings deductible at the MTR Contributions deductible at 75% of the MTR Credit equal to 15.75% of pre-tax contributions Credit equal to 2.07% of average earnings Contributions taxed at a rate equal to 25% of the MTR 100% of contributions deductible at the MTR (=EET) Contributions deductible at the MTR without any limit Contributions deductible at the MTR Credit equal to 21% of pre-tax contributions Credit equal to 2.76% of average earnings Contributions not taxed Notes: "E" stands for exempt, "T" for taxed and "MTR" for marginal tax rate. The calculations assume an average earner contributing 10% of wages from age 20 to 64 and withdrawing benefits from age 65 to 84 as a fixed-payment annuity. 37. When changing the tax relief on contributions from full tax deductibility to partial tax relief, the decrease in the tax advantage on contributions is partially compensated by a lower taxation of withdrawals. Figure 6 shows the decomposition of the overall tax advantage for the "EET" tax regime and equivalent approaches to provide partial tax relief on contributions (i.e. providing the same overall tax advantage). A change in the tax treatment of contributions (without changing the tax treatment of returns on investment and of withdrawals) has obviously an impact on the tax advantage on contributions, but also on the tax advantage on withdrawals, independently of the approach used to provide reduced tax relief on contributions. For example, when only 50% of contributions are tax deductible, the tax advantage on contributions decreases by half (from 30% for the "EET" tax regime to 15% for the partial deductibility approach), the tax advantage on returns remains constant (21%) and the tax advantage on withdrawals increases (from -30% to - 26%). Indeed, a reduction in the tax relief on contributions lowers the level of after-tax contributions paid in the pension plan, and as a consequence it also lowers the level of assets accumulated, the future level of pension benefits, and the amount of tax due on the latter.

15 DAF/AS/PEN/WD(2017)11 15 Figure 6. Overall tax advantage for the "EET" tax regime and different approaches to provide partial tax relief on contributions Taxes saved over a lifetime, as a percentage of the present value of contributions Contributions Returns Withdrawals Overall tax advantage 55% 45% 35% 25% 15% 5% -5% -15% -25% -35% EET Partial deduction Limited Deduction deductibility at fixed rate Fixed tax rate Tax credit, % Tax credit, flat Notes: "E" stands for exempt and "T" for taxed. The calculations assumes an individual contributing 10% of wages from age 20 to 64 and withdrawing benefits from age 65 to 84 as a fixed-payment annuity. All the approaches to provide partial tax relief on contributions are equivalent in terms of overall tax advantage and correspond to the case where 50% of contributions are tax deductible (see Table 1). 38. Tax credits can provide the same overall tax advantage than other approaches to provide partial tax relief on contributions with a lower tax advantage on contributions. As shown in Figure 6, as compared to other equivalent approaches to provide partial the tax relief on contributions, tax credits lead to a smaller tax advantage on contributions (10.5%, as opposed to 15% for the other approaches). This is compensated by lower taxes paid on withdrawals (-21% as opposed to -26% for the other approaches) to achieve the same overall tax advantage. Indeed, with tax credits, the amounted invested after tax in the pension plan is lower than with tax deductions or rate reliefs. With tax credits, contributions are first taxed at the individual's marginal tax rate and then the treasury reduces the amount of income tax due by the value of the tax credit. However, that tax credit is not invested in the plan. In contrast, with tax deductions and rate reliefs, contributions are only partially taxed and the amount corresponding to the tax saved on contributions is invested in the plan. Therefore, tax credits lead to lower assets accumulated at retirement and lower pension benefits, which means that the tax collected on withdrawals is also lower. As a consequence, it is not necessary with tax credits to give as much tax relief on contributions to achieve the same overall tax advantage. 39. In addition, tax credits reduce the level of after-tax pension benefits that individuals receive during retirement. As explained above, tax credits do not get invested in the pension plan. This translates into lower assets accumulated at retirement and then lower before-tax pension benefits, as compared to other approaches providing the same overall tax advantage. However, even after accounting for taxes due during the pay-out phase, after-tax pension benefits remain lower with tax credits. Therefore, individuals are

16 16 DAF/AS/PEN/WD(2017)11 worse-off with tax credits, even though they can save the same amount of taxes than with other approaches. 40. A fixed tax rate on contributions makes high-income earners better-off, while a tax credit expressed as a flat-rate amount and a tax deduction at fixed rate make lowincome earners better-off when tax rates increase with taxable income. Figure 7 illustrates how the overall tax advantage changes with the level of income for the "EET" tax regime and different approaches to provide partial tax relief on contributions. The different approaches are defined so that the average earner gets the same overall tax advantage, which corresponds to deducting 50% of contributions. None of the approaches are income neutral. Figure 7. Overall tax advantage for the "EET" tax regime and different approaches to provide partial tax relief on contributions, by income level Taxes saved over a lifetime, as a percentage of the present value of contributions EET Partial deduction Limited deductibility Tax credit (%) Tax credit (flat) Deduction fixed rate 35% 30% 25% 20% 15% 10% 5% 0% -5% -10% -15% Fixed tax rate Income level (multiple of average earnings) Notes: "E" stands for exempt and "T" for taxed. The calculations assumes an individual contributing 10% of wages from age 20 to 64 and withdrawing benefits from age 65 to 84 as a fixed-payment annuity. The tax system is such that tax rates increase with taxable income. All the strategies are equivalent in terms of overall tax advantage for the average earner and correspond to the case where 50% of contributions are tax deductible (see Table 1). 41. Reducing tax relief on contributions through partial tax deductibility lowers the overall tax advantage for all income groups, but hits middle- and high-income earners harder than low-income earners. As shown in Figure 7, the gap in the overall tax advantage between full tax deductibility ("EET") and partial deductibility of contributions increases with income. Indeed, low-income earners are the ones who benefit the least from the "EET" tax regime in the first place, because they already pay little or no income tax before the deduction of contributions. Reducing the portion of contributions that can be deducted from taxable income therefore has a lower impact for low-income earners.

17 DAF/AS/PEN/WD(2017) Introducing a ceiling on tax-deductible contributions reduces the amount of taxes saved only for individuals who contribute beyond the ceiling. The ceiling has no impact on the overall tax advantage as long as contributions remain below it. As the analysis assumes that all individuals contribute the same proportion of their earnings (10%), the likelihood of reaching the ceiling increases with income. In addition, as the ceiling is set as a nominal flat-rate amount, the part of taxable contributions increases with the income level, thereby reducing the tax advantage on contributions and the overall tax advantage as income increases. 43. Tax credits expressed as a proportion of contributions and tax deductions at fixed rate allow smoothing out the overall tax advantage across income groups. With both approaches, the tax advantage on contributions is income neutral because the tax relief is provided at the same rate for everyone (the tax credit rate and the deduction rate respectively are fixed). However, very low-income earners who pay little or no income tax benefit less from the tax credit because of its non-refundable nature. It is also interesting to note that low-income earners would save more taxes by deducting their contributions at a fixed rate than by deducting them at their marginal rate (as with the "EET" tax regime) if the fixed deduction rate is higher than their marginal income tax rate. 44. Tax credits expressed as a flat-rate amount target tax advantages at low-income earners. As the value of the tax credit is fixed across the income scale, its value in relative terms declines with the income level. Unfortunately, individuals paying little or no income tax do not benefit as much because the credit is non-refundable. 45. On the contrary, taxing contributions at a fixed rate targets tax advantages at highincome earners. Indeed, as the income level increases, the gap between the fixed tax rate and the individual's marginal income tax rate increases, leading to a larger tax advantage on contributions and thereby a larger overall tax advantage. However, this approach could act as a disincentive for very low-income earners if the fixed tax rate exceeds their marginal income tax rate (negative overall tax advantage for individuals earning 20% of average earnings, Figure 7). 46. Finally, most of the approaches to provide partial tax relief on contributions are income neutral in tax systems where individuals have their entire income taxed at the same rate. Introducing a tax-deducibility ceiling or a tax credit calculated as a flat-rate amount would still offer the largest overall tax advantage to low-income earners. In addition, in personal income tax systems where all individuals pay tax on their first unit of income, even very low-income earners can benefit from tax credits. 4. Evolution of the fiscal cost when changing the approach to design incentives for retirement savings 47. This section looks at the evolution of the fiscal cost for the treasury when changing the approach to design incentives for retirement savings. The previous section has looked at the tax advantage in terms of taxes saved that alternative approaches to design incentives for retirement savings provide to individuals. This section assesses the fiscal cost for the treasury in order to have an overall view of the impact of these alternative approaches. 48. The analysis focuses on the transition from the most common approach used currently (exempting contributions and returns on investment while taxing withdrawals, "EET") to alternative approaches. The fiscal cost is assessed through the calculation of

18 18 DAF/AS/PEN/WD(2017)11 the net tax expenditure related to the preferential tax treatment of retirement savings and of direct spending related to financial incentives. The methodology used to calculate the net tax expenditure is the same as the one used in previous papers. The analysis calculates the net tax expenditure associated with the tax treatment of private pension plans, today and into the future, based on the revenue foregone method and using the cash-flow approach. The revenue foregone method measures the amount by which tax revenues are reduced by a particular tax concession, under the assumption of unchanged behaviour from the taxpayer. The cash-flow approach estimates the effect on government cashflows in a given year. The net tax expenditure in a given year is calculated as the net amount of tax revenue foregone on contributions, tax revenue foregone on accrued income and tax revenue collected on withdrawals. The section first assesses the impact of a transition from the "EET" tax regime to another tax or financial incentive and then looks at the impact of reducing tax relief on contributions following different approaches. Transition to alternative tax and financial incentives 49. In the long term, tax incentives that provide a higher overall tax advantage to the individual than the "EET" tax regime also entail a higher fiscal cost for the treasury. Figure 8 shows projections of the net tax expenditure and its components (revenue foregone on contributions, revenue foregone on returns and revenue collected on withdrawals) for a pension plan with "EET" tax incentives. The pension plan is mature in year t and replaced five years later, at, by an alternative tax incentive approach. 8 The left panel of Figure 8 shows the total net tax expenditure (solid line), as well as the net tax expenditure resulting from the "EET" tax regime (blue dotted line) and the alternative tax regime (red dotted line). The right panel shows tax revenue foregone on contribution (blue line), tax revenue foregone on returns (green line) and tax collected on withdrawals (red line) resulting from the transition between the two regimes. Tax regimes providing a higher tax advantage to individuals than the current "EET" approach translate into a higher fiscal cost once the new system has reached maturity. For example, in the case of the tax regime where none of the flows are taxed ("EEE" tax regime), the net tax expenditure increases as the proportion of taxable pension income decreases over time. 50. In the short term however, the fiscal cost may actually decline temporarily. The transition to either of the "ETE", "TET" or "TTE" tax regimes would eventually translate into a larger net tax expenditure in the long run, but in the short term, the net tax expenditure would actually go down. This is due to the phasing out of the "EET" tax regime. Indeed, for the three alternative tax incentives, some of the fiscal costs are reduced overnight when the new system is introduced at, as contributions stop being paid in the "EET" plan and start building assets in the new plan: the treasury starts collecting taxes on returns on investment for the "EtE" and "tte" tax regimes; there is a reduction in tax revenue foregone on contributions for the "tet" and "tte" tax regimes. At the same time, revenues collected on withdrawals are still high at, as all retirees paying taxes on their pension benefits have contributed to the "EET" system. Over time, tax revenue collected on withdrawals declines slowly (to zero for the ETE" and "TTE" tax regimes), resulting in an increase in the net tax expenditure. 8 The tax parameters are the same as those presented in Box 1.